From 3333e9ac0f80a2a9684fdfd29dd8ed7e253559c4 Mon Sep 17 00:00:00 2001
From: DCCONSTRUCTIONS <dcconstructions@MacBook-Pro-DCCONSTRUCTIONS-2.local>
Date: Sat, 25 Jul 2026 10:26:07 +0300
Subject: [PATCH] feat(lidar): add RAVNOVES field review

---
 .../src/core/lidar/replayQuality.ts           | 500 ++++++++++
 .../control-station/src/styles/responsive.css |  32 +
 .../control-station/src/styles/workspaces.css | 212 +++++
 .../src/workspaces/LidarGroundPointCloud.tsx  |  36 +-
 .../src/workspaces/LidarQualityWorkspace.tsx  | 275 +++++-
 .../test/lidarReplayQuality.test.mjs          | 213 +++++
 docs/10_EXTERNAL_PERCEPTION_WORKER.md         |  11 +
 docs/13_LIDAR_WORKER_PRODUCT_AND_ROADMAP.md   |  22 +
 .../perception/run_lidar_field_review.py      | 182 ++++
 src/k1link/compute/__init__.py                |  30 +
 src/k1link/compute/lidar_field_review.py      | 891 ++++++++++++++++++
 src/k1link/web/lidar_api.py                   | 147 ++-
 tests/test_lidar_field_review.py              | 238 +++++
 13 files changed, 2775 insertions(+), 14 deletions(-)
 create mode 100644 experiments/perception/run_lidar_field_review.py
 create mode 100644 src/k1link/compute/lidar_field_review.py
 create mode 100644 tests/test_lidar_field_review.py

diff --git a/apps/control-station/src/core/lidar/replayQuality.ts b/apps/control-station/src/core/lidar/replayQuality.ts
index 583aa92..cf4130a 100644
--- a/apps/control-station/src/core/lidar/replayQuality.ts
+++ b/apps/control-station/src/core/lidar/replayQuality.ts
@@ -147,6 +147,106 @@ export interface LidarGroundFrame {
   groundTruth: false;
 }
 
+export interface LidarFieldReviewWindowSummary {
+  index: number;
+  key: string;
+  label: string;
+  startSeconds: number;
+  endSeconds: number;
+  midpointSeconds: number;
+  sourceLidarSamples: number;
+  sourcePointCount: number;
+  displayPointCount: number;
+  sourceFrameStart: number;
+  sourceFrameEnd: number;
+  previewSourceFrameIndex: number;
+  previewSessionSeconds: number;
+}
+
+export interface LidarFieldReview {
+  reviewId: string;
+  displayName: string;
+  sessionId: string;
+  sourcePackId: string;
+  status: "diagnostic-only";
+  source: {
+    timelineStartSeconds: number;
+    timelineEndSeconds: number;
+    availableLidarFrames: number;
+    pointCount: number;
+    representation: "legacy-e10-vendor-map-with-pose";
+    intensityAvailable: false;
+    rawScanAccepted: false;
+  };
+  selection: {
+    purpose: "operator-readable-central-urban-field-review";
+    defaultWindowIndex: number;
+    accumulation: "per-source-frame masks accumulated in map frame";
+    maximumPointsPerWindow: number;
+  };
+  windows: LidarFieldReviewWindowSummary[];
+  metrics: {
+    sourceSamples: number;
+    current: {
+      provider: LidarGroundProviderSummary;
+      groundFraction: LidarDistribution;
+      latencyMs: LidarDistribution;
+    };
+    candidate: {
+      provider: LidarGroundProviderSummary;
+      groundFraction: LidarDistribution;
+      latencyMs: LidarDistribution;
+    };
+    comparison: {
+      algorithmGroundIou: LidarDistribution;
+      groundDisagreementFraction: LidarDistribution;
+      isAccuracyMetric: false;
+    };
+  };
+  decision: {
+    status: "visual-review-only";
+    productionPromotion: false;
+    reasons: string[];
+  };
+  createdAtUtc: string | null;
+  groundTruth: false;
+  authority: {
+    commandsEnabled: false;
+    navigationOrSafetyAccepted: false;
+  };
+}
+
+export interface LidarFieldReviewCatalog {
+  configured: boolean;
+  validTotal: number;
+  invalidTotal: number;
+  items: LidarFieldReview[];
+}
+
+export interface LidarFieldReviewWindow {
+  reviewId: string;
+  displayName: string;
+  sessionId: string;
+  sourcePackId: string;
+  windowIndex: number;
+  windowCount: number;
+  window: LidarFieldReviewWindowSummary;
+  pointCount: number;
+  coordinateFrame: "map";
+  distanceUnit: "m";
+  pointsXyzM: Array<[number, number, number]>;
+  intensity0To255: null;
+  intensity: {
+    available: false;
+    reason: string;
+  };
+  masks: LidarGroundFrame["masks"];
+  counts: LidarGroundFrame["counts"];
+  previewUrl: string;
+  groundTruth: false;
+  authority: LidarFieldReview["authority"];
+}
+
 export class LidarReplayContractError extends Error {}
 
 export class LidarReplayApiError extends Error {
@@ -162,7 +262,10 @@ type LidarFetch = (
 
 const SAFE_PACK_ID = /^lidar-replay-pack-[a-f0-9]{64}$/;
 const SAFE_GROUND_BENCHMARK_ID = /^ground-benchmark-[a-f0-9]{64}$/;
+const SAFE_FIELD_REVIEW_ID = /^lidar-field-review-[a-f0-9]{64}$/;
+const SAFE_E10_PACK_ID = /^e10-lidar-pack-[a-f0-9]{64}$/;
 const SAFE_ID = /^[A-Za-z0-9][A-Za-z0-9._:/-]{0,159}$/;
+const SAFE_FIELD_KEY = /^[a-z0-9][a-z0-9-]{0,63}$/;
 const SHA256 = /^[a-f0-9]{64}$/;
 const GIT_SHA1 = /^[a-f0-9]{40}$/;
 
@@ -641,6 +744,358 @@ export function parseLidarGroundFrame(value: unknown): LidarGroundFrame {
   };
 }
 
+function fieldReviewWindowSummary(
+  value: unknown,
+  expectedIndex: number,
+): LidarFieldReviewWindowSummary {
+  const source = record(value, `field-review window ${expectedIndex}`);
+  const index = integer(source.index, "window.index");
+  const startSeconds = number(source.start_seconds, "window.start_seconds") ?? 0;
+  const endSeconds = number(source.end_seconds, "window.end_seconds") ?? 0;
+  if (index !== expectedIndex || startSeconds >= endSeconds) {
+    throw new LidarReplayContractError("LiDAR field-review window несовместим");
+  }
+  return {
+    index,
+    key: string(source.key, "window.key", SAFE_FIELD_KEY),
+    label: string(source.label, "window.label"),
+    startSeconds,
+    endSeconds,
+    midpointSeconds:
+      number(source.midpoint_seconds, "window.midpoint_seconds") ?? 0,
+    sourceLidarSamples: integer(
+      source.source_lidar_samples,
+      "window.source_lidar_samples",
+    ),
+    sourcePointCount: integer(
+      source.source_point_count,
+      "window.source_point_count",
+    ),
+    displayPointCount: integer(
+      source.display_point_count,
+      "window.display_point_count",
+    ),
+    sourceFrameStart: integer(
+      source.source_frame_start,
+      "window.source_frame_start",
+    ),
+    sourceFrameEnd: integer(
+      source.source_frame_end,
+      "window.source_frame_end",
+    ),
+    previewSourceFrameIndex: integer(
+      source.preview_source_frame_index,
+      "window.preview_source_frame_index",
+    ),
+    previewSessionSeconds:
+      number(
+        source.preview_session_seconds,
+        "window.preview_session_seconds",
+      ) ?? 0,
+  };
+}
+
+function fieldReviewBranch(
+  value: unknown,
+  label: string,
+): LidarFieldReview["metrics"]["current"] {
+  const source = record(value, label);
+  return {
+    provider: groundProvider(source.provider, `${label}.provider`),
+    groundFraction: distribution(
+      source.ground_fraction,
+      `${label}.ground_fraction`,
+    ),
+    latencyMs: distribution(source.latency_ms, `${label}.latency_ms`),
+  };
+}
+
+function fieldReview(value: unknown): LidarFieldReview {
+  const source = record(value, "LiDAR field review");
+  const sourceEvidence = record(source.source, "field-review source");
+  const selection = record(source.selection, "field-review selection");
+  const sampling = record(selection.sampling, "field-review sampling");
+  const metrics = record(source.metrics, "field-review metrics");
+  const comparison = record(metrics.comparison, "field-review comparison");
+  const decision = record(source.decision, "field-review decision");
+  const authority = record(source.authority, "field-review authority");
+  if (
+    source.status !== "diagnostic-only"
+    || source.ground_truth !== false
+    || sourceEvidence.representation !== "legacy-e10-vendor-map-with-pose"
+    || sourceEvidence.intensity_available !== false
+    || sourceEvidence.raw_scan_accepted !== false
+    || selection.purpose !== "operator-readable-central-urban-field-review"
+    || selection.accumulation !== "per-source-frame masks accumulated in map frame"
+    || sampling.method !== "uniform-point-index-per-window"
+    || comparison.is_accuracy_metric !== false
+    || decision.status !== "visual-review-only"
+    || decision.production_promotion !== false
+    || authority.commands_enabled !== false
+    || authority.navigation_or_safety_accepted !== false
+  ) {
+    throw new LidarReplayContractError(
+      "LiDAR field review завышает readiness или меняет evidence",
+    );
+  }
+  const windows = array(source.windows, "field-review windows").map(
+    fieldReviewWindowSummary,
+  );
+  const defaultWindowIndex = integer(
+    selection.default_window_index,
+    "selection.default_window_index",
+  );
+  const maximumPointsPerWindow = integer(
+    sampling.maximum_points_per_window,
+    "sampling.maximum_points_per_window",
+  );
+  if (
+    windows.length < 1
+    || defaultWindowIndex >= windows.length
+    || maximumPointsPerWindow < 1
+    || maximumPointsPerWindow > 80_000
+    || windows.some(
+      (window) =>
+        window.sourceLidarSamples < 1
+        || window.sourcePointCount < window.displayPointCount
+        || window.displayPointCount < 1
+        || window.displayPointCount > maximumPointsPerWindow,
+    )
+  ) {
+    throw new LidarReplayContractError("LiDAR field-review selection несовместим");
+  }
+  return {
+    reviewId: string(source.review_id, "review_id", SAFE_FIELD_REVIEW_ID),
+    displayName: string(source.display_name, "display_name"),
+    sessionId: string(source.session_id, "session_id", SAFE_ID),
+    sourcePackId: string(
+      source.source_pack_id,
+      "source_pack_id",
+      SAFE_E10_PACK_ID,
+    ),
+    status: "diagnostic-only",
+    source: {
+      timelineStartSeconds:
+        number(
+          sourceEvidence.timeline_start_seconds,
+          "source.timeline_start_seconds",
+        ) ?? 0,
+      timelineEndSeconds:
+        number(
+          sourceEvidence.timeline_end_seconds,
+          "source.timeline_end_seconds",
+        ) ?? 0,
+      availableLidarFrames: integer(
+        sourceEvidence.available_lidar_frames,
+        "source.available_lidar_frames",
+      ),
+      pointCount: integer(sourceEvidence.point_count, "source.point_count"),
+      representation: "legacy-e10-vendor-map-with-pose",
+      intensityAvailable: false,
+      rawScanAccepted: false,
+    },
+    selection: {
+      purpose: "operator-readable-central-urban-field-review",
+      defaultWindowIndex,
+      accumulation: "per-source-frame masks accumulated in map frame",
+      maximumPointsPerWindow,
+    },
+    windows,
+    metrics: {
+      sourceSamples: integer(metrics.source_samples, "metrics.source_samples"),
+      current: fieldReviewBranch(metrics.current, "metrics.current"),
+      candidate: fieldReviewBranch(metrics.candidate, "metrics.candidate"),
+      comparison: {
+        algorithmGroundIou: distribution(
+          comparison.algorithm_to_algorithm_ground_iou,
+          "metrics.comparison.algorithm_ground_iou",
+        ),
+        groundDisagreementFraction: distribution(
+          comparison.ground_disagreement_fraction,
+          "metrics.comparison.ground_disagreement_fraction",
+        ),
+        isAccuracyMetric: false,
+      },
+    },
+    decision: {
+      status: "visual-review-only",
+      productionPromotion: false,
+      reasons: array(decision.reasons, "decision.reasons").map((reason) =>
+        string(reason, "decision.reason")
+      ),
+    },
+    createdAtUtc:
+      source.created_at_utc === null || source.created_at_utc === undefined
+        ? null
+        : string(source.created_at_utc, "created_at_utc"),
+    groundTruth: false,
+    authority: {
+      commandsEnabled: false,
+      navigationOrSafetyAccepted: false,
+    },
+  };
+}
+
+export function parseLidarFieldReviewCatalog(
+  value: unknown,
+): LidarFieldReviewCatalog {
+  const source = record(value, "LiDAR field-review catalog");
+  if (
+    source.schema_version !== "missioncore.lidar-field-review-catalog/v1"
+    || source.access !== "read-only"
+  ) {
+    throw new LidarReplayContractError("LiDAR field-review catalog несовместим");
+  }
+  return {
+    configured: boolean(source.configured, "configured"),
+    validTotal: integer(source.valid_total, "valid_total"),
+    invalidTotal: integer(source.invalid_total, "invalid_total"),
+    items: array(source.items, "items").map(fieldReview),
+  };
+}
+
+export function parseLidarFieldReviewWindow(
+  value: unknown,
+): LidarFieldReviewWindow {
+  const source = record(value, "LiDAR field-review window");
+  const reviewId = string(
+    source.review_id,
+    "review_id",
+    SAFE_FIELD_REVIEW_ID,
+  );
+  if (
+    source.schema_version !== "missioncore.lidar-field-review-window/v1"
+    || source.access !== "read-only"
+    || source.ground_truth !== false
+    || source.coordinate_frame !== "map"
+    || source.distance_unit !== "m"
+  ) {
+    throw new LidarReplayContractError("LiDAR field-review window несовместим");
+  }
+  const authority = record(source.authority, "authority");
+  const intensity = record(source.intensity, "intensity");
+  if (
+    authority.commands_enabled !== false
+    || authority.navigation_or_safety_accepted !== false
+    || intensity.available !== false
+  ) {
+    throw new LidarReplayContractError("LiDAR field-review authority несовместим");
+  }
+  const pointCount = integer(source.point_count, "point_count");
+  if (pointCount < 1 || pointCount > 80_000) {
+    throw new LidarReplayContractError("LiDAR field-review window слишком большой");
+  }
+  const points = array(source.points_xyz_m, "points_xyz_m");
+  if (points.length !== pointCount) {
+    throw new LidarReplayContractError("Количество field-review points не совпадает");
+  }
+  const pointsXyzM = points.map((value, index): [number, number, number] => {
+    const tuple = array(value, `points_xyz_m[${index}]`);
+    if (tuple.length !== 3) {
+      throw new LidarReplayContractError("LiDAR point должен содержать XYZ");
+    }
+    return [
+      number(tuple[0], `points_xyz_m[${index}].x`) ?? 0,
+      number(tuple[1], `points_xyz_m[${index}].y`) ?? 0,
+      number(tuple[2], `points_xyz_m[${index}].z`) ?? 0,
+    ];
+  });
+  const masks = record(source.masks, "masks");
+  const currentGround = groundMask(
+    masks.current_ground,
+    "masks.current_ground",
+    pointCount,
+  );
+  const currentAssigned = groundMask(
+    masks.current_assigned,
+    "masks.current_assigned",
+    pointCount,
+  );
+  const candidateGround = groundMask(
+    masks.candidate_ground,
+    "masks.candidate_ground",
+    pointCount,
+  );
+  const candidateAssigned = groundMask(
+    masks.candidate_assigned,
+    "masks.candidate_assigned",
+    pointCount,
+  );
+  const disagreement = groundMask(
+    masks.disagreement,
+    "masks.disagreement",
+    pointCount,
+  );
+  const counts = record(source.counts, "counts");
+  const parsedCounts = {
+    currentGround: integer(counts.current_ground, "counts.current_ground"),
+    candidateGround: integer(
+      counts.candidate_ground,
+      "counts.candidate_ground",
+    ),
+    disagreement: integer(counts.disagreement, "counts.disagreement"),
+  };
+  const windowIndex = integer(source.window_index, "window_index");
+  const windowCount = integer(source.window_count, "window_count");
+  const window = fieldReviewWindowSummary(source.window, windowIndex);
+  const expectedPreviewUrl =
+    `/api/v1/lidar/field-reviews/${reviewId}/windows/${windowIndex}/preview`;
+  if (
+    windowCount < 1
+    || windowIndex >= windowCount
+    || window.displayPointCount !== pointCount
+    || parsedCounts.currentGround
+      !== currentGround.reduce((sum, item) => sum + item, 0)
+    || parsedCounts.candidateGround
+      !== candidateGround.reduce((sum, item) => sum + item, 0)
+    || parsedCounts.disagreement
+      !== disagreement.reduce((sum, item) => sum + item, 0)
+    || disagreement.some(
+      (item, index) =>
+        item !== Number(currentGround[index] !== candidateGround[index]),
+    )
+    || source.preview_url !== expectedPreviewUrl
+  ) {
+    throw new LidarReplayContractError("LiDAR field-review content несовместим");
+  }
+  return {
+    reviewId,
+    displayName: string(source.display_name, "display_name"),
+    sessionId: string(source.session_id, "session_id", SAFE_ID),
+    sourcePackId: string(
+      source.source_pack_id,
+      "source_pack_id",
+      SAFE_E10_PACK_ID,
+    ),
+    windowIndex,
+    windowCount,
+    window,
+    pointCount,
+    coordinateFrame: "map",
+    distanceUnit: "m",
+    pointsXyzM,
+    intensity0To255: null,
+    intensity: {
+      available: false,
+      reason: string(intensity.reason, "intensity.reason"),
+    },
+    masks: {
+      currentGround,
+      currentAssigned,
+      candidateGround,
+      candidateAssigned,
+      disagreement,
+    },
+    counts: parsedCounts,
+    previewUrl: expectedPreviewUrl,
+    groundTruth: false,
+    authority: {
+      commandsEnabled: false,
+      navigationOrSafetyAccepted: false,
+    },
+  };
+}
+
 async function responseJson(
   response: Response,
   fallback: string,
@@ -739,3 +1194,48 @@ export async function fetchLidarGroundFrame(
     await responseJson(response, "Не удалось получить LiDAR ground frame."),
   );
 }
+
+export async function fetchLidarFieldReviews(
+  options: { signal?: AbortSignal; fetcher?: LidarFetch } = {},
+): Promise<LidarFieldReviewCatalog> {
+  const fetcher = options.fetcher ?? fetch;
+  const response = await fetcher("/api/v1/lidar/field-reviews?limit=10", {
+    method: "GET",
+    headers: { Accept: "application/json" },
+    signal: options.signal,
+  });
+  return parseLidarFieldReviewCatalog(
+    await responseJson(response, "Не удалось получить полевой LiDAR review."),
+  );
+}
+
+export async function fetchLidarFieldReviewWindow(
+  reviewId: string,
+  windowIndex: number,
+  options: { signal?: AbortSignal; fetcher?: LidarFetch } = {},
+): Promise<LidarFieldReviewWindow> {
+  if (
+    !SAFE_FIELD_REVIEW_ID.test(reviewId)
+    || !Number.isInteger(windowIndex)
+    || windowIndex < 0
+  ) {
+    throw new LidarReplayContractError(
+      "Некорректное окно полевого LiDAR review",
+    );
+  }
+  const fetcher = options.fetcher ?? fetch;
+  const response = await fetcher(
+    `/api/v1/lidar/field-reviews/${reviewId}/windows/${windowIndex}`,
+    {
+      method: "GET",
+      headers: { Accept: "application/json" },
+      signal: options.signal,
+    },
+  );
+  return parseLidarFieldReviewWindow(
+    await responseJson(
+      response,
+      "Не удалось получить окно полевого LiDAR review.",
+    ),
+  );
+}
diff --git a/apps/control-station/src/styles/responsive.css b/apps/control-station/src/styles/responsive.css
index 46273f8..459c1b3 100644
--- a/apps/control-station/src/styles/responsive.css
+++ b/apps/control-station/src/styles/responsive.css
@@ -18,6 +18,23 @@
     grid-template-columns: 1fr;
   }
 
+  .lidar-field-review__source {
+    grid-template-columns: repeat(2, minmax(0, 1fr));
+  }
+
+  .lidar-field-window-list {
+    grid-template-columns: repeat(3, minmax(0, 1fr));
+  }
+
+  .lidar-field-stage {
+    grid-template-columns: 1fr;
+  }
+
+  .lidar-field-camera img {
+    min-height: 0;
+    aspect-ratio: 4 / 3;
+  }
+
   .polygon-run-providers {
     grid-template-columns: repeat(2, minmax(0, 1fr));
   }
@@ -202,6 +219,21 @@
     display: none;
   }
 
+  .lidar-field-review__source,
+  .lidar-field-window-list {
+    grid-template-columns: 1fr;
+  }
+
+  .lidar-field-cloud > header,
+  .lidar-field-review__explanation {
+    align-items: stretch;
+    grid-template-columns: 1fr;
+  }
+
+  .lidar-field-cloud > header {
+    flex-direction: column;
+  }
+
   .polygon-run-identity dl {
     grid-template-columns: 1fr;
   }
diff --git a/apps/control-station/src/styles/workspaces.css b/apps/control-station/src/styles/workspaces.css
index 6e59507..550e19c 100644
--- a/apps/control-station/src/styles/workspaces.css
+++ b/apps/control-station/src/styles/workspaces.css
@@ -954,6 +954,218 @@
   margin-top: 0.8rem;
 }
 
+.lidar-field-review {
+  min-width: 0;
+  border-color: rgb(74 215 255 / 0.22);
+  background:
+    radial-gradient(circle at 8% 0%, rgb(56 124 255 / 0.13), transparent 30rem),
+    var(--station-panel);
+}
+
+.lidar-field-review__heading p,
+.lidar-field-review__explanation p,
+.lidar-field-review__empty p {
+  margin: 0.28rem 0 0;
+  max-width: 49rem;
+  color: var(--nodedc-text-muted);
+  font-size: 0.64rem;
+  line-height: 1.5;
+}
+
+.lidar-field-review__source {
+  display: grid;
+  grid-template-columns: minmax(13rem, 1.4fr) repeat(3, minmax(0, 0.72fr));
+  gap: 0.55rem;
+  margin-top: 1rem;
+}
+
+.lidar-field-review__source > div {
+  display: grid;
+  min-width: 0;
+  gap: 0.28rem;
+  border: 1px solid var(--station-hairline);
+  border-radius: 0.72rem;
+  background: rgb(255 255 255 / 0.025);
+  padding: 0.64rem 0.7rem;
+}
+
+.lidar-field-review__source span,
+.lidar-field-window-list small,
+.lidar-field-stage__label span,
+.lidar-field-camera footer,
+.lidar-field-stage__pending {
+  color: var(--nodedc-text-muted);
+  font-size: 0.59rem;
+  line-height: 1.4;
+}
+
+.lidar-field-review__source strong {
+  overflow: hidden;
+  color: var(--nodedc-text-primary);
+  font-size: 0.68rem;
+  text-overflow: ellipsis;
+  white-space: nowrap;
+}
+
+.lidar-field-window-list {
+  display: grid;
+  grid-template-columns: repeat(5, minmax(0, 1fr));
+  gap: 0.48rem;
+  margin-top: 0.65rem;
+}
+
+.lidar-field-window-list > button {
+  display: flex;
+  min-width: 0;
+  align-items: flex-start;
+  gap: 0.5rem;
+  border: 1px solid var(--station-hairline);
+  border-radius: 0.78rem;
+  background: rgb(255 255 255 / 0.02);
+  padding: 0.62rem;
+  color: inherit;
+  font: inherit;
+  text-align: left;
+  cursor: pointer;
+}
+
+.lidar-field-window-list > button:hover,
+.lidar-field-window-list > button[data-selected="true"] {
+  border-color: rgb(74 215 255 / 0.48);
+  background: rgb(74 215 255 / 0.08);
+}
+
+.lidar-field-window-list > button > span {
+  display: grid;
+  flex: 0 0 auto;
+  width: 1.35rem;
+  height: 1.35rem;
+  place-items: center;
+  border-radius: 50%;
+  background: rgb(74 215 255 / 0.12);
+  color: #78e3ff;
+  font-size: 0.59rem;
+  font-weight: 700;
+}
+
+.lidar-field-window-list > button > div {
+  display: grid;
+  min-width: 0;
+  gap: 0.18rem;
+}
+
+.lidar-field-window-list strong {
+  color: var(--nodedc-text-primary);
+  font-size: 0.61rem;
+  line-height: 1.35;
+}
+
+.lidar-field-stage {
+  display: grid;
+  grid-template-columns: minmax(18rem, 0.68fr) minmax(0, 1.32fr);
+  gap: 0.65rem;
+  margin-top: 0.7rem;
+}
+
+.lidar-field-camera,
+.lidar-field-cloud {
+  overflow: hidden;
+  min-width: 0;
+  border: 1px solid rgb(255 255 255 / 0.09);
+  border-radius: 0.9rem;
+  background: #071018;
+}
+
+.lidar-field-camera {
+  display: grid;
+  grid-template-rows: auto minmax(0, 1fr) auto;
+}
+
+.lidar-field-camera > .lidar-field-stage__label,
+.lidar-field-cloud > header {
+  min-height: 3.45rem;
+  border-bottom: 1px solid rgb(255 255 255 / 0.08);
+  padding: 0.66rem 0.72rem;
+}
+
+.lidar-field-stage__label {
+  display: grid;
+  gap: 0.18rem;
+}
+
+.lidar-field-stage__label strong {
+  color: var(--nodedc-text-primary);
+  font-size: 0.68rem;
+}
+
+.lidar-field-camera img {
+  display: block;
+  width: 100%;
+  height: 100%;
+  min-height: 25rem;
+  object-fit: contain;
+}
+
+.lidar-field-camera footer {
+  display: flex;
+  justify-content: space-between;
+  gap: 0.5rem;
+  border-top: 1px solid rgb(255 255 255 / 0.08);
+  padding: 0.52rem 0.7rem;
+}
+
+.lidar-field-stage__pending {
+  display: grid;
+  min-height: 25rem;
+  place-items: center;
+  padding: 1rem;
+  text-align: center;
+}
+
+.lidar-field-cloud > header {
+  display: flex;
+  align-items: center;
+  justify-content: space-between;
+  gap: 0.65rem;
+}
+
+.lidar-field-cloud .lidar-ground-scene {
+  min-height: 28rem;
+  border: 0;
+  border-radius: 0;
+}
+
+.lidar-field-cloud .lidar-ground-scene-placeholder {
+  min-height: 28rem;
+  border: 0;
+  border-radius: 0;
+}
+
+.lidar-field-review__explanation {
+  display: grid;
+  grid-template-columns: auto minmax(15rem, 1fr) auto;
+  align-items: center;
+  gap: 0.65rem;
+  margin-top: 0.68rem;
+  border-top: 1px solid var(--station-hairline);
+  padding-top: 0.68rem;
+}
+
+.lidar-field-review__explanation p {
+  margin: 0;
+}
+
+.lidar-field-review__empty {
+  display: flex;
+  align-items: center;
+  gap: 0.65rem;
+  margin-top: 0.8rem;
+}
+
+.lidar-field-review__empty p {
+  margin: 0;
+}
+
 .lidar-ground-benchmark {
   min-width: 0;
 }
diff --git a/apps/control-station/src/workspaces/LidarGroundPointCloud.tsx b/apps/control-station/src/workspaces/LidarGroundPointCloud.tsx
index e89aa15..090eb6d 100644
--- a/apps/control-station/src/workspaces/LidarGroundPointCloud.tsx
+++ b/apps/control-station/src/workspaces/LidarGroundPointCloud.tsx
@@ -2,16 +2,27 @@ import { useEffect, useRef, useState } from "react";
 import * as THREE from "three";
 import { OrbitControls } from "three/addons/controls/OrbitControls.js";
 
-import type { LidarGroundFrame } from "../core/lidar/replayQuality";
-
 export type LidarGroundViewMode =
   | "intensity"
   | "current"
   | "candidate"
   | "disagreement";
 
+export interface LidarGroundPointCloudFrame {
+  pointCount: number;
+  pointsXyzM: Array<[number, number, number]>;
+  intensity0To255: number[] | null;
+  masks: {
+    currentGround: number[];
+    currentAssigned: number[];
+    candidateGround: number[];
+    candidateAssigned: number[];
+    disagreement: number[];
+  };
+}
+
 interface LidarGroundPointCloudProps {
-  frame: LidarGroundFrame;
+  frame: LidarGroundPointCloudFrame;
   mode: LidarGroundViewMode;
 }
 
@@ -28,7 +39,7 @@ function setRgb(
 }
 
 function frameColors(
-  frame: LidarGroundFrame,
+  frame: LidarGroundPointCloudFrame,
   mode: LidarGroundViewMode,
 ): Float32Array {
   const colors = new Float32Array(frame.pointCount * 3);
@@ -38,7 +49,7 @@ function frameColors(
     const candidate = frame.masks.candidateGround[index] === 1;
     const candidateAssigned = frame.masks.candidateAssigned[index] === 1;
     if (mode === "intensity") {
-      const intensity = frame.intensity0To255[index] / 255;
+      const intensity = (frame.intensity0To255?.[index] ?? 96) / 255;
       setRgb(
         colors,
         offset,
@@ -88,6 +99,8 @@ export function LidarGroundPointCloud({
   const materialRef = useRef<THREE.PointsMaterial | null>(null);
   const cameraRef = useRef<THREE.PerspectiveCamera | null>(null);
   const controlsRef = useRef<OrbitControls | null>(null);
+  const fogRef = useRef<THREE.FogExp2 | null>(null);
+  const gridRef = useRef<THREE.GridHelper | null>(null);
   const [renderError, setRenderError] = useState<string | null>(null);
 
   useEffect(() => {
@@ -115,7 +128,9 @@ export function LidarGroundPointCloud({
     host.prepend(renderer.domElement);
 
     const scene = new THREE.Scene();
-    scene.fog = new THREE.FogExp2(0x071018, 0.035);
+    const fog = new THREE.FogExp2(0x071018, 0.035);
+    scene.fog = fog;
+    fogRef.current = fog;
     const camera = new THREE.PerspectiveCamera(48, 1, 0.01, 1_000);
     camera.position.set(6, 4.5, 6);
     cameraRef.current = camera;
@@ -147,6 +162,7 @@ export function LidarGroundPointCloud({
     scene.add(new THREE.Points(geometry, material));
 
     const grid = new THREE.GridHelper(24, 48, 0x3c7cff, 0x233747);
+    gridRef.current = grid;
     const gridMaterials = Array.isArray(grid.material)
       ? grid.material
       : [grid.material];
@@ -198,6 +214,8 @@ export function LidarGroundPointCloud({
       materialRef.current = null;
       cameraRef.current = null;
       controlsRef.current = null;
+      fogRef.current = null;
+      gridRef.current = null;
     };
   }, []);
 
@@ -206,7 +224,9 @@ export function LidarGroundPointCloud({
     const material = materialRef.current;
     const camera = cameraRef.current;
     const controls = controlsRef.current;
-    if (!geometry || !material || !camera || !controls) return;
+    const fog = fogRef.current;
+    const grid = gridRef.current;
+    if (!geometry || !material || !camera || !controls || !fog || !grid) return;
 
     const positions = new Float32Array(frame.pointCount * 3);
     let minimumX = Number.POSITIVE_INFINITY;
@@ -235,6 +255,8 @@ export function LidarGroundPointCloud({
     geometry.computeBoundingSphere();
     const radius = Math.max(geometry.boundingSphere?.radius ?? 1, 0.2);
     material.size = THREE.MathUtils.clamp(radius / 155, 0.014, 0.075);
+    fog.density = THREE.MathUtils.clamp(0.18 / radius, 0.0008, 0.035);
+    grid.scale.setScalar(Math.max(radius / 12, 1));
 
     const targetHeight = Math.max((maximumZ - minimumZ) * 0.35, 0.15);
     const distance = Math.max(radius * 1.8, 1.2);
diff --git a/apps/control-station/src/workspaces/LidarQualityWorkspace.tsx b/apps/control-station/src/workspaces/LidarQualityWorkspace.tsx
index 3b84aee..c5047f3 100644
--- a/apps/control-station/src/workspaces/LidarQualityWorkspace.tsx
+++ b/apps/control-station/src/workspaces/LidarQualityWorkspace.tsx
@@ -6,10 +6,14 @@ import {
 } from "@nodedc/ui-react";
 
 import {
+  fetchLidarFieldReviews,
+  fetchLidarFieldReviewWindow,
   fetchLidarGroundFrame,
   fetchLidarGroundBenchmarks,
   fetchLidarReplayCatalog,
   fetchLidarReplayDetail,
+  type LidarFieldReview,
+  type LidarFieldReviewWindow,
   type LidarGroundBenchmark,
   type LidarGroundFrame,
   type LidarReplayCatalog,
@@ -69,6 +73,14 @@ export function LidarQualityWorkspace({
   const [groundFrameError, setGroundFrameError] = useState<string | null>(null);
   const [groundViewMode, setGroundViewMode] =
     useState<LidarGroundViewMode>("disagreement");
+  const [fieldReview, setFieldReview] = useState<LidarFieldReview | null>(null);
+  const [fieldWindow, setFieldWindow] =
+    useState<LidarFieldReviewWindow | null>(null);
+  const [fieldWindowIndex, setFieldWindowIndex] = useState(0);
+  const [fieldLoading, setFieldLoading] = useState(true);
+  const [fieldError, setFieldError] = useState<string | null>(null);
+  const [fieldViewMode, setFieldViewMode] =
+    useState<LidarGroundViewMode>("disagreement");
   const [selectedPackId, setSelectedPackId] = useState<string | null>(null);
   const [loading, setLoading] = useState(true);
   const [error, setError] = useState<string | null>(null);
@@ -146,7 +158,56 @@ export function LidarQualityWorkspace({
     return () => controller.abort();
   }, [groundBenchmark, groundFrameIndex]);
 
+  useEffect(() => {
+    const controller = new AbortController();
+    setFieldLoading(true);
+    setFieldError(null);
+    void fetchLidarFieldReviews({ signal: controller.signal })
+      .then((nextCatalog) => {
+        if (controller.signal.aborted) return;
+        const nextReview = nextCatalog.items[0] ?? null;
+        setFieldReview(nextReview);
+        setFieldWindow(null);
+        setFieldWindowIndex(nextReview?.selection.defaultWindowIndex ?? 0);
+      })
+      .catch((loadError) => {
+        if (controller.signal.aborted) return;
+        setFieldReview(null);
+        setFieldWindow(null);
+        setFieldError(errorMessage(loadError));
+      })
+      .finally(() => {
+        if (!controller.signal.aborted) setFieldLoading(false);
+      });
+    return () => controller.abort();
+  }, [reloadGeneration]);
+
+  useEffect(() => {
+    if (!fieldReview) return;
+    const controller = new AbortController();
+    setFieldLoading(true);
+    setFieldError(null);
+    void fetchLidarFieldReviewWindow(
+      fieldReview.reviewId,
+      fieldWindowIndex,
+      { signal: controller.signal },
+    )
+      .then((window) => {
+        if (!controller.signal.aborted) setFieldWindow(window);
+      })
+      .catch((loadError) => {
+        if (controller.signal.aborted) return;
+        setFieldWindow(null);
+        setFieldError(errorMessage(loadError));
+      })
+      .finally(() => {
+        if (!controller.signal.aborted) setFieldLoading(false);
+      });
+    return () => controller.abort();
+  }, [fieldReview, fieldWindowIndex]);
+
   const groundNormalization = groundBenchmark?.inputDomain.normalization ?? null;
+  const selectedFieldWindow = fieldReview?.windows[fieldWindowIndex] ?? null;
 
   return (
     <div className="standard-workspace lidar-quality-workspace">
@@ -191,12 +252,213 @@ export function LidarQualityWorkspace({
         </GlassSurface>
       ) : (
         <>
+          <GlassSurface className="lidar-field-review" padding="lg">
+            <header className="panel-heading lidar-field-review__heading">
+              <div>
+                <span className="section-eyebrow">ПОЛЕВОЙ REVIEW · RAVNOVES00</span>
+                <h2>
+                  {fieldReview?.displayName
+                    ?? "Центральный городской интервал"}
+                </h2>
+                <p>
+                  Дорога, дома, автомобили и растительность. Облако накоплено
+                  по исходным LiDAR-сэмплам выбранного окна, а кадр камеры
+                  фиксирует контекст сцены.
+                </p>
+              </div>
+              <StatusBadge tone={fieldError ? "danger" : "warning"}>
+                {fieldError ? "Review недоступен" : "Visual review only"}
+              </StatusBadge>
+            </header>
+
+            {fieldReview ? (
+              <>
+                <div
+                  className="lidar-field-review__source"
+                  aria-label="Источник полевого LiDAR review"
+                >
+                  <div>
+                    <span>Запись</span>
+                    <strong>{fieldReview.sessionId}</strong>
+                  </div>
+                  <div>
+                    <span>Интервал записи</span>
+                    <strong>
+                      {formatNumber(fieldReview.source.timelineStartSeconds, 2)}
+                      {"–"}
+                      {formatNumber(fieldReview.source.timelineEndSeconds, 2)} с
+                    </strong>
+                  </div>
+                  <div>
+                    <span>LiDAR-сэмплов</span>
+                    <strong>
+                      {fieldReview.source.availableLidarFrames.toLocaleString("ru-RU")}
+                    </strong>
+                  </div>
+                  <div>
+                    <span>Исходных точек</span>
+                    <strong>
+                      {fieldReview.source.pointCount.toLocaleString("ru-RU")}
+                    </strong>
+                  </div>
+                </div>
+
+                <div
+                  className="lidar-field-window-list"
+                  role="group"
+                  aria-label="Полевые сцены RAVNOVES00"
+                >
+                  {fieldReview.windows.map((window) => (
+                    <button
+                      type="button"
+                      key={window.key}
+                      data-selected={
+                        window.index === fieldWindowIndex ? "true" : undefined
+                      }
+                      onClick={() => {
+                        setFieldWindow(null);
+                        setFieldWindowIndex(window.index);
+                      }}
+                    >
+                      <span>{window.index + 1}</span>
+                      <div>
+                        <strong>{window.label}</strong>
+                        <small>
+                          {formatNumber(window.startSeconds, 0)}
+                          {"–"}
+                          {formatNumber(window.endSeconds, 0)} с ·{" "}
+                          {window.sourceLidarSamples} сканов
+                        </small>
+                      </div>
+                    </button>
+                  ))}
+                </div>
+
+                <section className="lidar-field-stage">
+                  <div className="lidar-field-camera">
+                    <div className="lidar-field-stage__label">
+                      <span>КОНТЕКСТ КАМЕРЫ</span>
+                      <strong>
+                        {selectedFieldWindow?.label ?? "Выбранная сцена"}
+                      </strong>
+                    </div>
+                    {fieldWindow ? (
+                      <img
+                        src={fieldWindow.previewUrl}
+                        alt={`Кадр камеры: ${fieldWindow.window.label}`}
+                      />
+                    ) : (
+                      <div className="lidar-field-stage__pending">
+                        {fieldError ?? "Загружаем кадр выбранной сцены…"}
+                      </div>
+                    )}
+                    <footer>
+                      <span>
+                        Кадр {selectedFieldWindow?.previewSourceFrameIndex ?? "—"}
+                      </span>
+                      <span>
+                        t = {formatNumber(
+                          selectedFieldWindow?.previewSessionSeconds ?? null,
+                          2,
+                        )}{" "}
+                        с
+                      </span>
+                    </footer>
+                  </div>
+
+                  <div className="lidar-field-cloud">
+                    <header>
+                      <div className="lidar-field-stage__label">
+                        <span>НАКОПЛЕННОЕ MAP-ОБЛАКО</span>
+                        <strong>
+                          {selectedFieldWindow
+                            ? `${selectedFieldWindow.sourcePointCount.toLocaleString(
+                                "ru-RU",
+                              )} исходных · ${selectedFieldWindow.displayPointCount.toLocaleString(
+                                "ru-RU",
+                              )} показано`
+                            : "Ожидание данных"}
+                        </strong>
+                      </div>
+                      <div
+                        className="lidar-ground-modes"
+                        role="group"
+                        aria-label="Режим окраски полевого LiDAR"
+                      >
+                        {([
+                          ["current", "Current"],
+                          ["candidate", "Patchwork++"],
+                          ["disagreement", "Расхождения"],
+                        ] as const).map(([mode, label]) => (
+                          <button
+                            type="button"
+                            key={mode}
+                            data-active={
+                              fieldViewMode === mode ? "true" : undefined
+                            }
+                            onClick={() => setFieldViewMode(mode)}
+                          >
+                            {label}
+                          </button>
+                        ))}
+                      </div>
+                    </header>
+                    {fieldWindow ? (
+                      <LidarGroundPointCloud
+                        frame={fieldWindow}
+                        mode={fieldViewMode}
+                      />
+                    ) : (
+                      <div className="lidar-ground-scene-placeholder">
+                        <StatusBadge tone={fieldError ? "danger" : "accent"}>
+                          {fieldError ? "Ошибка" : "Загрузка"}
+                        </StatusBadge>
+                        <span>
+                          {fieldError ?? "Готовим накопленное облако…"}
+                        </span>
+                      </div>
+                    )}
+                  </div>
+                </section>
+
+                <div className="lidar-field-review__explanation">
+                  <StatusBadge tone="warning">Не accuracy</StatusBadge>
+                  <p>
+                    Маски вычислены отдельно на каждом исходном скане и только
+                    затем сведены в map frame. Это legacy E10 vendor-map
+                    derivative без intensity и независимой ground-разметки,
+                    поэтому результат предназначен для визуального разбора, а
+                    не для production-gate.
+                  </p>
+                  <div className="lidar-ground-legend">
+                    <span><i data-color="shared" />Ground у обоих</span>
+                    <span><i data-color="current" />Только current</span>
+                    <span><i data-color="candidate" />Только Patchwork++</span>
+                    <span><i data-color="non-ground" />Оба non-ground</span>
+                  </div>
+                </div>
+              </>
+            ) : (
+              <div className="lidar-field-review__empty">
+                <StatusBadge tone={fieldError ? "danger" : "accent"}>
+                  {fieldError ? "Ошибка загрузки" : "Проверка evidence"}
+                </StatusBadge>
+                <p>
+                  {fieldError
+                    ?? (fieldLoading
+                      ? "Читаем RAVNOVES00 field-review артефакт…"
+                      : "Field-review артефакт ещё не опубликован.")}
+                </p>
+              </div>
+            )}
+          </GlassSurface>
+
           <section className="metrics-grid" aria-label="Качество LiDAR replay">
             <MetricCard
               featured
-              eyebrow="ТОЧЕЧНЫХ КАДРОВ"
+              eyebrow="ТЕХНИЧЕСКИХ КАДРОВ"
               value={detail.pack.pointFrames.toLocaleString("ru-RU")}
-              detail={`${detail.pack.points.toLocaleString("ru-RU")} точек сохранено`}
+              detail={`${detail.pack.points.toLocaleString("ru-RU")} точек · indoor contract slice`}
             />
             <MetricCard
               eyebrow="СРЕДНЕЕ ТОЧЕК"
@@ -358,12 +620,13 @@ export function LidarQualityWorkspace({
                   <header>
                     <div>
                       <span className="section-eyebrow">
-                        POINT-ALIGNED REVIEW
+                        ТЕХНИЧЕСКИЙ CONTRACT SLICE · VIEWER_LIVE
                       </span>
-                      <h3>Покадровое облако и маски</h3>
+                      <h3>Indoor-проверка point-aligned контракта</h3>
                       <p>
-                        Один и тот же map-frame XYZ, разные диагностические
-                        раскраски. Маски не изменяют replay.
+                        Этот короткий indoor-срез подтверждает выравнивание XYZ
+                        и масок, но не является полевой оценкой качества. Для
+                        улицы используйте RAVNOVES00 выше.
                       </p>
                     </div>
                     <div className="lidar-ground-frame-status">
diff --git a/apps/control-station/test/lidarReplayQuality.test.mjs b/apps/control-station/test/lidarReplayQuality.test.mjs
index 171219a..01a8ded 100644
--- a/apps/control-station/test/lidarReplayQuality.test.mjs
+++ b/apps/control-station/test/lidarReplayQuality.test.mjs
@@ -8,10 +8,14 @@ let parseLidarReplayCatalog;
 let parseLidarReplayDetail;
 let parseLidarGroundBenchmarkCatalog;
 let parseLidarGroundFrame;
+let parseLidarFieldReviewCatalog;
+let parseLidarFieldReviewWindow;
 let fetchLidarReplayCatalog;
 let fetchLidarReplayDetail;
 let fetchLidarGroundBenchmarks;
 let fetchLidarGroundFrame;
+let fetchLidarFieldReviews;
+let fetchLidarFieldReviewWindow;
 let LidarReplayContractError;
 let workspaceById;
 
@@ -26,10 +30,14 @@ before(async () => {
     parseLidarReplayDetail,
     parseLidarGroundBenchmarkCatalog,
     parseLidarGroundFrame,
+    parseLidarFieldReviewCatalog,
+    parseLidarFieldReviewWindow,
     fetchLidarReplayCatalog,
     fetchLidarReplayDetail,
     fetchLidarGroundBenchmarks,
     fetchLidarGroundFrame,
+    fetchLidarFieldReviews,
+    fetchLidarFieldReviewWindow,
     LidarReplayContractError,
   } = await server.ssrLoadModule("/src/core/lidar/replayQuality.ts"));
   ({ workspaceById } = await server.ssrLoadModule("/src/productModel.ts"));
@@ -40,6 +48,8 @@ after(async () => {
 });
 
 const packId = `lidar-replay-pack-${"a".repeat(64)}`;
+const fieldReviewId = `lidar-field-review-${"d".repeat(64)}`;
+const e10PackId = `e10-lidar-pack-${"e".repeat(64)}`;
 
 function summary(overrides = {}) {
   return {
@@ -267,6 +277,145 @@ function groundFrame(overrides = {}) {
   };
 }
 
+function fieldReviewWindowSummary(overrides = {}) {
+  return {
+    index: 0,
+    key: "intersection-facades",
+    label: "Перекрёсток, дорога и фасады",
+    start_seconds: 145,
+    end_seconds: 151,
+    midpoint_seconds: 148,
+    source_lidar_samples: 56,
+    source_point_count: 98538,
+    display_point_count: 3,
+    source_frame_start: 1097,
+    source_frame_end: 1156,
+    preview_source_frame_index: 1120,
+    preview_session_seconds: 147.451857292,
+    ...overrides,
+  };
+}
+
+function fieldReviewCatalog(overrides = {}) {
+  const provider = (providerId, sourceCommit = undefined) => ({
+    provider_id: providerId,
+    source_commit: sourceCommit,
+  });
+  const branch = (providerValue) => ({
+    provider: providerValue,
+    ground_fraction: distribution(),
+    latency_ms: distribution(),
+  });
+  return {
+    schema_version: "missioncore.lidar-field-review-catalog/v1",
+    configured: true,
+    valid_total: 1,
+    invalid_total: 0,
+    access: "read-only",
+    items: [{
+      review_id: fieldReviewId,
+      display_name: "RAVNOVES00 · центральный городской интервал",
+      session_id: "20260720T065719Z_viewer_live",
+      source_pack_id: e10PackId,
+      status: "diagnostic-only",
+      source: {
+        timeline_start_seconds: 135.365857292,
+        timeline_end_seconds: 195.334857292,
+        available_lidar_frames: 526,
+        point_count: 1182292,
+        representation: "legacy-e10-vendor-map-with-pose",
+        intensity_available: false,
+        raw_scan_accepted: false,
+      },
+      selection: {
+        purpose: "operator-readable-central-urban-field-review",
+        default_window_index: 0,
+        accumulation: "per-source-frame masks accumulated in map frame",
+        sampling: {
+          method: "uniform-point-index-per-window",
+          maximum_points_per_window: 80000,
+        },
+      },
+      windows: [fieldReviewWindowSummary()],
+      metrics: {
+        source_samples: 56,
+        current: branch(
+          provider("missioncore-local-percentile-ground/v1"),
+        ),
+        candidate: branch(
+          provider(
+            "patchworkpp/v1.4.1",
+            "3e6903a1d5537a4cc2ace897b0bbb98a92d6014c",
+          ),
+        ),
+        comparison: {
+          algorithm_to_algorithm_ground_iou: distribution(),
+          ground_disagreement_fraction: distribution(),
+          is_accuracy_metric: false,
+        },
+      },
+      decision: {
+        status: "visual-review-only",
+        production_promotion: false,
+        reasons: ["legacy derivative does not retain intensity"],
+      },
+      created_at_utc: "2026-07-25T02:00:00Z",
+      ground_truth: false,
+      authority: {
+        commands_enabled: false,
+        navigation_or_safety_accepted: false,
+      },
+    }],
+    ...overrides,
+  };
+}
+
+function fieldReviewWindow(overrides = {}) {
+  return {
+    schema_version: "missioncore.lidar-field-review-window/v1",
+    review_id: fieldReviewId,
+    display_name: "RAVNOVES00 · центральный городской интервал",
+    session_id: "20260720T065719Z_viewer_live",
+    source_pack_id: e10PackId,
+    window_index: 0,
+    window_count: 1,
+    window: fieldReviewWindowSummary(),
+    point_count: 3,
+    coordinate_frame: "map",
+    distance_unit: "m",
+    intensity: {
+      available: false,
+      reason: "E10 derivative did not retain rgbi/intensity",
+    },
+    points_xyz_m: [
+      [0, 0, 0],
+      [1, 0, 0.1],
+      [0, 1, 0.5],
+    ],
+    masks: {
+      current_ground: [1, 1, 0],
+      current_assigned: [1, 1, 1],
+      candidate_ground: [1, 0, 0],
+      candidate_assigned: [1, 1, 1],
+      disagreement: [0, 1, 0],
+    },
+    counts: {
+      current_ground: 2,
+      candidate_ground: 1,
+      disagreement: 1,
+    },
+    preview_url:
+      `/api/v1/lidar/field-reviews/${fieldReviewId}/windows/0/preview`,
+    access: "read-only",
+    ground_truth: false,
+    authority: {
+      commands_enabled: false,
+      navigation_or_safety_accepted: false,
+    },
+    ...overrides,
+  };
+}
+
 function jsonResponse(payload, status = 200) {
   return new Response(JSON.stringify(payload), {
     status,
@@ -359,6 +508,49 @@ test("ground frame stays point-aligned, bounded and path-free", () => {
   );
 });
 
+test("field review identifies RAVNOVES00 source and stays visual-only", () => {
+  const parsedCatalog = parseLidarFieldReviewCatalog(fieldReviewCatalog());
+  const parsedWindow = parseLidarFieldReviewWindow(fieldReviewWindow());
+
+  assert.equal(parsedCatalog.items[0].sourcePackId, e10PackId);
+  assert.equal(parsedCatalog.items[0].source.availableLidarFrames, 526);
+  assert.equal(parsedCatalog.items[0].selection.defaultWindowIndex, 0);
+  assert.equal(parsedWindow.window.label, "Перекрёсток, дорога и фасады");
+  assert.equal(parsedWindow.intensity0To255, null);
+  assert.equal(parsedWindow.counts.disagreement, 1);
+  assert.equal("path" in parsedWindow, false);
+
+  const promoted = fieldReviewCatalog();
+  promoted.items[0].decision.production_promotion = true;
+  assert.throws(
+    () => parseLidarFieldReviewCatalog(promoted),
+    LidarReplayContractError,
+  );
+});
+
+test("field-review window refuses forged preview and point masks", () => {
+  assert.throws(
+    () => parseLidarFieldReviewWindow(fieldReviewWindow({
+      preview_url: "https://example.invalid/frame.jpg",
+    })),
+    LidarReplayContractError,
+  );
+  assert.throws(
+    () => parseLidarFieldReviewWindow(fieldReviewWindow({
+      masks: {
+        ...fieldReviewWindow().masks,
+        disagreement: [0, 0, 0],
+      },
+      counts: {
+        current_ground: 2,
+        candidate_ground: 1,
+        disagreement: 0,
+      },
+    })),
+    LidarReplayContractError,
+  );
+});
+
 test("LiDAR fetchers use read-only endpoints and workspace is registered", async () => {
   const calls = [];
   const fetcher = async (input, init) => {
@@ -368,6 +560,11 @@ test("LiDAR fetchers use read-only endpoints and workspace is registered", async
         ? jsonResponse(groundFrame())
         : jsonResponse(groundCatalog());
     }
+    if (String(input).includes("field-reviews")) {
+      return String(input).includes("/windows/")
+        ? jsonResponse(fieldReviewWindow())
+        : jsonResponse(fieldReviewCatalog());
+    }
     return String(input).includes(packId)
       ? jsonResponse(detail())
       : jsonResponse(catalog());
@@ -380,11 +577,19 @@ test("LiDAR fetchers use read-only endpoints and workspace is registered", async
     0,
     { fetcher },
   );
+  const fieldCatalog = await fetchLidarFieldReviews({ fetcher });
+  const fieldWindow = await fetchLidarFieldReviewWindow(
+    fieldReviewId,
+    0,
+    { fetcher },
+  );
 
   assert.equal(parsedCatalog.validTotal, 1);
   assert.equal(parsedDetail.pack.packId, packId);
   assert.equal(ground.validTotal, 1);
   assert.equal(frame.pointCount, 3);
+  assert.equal(fieldCatalog.items[0].windows[0].sourceLidarSamples, 56);
+  assert.equal(fieldWindow.previewUrl.endsWith("/preview"), true);
   assert.deepEqual(calls, [
     { input: "/api/v1/lidar/replay-packs?limit=50", method: "GET" },
     { input: `/api/v1/lidar/replay-packs/${packId}`, method: "GET" },
@@ -396,6 +601,14 @@ test("LiDAR fetchers use read-only endpoints and workspace is registered", async
       input: `/api/v1/lidar/ground-benchmarks/ground-benchmark-${"c".repeat(64)}/frames/0`,
       method: "GET",
     },
+    {
+      input: "/api/v1/lidar/field-reviews?limit=10",
+      method: "GET",
+    },
+    {
+      input: `/api/v1/lidar/field-reviews/${fieldReviewId}/windows/0`,
+      method: "GET",
+    },
   ]);
   assert.equal(workspaceById("lidar-quality").root, "data");
   assert.equal(workspaceById("lidar-quality").kind, "lidar-quality");
diff --git a/docs/10_EXTERNAL_PERCEPTION_WORKER.md b/docs/10_EXTERNAL_PERCEPTION_WORKER.md
index ea70fcc..f6e8d91 100644
--- a/docs/10_EXTERNAL_PERCEPTION_WORKER.md
+++ b/docs/10_EXTERNAL_PERCEPTION_WORKER.md
@@ -31,6 +31,17 @@ that evidence in Three.js with unrestricted orbit, pan and zoom plus
 intensity/current/candidate/disagreement color modes. The browser never runs
 Patchwork++, parses private firmware or receives a filesystem path.
 
+`missioncore.lidar-field-review/v1` is the separate operator-readable field
+surface. The read-only `/api/v1/lidar/field-reviews` catalog names the exact
+RAVNOVES00 source, timeline and five central urban windows. The bounded
+`/api/v1/lidar/field-reviews/{review_id}/windows/{window_index}` response
+contains a deterministic accumulated map cloud plus masks that were computed
+per original LiDAR sample before accumulation. Its sibling `/preview` endpoint
+returns the content-bound camera JPEG for that window. No source path,
+Patchwork++ binary, worker control or command authority crosses this boundary.
+The legacy E10 source has no intensity, so the API declares it unavailable
+instead of synthesizing a value.
+
 Static K1 3.0.2 firmware evidence confirms that the appliance internally uses
 a Livox MID-360 point/IMU path with richer timestamp/ring semantics and a
 configured MQTT point-cloud downsample factor of four. This creates a concrete
diff --git a/docs/13_LIDAR_WORKER_PRODUCT_AND_ROADMAP.md b/docs/13_LIDAR_WORKER_PRODUCT_AND_ROADMAP.md
index 578da76..b11b829 100644
--- a/docs/13_LIDAR_WORKER_PRODUCT_AND_ROADMAP.md
+++ b/docs/13_LIDAR_WORKER_PRODUCT_AND_ROADMAP.md
@@ -226,6 +226,8 @@ quality line. It is not repaired or hidden by replay.
   source point; raw replay remains unchanged.
 - [x] Add a bounded point-aligned frame API and browser 3D review for
   intensity, current ground, candidate ground and disagreement.
+- [x] Separate the 66-frame indoor contract slice from a named RAVNOVES00
+  field review with five camera-bound urban windows and accumulated map clouds.
 - [x] Record the K1 3.0.2 internal MID-360/LIO field path from static,
   redacted firmware evidence without changing device state.
 - [x] Bind physical height, applied vertical-origin offset and evidence class
@@ -255,6 +257,26 @@ latency remains 0.42 ms. Algorithm IoU is 4.07% p50 and disagreement is 19.13%
 p50. The run is useful for visual review, but its evidence class is
 `operator-estimated`, so input acceptance and production promotion stay false.
 
+The operator-facing field generation is
+`lidar-field-review-57f359dae336f06962e3a29e69e2da1bb8365f9af46d5d3173609f92d43db1ef`.
+It uses the immutable RAVNOVES00 E10 derivative
+`e10-lidar-pack-5da0396d32a27f9d1ca537cc2e8a371d386078d6f0dc71737b78620992af9625`
+over `135.365857292–195.334857292` session seconds. The source has 526
+available LiDAR samples and 1,182,292 points. Five central urban windows cover
+roads, facades, sidewalks, parked vehicles and vegetation. Each window:
+
+- retains the exact source interval and camera preview identity;
+- computes both ground masks independently on every original LiDAR sample;
+- accumulates those point-aligned results only afterwards in the map frame;
+- reports 53–68 source samples and 98,538–172,027 source points;
+- publishes a deterministic uniform display sample of at most 80,000 points.
+
+This field review fixes the operator-context problem of the indoor technical
+slice; it does not upgrade evidence. The E10 derivative remains intensity-free,
+vendor-mapped and without independent ground labels. Its contract is therefore
+`missioncore.lidar-field-review/v1`, `status=diagnostic-only`,
+`decision.status=visual-review-only` and `production_promotion=false`.
+
 The result is an evidence-backed **do-not-promote** decision for the current K1
 feed. Patchwork++ is fast, but its input model assumes a sensor-centric scan
 and physical sensor height. K1 `lio_pcl` is a vendor-mapped increment, its
diff --git a/experiments/perception/run_lidar_field_review.py b/experiments/perception/run_lidar_field_review.py
new file mode 100644
index 0000000..24c92e7
--- /dev/null
+++ b/experiments/perception/run_lidar_field_review.py
@@ -0,0 +1,182 @@
+#!/usr/bin/env python3
+from __future__ import annotations
+
+import argparse
+import hashlib
+import json
+import subprocess
+import tempfile
+from pathlib import Path
+
+from k1link.compute import (
+    PATCHWORKPP_SOURCE_COMMIT,
+    PATCHWORKPP_SOURCE_TAG,
+    RAVNOVES00_CENTRAL_WINDOWS,
+    E10LidarFieldSource,
+    GroundBenchmarkProfile,
+    LidarFieldReviewV1,
+    PatchworkPPGroundSegmenter,
+    build_lidar_field_review,
+)
+
+
+def _arguments() -> argparse.Namespace:
+    parser = argparse.ArgumentParser(
+        description=(
+            "Build accumulated RAVNOVES00 central-urban LiDAR ground review "
+            "with content-bound camera context previews."
+        )
+    )
+    parser.add_argument("source_pack", type=Path)
+    parser.add_argument("camera_epoch", type=Path)
+    parser.add_argument("output_root", type=Path)
+    parser.add_argument("--ffmpeg", default="ffmpeg")
+    parser.add_argument("--patchwork-module", default="pypatchworkpp")
+    parser.add_argument("--patchwork-source-tag", default=PATCHWORKPP_SOURCE_TAG)
+    parser.add_argument(
+        "--patchwork-source-commit",
+        default=PATCHWORKPP_SOURCE_COMMIT,
+    )
+    parser.add_argument("--sensor-height-m", type=float, default=1.27)
+    parser.add_argument("--map-vertical-origin-offset-m", type=float, default=1.27)
+    return parser.parse_args()
+
+
+def _preview_segments(epoch: Path) -> dict[str, Path]:
+    resolved = epoch.expanduser().resolve(strict=True)
+    init = resolved / "init.mp4"
+    index_path = resolved / "index.jsonl"
+    if not init.is_file() or init.is_symlink() or not index_path.is_file():
+        raise RuntimeError("Camera epoch is incomplete")
+    required = {
+        window.preview_source_frame_index + 1: window.key for window in RAVNOVES00_CENTRAL_WINDOWS
+    }
+    found: dict[str, Path] = {}
+    with index_path.open(encoding="utf-8") as stream:
+        for line in stream:
+            value = json.loads(line)
+            sequence = value.get("sequence") if isinstance(value, dict) else None
+            if sequence not in required:
+                continue
+            relative = value.get("path")
+            expected_sha256 = value.get("sha256")
+            expected_length = value.get("length")
+            if (
+                value.get("schema_version") != "missioncore.camera-recording-index/v1"
+                or relative != f"segments/{sequence}.m4s"
+                or not isinstance(expected_sha256, str)
+                or not isinstance(expected_length, int)
+            ):
+                raise RuntimeError("Camera preview segment index is invalid")
+            segment = resolved / relative
+            if (
+                segment.is_symlink()
+                or not segment.is_file()
+                or segment.stat().st_size != expected_length
+                or _sha256(segment) != expected_sha256
+            ):
+                raise RuntimeError("Camera preview segment failed integrity")
+            found[required[sequence]] = segment
+    if set(found) != set(required.values()):
+        raise RuntimeError("Camera preview segments are incomplete")
+    return found
+
+
+def _extract_previews(
+    ffmpeg: str,
+    epoch: Path,
+    output: Path,
+) -> dict[str, Path]:
+    init = epoch.expanduser().resolve(strict=True) / "init.mp4"
+    segments = _preview_segments(epoch)
+    previews: dict[str, Path] = {}
+    for key, segment in segments.items():
+        target = output / f"{key}.jpg"
+        subprocess.run(
+            [
+                ffmpeg,
+                "-hide_banner",
+                "-loglevel",
+                "error",
+                "-i",
+                f"concat:{init}|{segment}",
+                "-frames:v",
+                "1",
+                "-vf",
+                "scale=640:480",
+                "-q:v",
+                "3",
+                str(target),
+            ],
+            check=True,
+        )
+        if not target.is_file() or target.stat().st_size < 1_000:
+            raise RuntimeError("Camera preview extraction failed")
+        previews[key] = target
+    return previews
+
+
+def main() -> int:
+    arguments = _arguments()
+    profile = GroundBenchmarkProfile(
+        profile_id="ravnoves00-central-urban-operator-height-ground-review/v1",
+        patchwork_sensor_height_proxy_m=arguments.sensor_height_m,
+        patchwork_map_vertical_origin_offset_m=(arguments.map_vertical_origin_offset_m),
+        patchwork_height_evidence="operator-estimated",
+    )
+    source = E10LidarFieldSource(arguments.source_pack)
+    try:
+        patchwork = PatchworkPPGroundSegmenter.load(
+            profile=profile,
+            module_name=arguments.patchwork_module,
+            source_tag=arguments.patchwork_source_tag,
+            source_commit=arguments.patchwork_source_commit,
+        )
+        with tempfile.TemporaryDirectory(prefix="missioncore-lidar-field-review-") as value:
+            previews = _extract_previews(
+                arguments.ffmpeg,
+                arguments.camera_epoch,
+                Path(value),
+            )
+            output = build_lidar_field_review(
+                source,
+                arguments.output_root,
+                patchwork=patchwork,
+                profile=profile,
+                preview_paths=previews,
+            )
+    finally:
+        source.close()
+    review = LidarFieldReviewV1(output)
+    try:
+        print(
+            json.dumps(
+                {
+                    "review_id": review.review_id,
+                    "display_name": review.report["display_name"],
+                    "session_id": review.report["session_id"],
+                    "source": review.report["source"],
+                    "selection": review.report["selection"],
+                    "windows": review.report["windows"],
+                    "metrics": review.report["metrics"],
+                    "decision": review.report["decision"],
+                },
+                ensure_ascii=False,
+                indent=2,
+            )
+        )
+    finally:
+        review.close()
+    return 0
+
+
+def _sha256(path: Path) -> str:
+    digest = hashlib.sha256()
+    with path.open("rb") as stream:
+        while chunk := stream.read(1024 * 1024):
+            digest.update(chunk)
+    return digest.hexdigest()
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())
diff --git a/src/k1link/compute/__init__.py b/src/k1link/compute/__init__.py
index 4430145..4a74317 100644
--- a/src/k1link/compute/__init__.py
+++ b/src/k1link/compute/__init__.py
@@ -69,6 +69,22 @@ from .lidar_contract import (
     lidar_readiness_document,
     sensor_frame_xyzi,
 )
+from .lidar_field_review import (
+    E10_LIDAR_PACK_SCHEMA,
+    FIELD_REVIEW_ARRAYS_NAME,
+    FIELD_REVIEW_MANIFEST_NAME,
+    FIELD_REVIEW_REPORT_NAME,
+    LIDAR_FIELD_REVIEW_REPORT_SCHEMA,
+    LIDAR_FIELD_REVIEW_SCHEMA,
+    LIDAR_FIELD_REVIEW_WINDOW_SCHEMA,
+    MAX_FIELD_REVIEW_WINDOW_POINTS,
+    RAVNOVES00_CENTRAL_WINDOWS,
+    E10LidarFieldSource,
+    FieldReviewWindowSpec,
+    LidarFieldReviewV1,
+    build_lidar_field_review,
+    lidar_field_review_catalog_item,
+)
 from .lidar_ground import (
     DEFAULT_GROUND_BENCHMARK_PROFILE,
     LIDAR_GROUND_ANNOTATION_TEMPLATE_SCHEMA,
@@ -179,6 +195,9 @@ __all__ = [
     "LIDAR_GROUND_BENCHMARK_REPORT_SCHEMA",
     "LIDAR_GROUND_BENCHMARK_SCHEMA",
     "LIDAR_GROUND_FRAME_SCHEMA",
+    "LIDAR_FIELD_REVIEW_REPORT_SCHEMA",
+    "LIDAR_FIELD_REVIEW_SCHEMA",
+    "LIDAR_FIELD_REVIEW_WINDOW_SCHEMA",
     "LIDAR_EVIDENCE_PROFILE_SCHEMA",
     "LIDAR_EQUIVALENCE_REPORT_SCHEMA",
     "LIDAR_QUALITY_REPORT_SCHEMA",
@@ -262,6 +281,17 @@ __all__ = [
     "build_lidar_ground_annotation_template",
     "build_lidar_ground_benchmark",
     "lidar_ground_frame_detail",
+    "E10_LIDAR_PACK_SCHEMA",
+    "FIELD_REVIEW_ARRAYS_NAME",
+    "FIELD_REVIEW_MANIFEST_NAME",
+    "FIELD_REVIEW_REPORT_NAME",
+    "MAX_FIELD_REVIEW_WINDOW_POINTS",
+    "RAVNOVES00_CENTRAL_WINDOWS",
+    "E10LidarFieldSource",
+    "FieldReviewWindowSpec",
+    "LidarFieldReviewV1",
+    "build_lidar_field_review",
+    "lidar_field_review_catalog_item",
     "DetectionFrame",
     "ObjectDetection",
     "RecordedPerceptionOverlayError",
diff --git a/src/k1link/compute/lidar_field_review.py b/src/k1link/compute/lidar_field_review.py
new file mode 100644
index 0000000..76698c0
--- /dev/null
+++ b/src/k1link/compute/lidar_field_review.py
@@ -0,0 +1,891 @@
+from __future__ import annotations
+
+import hashlib
+import json
+import os
+import re
+import shutil
+from collections.abc import Mapping, Sequence
+from dataclasses import dataclass
+from datetime import UTC, datetime
+from pathlib import Path
+from typing import Any, Final
+
+import numpy as np
+import numpy.typing as npt
+
+from k1link.device_plugins.xgrids_k1.analyze import (
+    CalibratedProjectionError,
+    map_points_to_lidar,
+)
+
+from .lidar_ground import (
+    GroundBenchmarkProfile,
+    GroundSegmenter,
+    LidarGroundError,
+    LocalPercentileGroundSegmenter,
+)
+
+LIDAR_FIELD_REVIEW_SCHEMA: Final = "missioncore.lidar-field-review/v1"
+LIDAR_FIELD_REVIEW_REPORT_SCHEMA: Final = "missioncore.lidar-field-review-report/v1"
+LIDAR_FIELD_REVIEW_WINDOW_SCHEMA: Final = "missioncore.lidar-field-review-window/v1"
+E10_LIDAR_PACK_SCHEMA: Final = "missioncore.e10-lidar-replay-pack/v1"
+FIELD_REVIEW_ARRAYS_NAME: Final = "field-review.npz"
+FIELD_REVIEW_REPORT_NAME: Final = "field-review.json"
+FIELD_REVIEW_MANIFEST_NAME: Final = "manifest.json"
+MAX_FIELD_REVIEW_WINDOW_POINTS: Final = 80_000
+
+_E10_PACK_ID = re.compile(r"^e10-lidar-pack-[a-f0-9]{64}$")
+_FIELD_REVIEW_ID = re.compile(r"^lidar-field-review-[a-f0-9]{64}$")
+_SHA256 = re.compile(r"^[a-f0-9]{64}$")
+_SAFE_KEY = re.compile(r"^[a-z0-9][a-z0-9-]{0,63}$")
+
+
+@dataclass(frozen=True, slots=True)
+class FieldReviewWindowSpec:
+    key: str
+    label: str
+    start_seconds: float
+    end_seconds: float
+    preview_source_frame_index: int
+
+    def __post_init__(self) -> None:
+        if (
+            _SAFE_KEY.fullmatch(self.key) is None
+            or not self.label.strip()
+            or len(self.label) > 120
+            or not np.isfinite((self.start_seconds, self.end_seconds)).all()
+            or not 0 <= self.start_seconds < self.end_seconds
+            or self.preview_source_frame_index < 0
+        ):
+            raise LidarGroundError("LiDAR field-review window is invalid")
+
+    def to_dict(self) -> dict[str, object]:
+        return {
+            "key": self.key,
+            "label": self.label,
+            "start_seconds": self.start_seconds,
+            "end_seconds": self.end_seconds,
+            "preview_source_frame_index": self.preview_source_frame_index,
+        }
+
+
+RAVNOVES00_CENTRAL_WINDOWS: Final = (
+    FieldReviewWindowSpec(
+        key="intersection-facades",
+        label="Перекрёсток, дорога и фасады",
+        start_seconds=145.0,
+        end_seconds=151.0,
+        preview_source_frame_index=1120,
+    ),
+    FieldReviewWindowSpec(
+        key="crossing-parked-vehicles",
+        label="Переход и припаркованные машины",
+        start_seconds=153.0,
+        end_seconds=159.0,
+        preview_source_frame_index=1200,
+    ),
+    FieldReviewWindowSpec(
+        key="long-street",
+        label="Длинный фасад, тротуар и улица",
+        start_seconds=167.0,
+        end_seconds=175.0,
+        preview_source_frame_index=1350,
+    ),
+    FieldReviewWindowSpec(
+        key="sidewalk-vehicles",
+        label="Тротуар, дома и автомобили",
+        start_seconds=180.0,
+        end_seconds=187.0,
+        preview_source_frame_index=1480,
+    ),
+    FieldReviewWindowSpec(
+        key="street-vegetation",
+        label="Продолжение улицы и растительность",
+        start_seconds=188.0,
+        end_seconds=194.0,
+        preview_source_frame_index=1550,
+    ),
+)
+
+
+class E10LidarFieldSource:
+    """Strict reader for the immutable, intensity-free RAVNOVES00 E10 pack."""
+
+    def __init__(self, root: Path) -> None:
+        candidate = root.expanduser().absolute()
+        if candidate.is_symlink():
+            raise LidarGroundError("E10 LiDAR source cannot be a symlink")
+        self.root = candidate.resolve(strict=True)
+        if not self.root.is_dir() or _E10_PACK_ID.fullmatch(self.root.name) is None:
+            raise LidarGroundError("E10 LiDAR source id is invalid")
+        self.manifest = _read_json(self.root / "manifest.json")
+        self.identity = _object(self.manifest.get("identity"), "E10 LiDAR identity")
+        identity_sha256 = self.manifest.get("identity_sha256")
+        artifact = _object(self.manifest.get("artifact"), "E10 LiDAR artifact")
+        artifact_path = artifact.get("path")
+        if (
+            self.manifest.get("schema_version") != E10_LIDAR_PACK_SCHEMA
+            or self.identity.get("schema_version") != E10_LIDAR_PACK_SCHEMA
+            or not isinstance(identity_sha256, str)
+            or _SHA256.fullmatch(identity_sha256) is None
+            or hashlib.sha256(_canonical_json(self.identity)).hexdigest() != identity_sha256
+            or self.root.name != f"e10-lidar-pack-{identity_sha256}"
+            or self.manifest.get("pack_id") != self.root.name
+            or artifact_path != "lidar-pack.npz"
+        ):
+            raise LidarGroundError("E10 LiDAR source identity is invalid")
+        arrays_path = self.root / artifact_path
+        if (
+            arrays_path.is_symlink()
+            or not arrays_path.is_file()
+            or arrays_path.stat().st_size != artifact.get("byte_length")
+            or _sha256(arrays_path) != artifact.get("sha256")
+        ):
+            raise LidarGroundError("E10 LiDAR source artifact is invalid")
+        self.arrays = np.load(arrays_path, allow_pickle=False)
+        try:
+            self._validate_arrays()
+        except BaseException:
+            self.close()
+            raise
+        self.pack_id = self.root.name
+
+    @property
+    def frame_count(self) -> int:
+        return int(self.identity["frame_count"])
+
+    @property
+    def point_count(self) -> int:
+        return int(self.identity["point_count"])
+
+    def close(self) -> None:
+        self.arrays.close()
+
+    def _validate_arrays(self) -> None:
+        required = {
+            "frame_indices",
+            "source_frame_indices",
+            "session_seconds",
+            "sample_available",
+            "cloud_offsets",
+            "cloud_points_map",
+            "pose_positions_map",
+            "pose_quaternions_map_from_lidar",
+            "lidar_camera_delta_ms",
+            "pose_point_delta_ms",
+            "intrinsic_fx_fy_cx_cy",
+            "distortion_kb4",
+            "t_camera_from_lidar",
+        }
+        frame_count = _nonnegative_int(self.identity.get("frame_count"), "E10 frame count")
+        point_count = _nonnegative_int(self.identity.get("point_count"), "E10 point count")
+        available = self.arrays["sample_available"]
+        offsets = self.arrays["cloud_offsets"]
+        points = self.arrays["cloud_points_map"]
+        positions = self.arrays["pose_positions_map"]
+        quaternions = self.arrays["pose_quaternions_map_from_lidar"]
+        if (
+            set(self.arrays.files) != required
+            or self.arrays["frame_indices"].shape != (frame_count,)
+            or self.arrays["source_frame_indices"].shape != (frame_count,)
+            or self.arrays["session_seconds"].shape != (frame_count,)
+            or available.shape != (frame_count,)
+            or offsets.shape != (frame_count + 1,)
+            or points.shape != (point_count, 3)
+            or positions.shape != (frame_count, 3)
+            or quaternions.shape != (frame_count, 4)
+            or self.arrays["frame_indices"].dtype != np.dtype("<i8")
+            or self.arrays["source_frame_indices"].dtype != np.dtype("<i8")
+            or self.arrays["session_seconds"].dtype != np.dtype("<f8")
+            or available.dtype != np.dtype("?")
+            or offsets.dtype != np.dtype("<i8")
+            or points.dtype != np.dtype("<f4")
+            or positions.dtype != np.dtype("<f8")
+            or quaternions.dtype != np.dtype("<f8")
+            or not np.array_equal(
+                self.arrays["frame_indices"],
+                np.arange(frame_count, dtype=np.int64),
+            )
+            or not np.all(np.diff(self.arrays["source_frame_indices"]) > 0)
+            or not np.isfinite(self.arrays["session_seconds"]).all()
+            or not np.all(np.diff(self.arrays["session_seconds"]) > 0)
+            or offsets[0] != 0
+            or offsets[-1] != point_count
+            or np.any(np.diff(offsets) < 0)
+            or not np.isfinite(points).all()
+            or int(np.count_nonzero(available)) != self.identity.get("available_lidar_frames")
+        ):
+            raise LidarGroundError("E10 LiDAR source arrays are invalid")
+        counts = np.diff(offsets)
+        if (
+            np.any(counts[available] <= 0)
+            or np.any(counts[~available] != 0)
+            or not np.isfinite(positions[available]).all()
+            or not np.isfinite(quaternions[available]).all()
+        ):
+            raise LidarGroundError("E10 LiDAR source availability is invalid")
+
+
+class LidarFieldReviewV1:
+    """Strict reader for accumulated, path-free LiDAR field-review evidence."""
+
+    def __init__(self, root: Path) -> None:
+        candidate = root.expanduser().absolute()
+        if candidate.is_symlink():
+            raise LidarGroundError("LiDAR field review cannot be a symlink")
+        self.root = candidate.resolve(strict=True)
+        if not self.root.is_dir() or _FIELD_REVIEW_ID.fullmatch(self.root.name) is None:
+            raise LidarGroundError("LiDAR field-review id is invalid")
+        self.manifest = _read_json(self.root / FIELD_REVIEW_MANIFEST_NAME)
+        self.identity = _object(self.manifest.get("identity"), "field-review identity")
+        identity_sha256 = self.manifest.get("identity_sha256")
+        if (
+            self.manifest.get("schema_version") != LIDAR_FIELD_REVIEW_SCHEMA
+            or self.identity.get("schema_version") != LIDAR_FIELD_REVIEW_SCHEMA
+            or not isinstance(identity_sha256, str)
+            or _SHA256.fullmatch(identity_sha256) is None
+            or hashlib.sha256(_canonical_json(self.identity)).hexdigest() != identity_sha256
+            or self.root.name != f"lidar-field-review-{identity_sha256}"
+            or self.manifest.get("review_id") != self.root.name
+        ):
+            raise LidarGroundError("LiDAR field-review identity is invalid")
+        artifacts = _validate_artifacts(self.root, self.manifest.get("artifacts"))
+        self.arrays = np.load(artifacts["field-review"], allow_pickle=False)
+        self.report = _read_json(artifacts["field-review-report"])
+        self.preview_paths = {
+            role.removeprefix("preview-"): path
+            for role, path in artifacts.items()
+            if role.startswith("preview-")
+        }
+        try:
+            self._validate()
+        except BaseException:
+            self.close()
+            raise
+        self.review_id = self.root.name
+
+    def close(self) -> None:
+        self.arrays.close()
+
+    def _validate(self) -> None:
+        windows = _list(self.report.get("windows"), "field-review windows")
+        parsed_windows = [
+            _object(item, f"field-review window {index}") for index, item in enumerate(windows)
+        ]
+        source = _object(self.report.get("source"), "field-review source")
+        decision = _object(self.report.get("decision"), "field-review decision")
+        authority = _object(self.report.get("authority"), "field-review authority")
+        offsets = self.arrays["window_offsets"]
+        points = self.arrays["points_xyz_map"]
+        point_count = _nonnegative_int(
+            self.identity.get("display_point_count"),
+            "field-review point count",
+        )
+        required = {
+            "window_offsets",
+            "points_xyz_map",
+            "current_ground",
+            "current_assigned",
+            "candidate_ground",
+            "candidate_assigned",
+        }
+        if (
+            self.report.get("schema_version") != LIDAR_FIELD_REVIEW_REPORT_SCHEMA
+            or self.report.get("review_id") != self.root.name
+            or self.report.get("status") != "diagnostic-only"
+            or self.report.get("ground_truth") is not False
+            or source.get("representation") != "legacy-e10-vendor-map-with-pose"
+            or source.get("intensity_available") is not False
+            or source.get("raw_scan_accepted") is not False
+            or decision.get("status") != "visual-review-only"
+            or decision.get("production_promotion") is not False
+            or authority.get("commands_enabled") is not False
+            or authority.get("navigation_or_safety_accepted") is not False
+            or len(windows) != self.identity.get("window_count")
+            or set(self.preview_paths) != {str(item.get("key")) for item in parsed_windows}
+            or set(self.arrays.files) != required
+            or offsets.shape != (len(windows) + 1,)
+            or offsets.dtype != np.dtype("<i8")
+            or offsets[0] != 0
+            or offsets[-1] != point_count
+            or np.any(np.diff(offsets) <= 0)
+            or points.shape != (point_count, 3)
+            or points.dtype != np.dtype("<f4")
+            or not np.isfinite(points).all()
+            or _logical_sha256(self.arrays) != self.identity.get("logical_content_sha256")
+        ):
+            raise LidarGroundError("LiDAR field-review content is invalid")
+        for name in (
+            "current_ground",
+            "current_assigned",
+            "candidate_ground",
+            "candidate_assigned",
+        ):
+            value = self.arrays[name]
+            if value.shape != (point_count,) or value.dtype != np.dtype("u1") or np.any(value > 1):
+                raise LidarGroundError("LiDAR field-review mask is invalid")
+        for index, window in enumerate(parsed_windows):
+            start = int(offsets[index])
+            end = int(offsets[index + 1])
+            if (
+                window.get("index") != index
+                or _SAFE_KEY.fullmatch(str(window.get("key"))) is None
+                or window.get("display_point_count") != end - start
+                or not 0 < end - start <= MAX_FIELD_REVIEW_WINDOW_POINTS
+                or _nonnegative_int(
+                    window.get("source_lidar_samples"),
+                    "field-review source samples",
+                )
+                < 1
+                or _nonnegative_int(
+                    window.get("source_point_count"),
+                    "field-review source points",
+                )
+                < end - start
+                or not isinstance(window.get("label"), str)
+                or not window["label"].strip()
+            ):
+                raise LidarGroundError("LiDAR field-review window is invalid")
+
+    def window_detail(self, window_index: int) -> dict[str, object]:
+        windows = _list(self.report["windows"], "field-review windows")
+        if not 0 <= window_index < len(windows):
+            raise IndexError(window_index)
+        window = _object(windows[window_index], "field-review window")
+        offsets = self.arrays["window_offsets"]
+        start = int(offsets[window_index])
+        end = int(offsets[window_index + 1])
+        current_ground = self.arrays["current_ground"][start:end]
+        candidate_ground = self.arrays["candidate_ground"][start:end]
+        disagreement = (current_ground != candidate_ground).astype(np.uint8)
+        return {
+            "schema_version": LIDAR_FIELD_REVIEW_WINDOW_SCHEMA,
+            "review_id": self.review_id,
+            "display_name": self.report["display_name"],
+            "session_id": self.report["session_id"],
+            "source_pack_id": self.report["source_pack_id"],
+            "window_index": window_index,
+            "window_count": len(windows),
+            "window": window,
+            "point_count": end - start,
+            "coordinate_frame": "map",
+            "distance_unit": "m",
+            "intensity": {
+                "available": False,
+                "reason": "E10 derivative did not retain rgbi/intensity",
+            },
+            "points_xyz_m": self.arrays["points_xyz_map"][start:end].astype(np.float64).tolist(),
+            "masks": {
+                "current_ground": current_ground.astype(np.int64).tolist(),
+                "current_assigned": self.arrays["current_assigned"][start:end]
+                .astype(np.int64)
+                .tolist(),
+                "candidate_ground": candidate_ground.astype(np.int64).tolist(),
+                "candidate_assigned": self.arrays["candidate_assigned"][start:end]
+                .astype(np.int64)
+                .tolist(),
+                "disagreement": disagreement.astype(np.int64).tolist(),
+            },
+            "counts": {
+                "current_ground": int(np.count_nonzero(current_ground)),
+                "candidate_ground": int(np.count_nonzero(candidate_ground)),
+                "disagreement": int(np.count_nonzero(disagreement)),
+            },
+            "preview_url": (
+                f"/api/v1/lidar/field-reviews/{self.review_id}/windows/{window_index}/preview"
+            ),
+            "access": "read-only",
+            "ground_truth": False,
+            "authority": {
+                "commands_enabled": False,
+                "navigation_or_safety_accepted": False,
+            },
+        }
+
+
+def build_lidar_field_review(
+    source: E10LidarFieldSource,
+    output_root: Path,
+    *,
+    patchwork: GroundSegmenter,
+    profile: GroundBenchmarkProfile,
+    preview_paths: Mapping[str, Path],
+    windows: Sequence[FieldReviewWindowSpec] = RAVNOVES00_CENTRAL_WINDOWS,
+    display_name: str = "RAVNOVES00 · центральный городской интервал",
+    default_window_index: int = 2,
+    maximum_display_points: int = MAX_FIELD_REVIEW_WINDOW_POINTS,
+) -> Path:
+    """Build accumulated field windows without upgrading legacy input evidence."""
+
+    if (
+        not windows
+        or not 0 <= default_window_index < len(windows)
+        or not 1 <= maximum_display_points <= MAX_FIELD_REVIEW_WINDOW_POINTS
+        or set(preview_paths) != {window.key for window in windows}
+    ):
+        raise LidarGroundError("LiDAR field-review build configuration is invalid")
+    times = source.arrays["session_seconds"]
+    available = source.arrays["sample_available"]
+    source_offsets = source.arrays["cloud_offsets"]
+    points_map = source.arrays["cloud_points_map"]
+    positions = source.arrays["pose_positions_map"]
+    quaternions = source.arrays["pose_quaternions_map_from_lidar"]
+    source_frame_indices = source.arrays["source_frame_indices"]
+    current = LocalPercentileGroundSegmenter(profile)
+    window_arrays: list[dict[str, npt.NDArray[Any]]] = []
+    window_reports: list[dict[str, object]] = []
+    current_latency: list[float] = []
+    candidate_latency: list[float] = []
+    current_fraction: list[float] = []
+    candidate_fraction: list[float] = []
+    disagreement_fraction: list[float] = []
+    algorithm_iou: list[float] = []
+    selected_source_point_count = 0
+
+    for window_index, spec in enumerate(windows):
+        source_rows = np.flatnonzero(
+            available & (times >= spec.start_seconds) & (times <= spec.end_seconds)
+        )
+        if source_rows.size == 0:
+            raise LidarGroundError("LiDAR field-review window has no source samples")
+        collected_points: list[npt.NDArray[np.float32]] = []
+        collected_current: list[npt.NDArray[np.uint8]] = []
+        collected_current_assigned: list[npt.NDArray[np.uint8]] = []
+        collected_candidate: list[npt.NDArray[np.uint8]] = []
+        collected_candidate_assigned: list[npt.NDArray[np.uint8]] = []
+        source_point_count = 0
+        for source_row in source_rows:
+            start = int(source_offsets[source_row])
+            end = int(source_offsets[source_row + 1])
+            cloud = np.asarray(points_map[start:end], dtype=np.float32)
+            source_point_count += cloud.shape[0]
+            current_input = np.zeros((cloud.shape[0], 4), dtype=np.float32)
+            current_input[:, :3] = cloud
+            current_result = current.segment(current_input)
+            try:
+                position = positions[source_row]
+                orientation = quaternions[source_row]
+                points_sensor = map_points_to_lidar(
+                    cloud,
+                    position_map_xyz=(
+                        float(position[0]),
+                        float(position[1]),
+                        float(position[2]),
+                    ),
+                    orientation_map_from_lidar_xyzw=(
+                        float(orientation[0]),
+                        float(orientation[1]),
+                        float(orientation[2]),
+                        float(orientation[3]),
+                    ),
+                )
+            except CalibratedProjectionError as exc:
+                raise LidarGroundError("LiDAR field-review pose conversion failed") from exc
+            candidate_input = np.zeros((cloud.shape[0], 4), dtype=np.float32)
+            candidate_input[:, :3] = points_sensor.astype(np.float32)
+            if profile.patchwork_map_vertical_origin_offset_m:
+                candidate_input[:, 2] -= profile.patchwork_map_vertical_origin_offset_m
+            candidate_result = patchwork.segment(candidate_input)
+            _segmentation(
+                current_result.ground_mask,
+                current_result.assigned_mask,
+                cloud.shape[0],
+                "current",
+            )
+            _segmentation(
+                candidate_result.ground_mask,
+                candidate_result.assigned_mask,
+                cloud.shape[0],
+                "candidate",
+            )
+            collected_points.append(cloud)
+            collected_current.append(current_result.ground_mask.astype(np.uint8))
+            collected_current_assigned.append(current_result.assigned_mask.astype(np.uint8))
+            collected_candidate.append(candidate_result.ground_mask.astype(np.uint8))
+            collected_candidate_assigned.append(candidate_result.assigned_mask.astype(np.uint8))
+            current_latency.append(current_result.latency_ms)
+            candidate_latency.append(candidate_result.latency_ms)
+            current_fraction.append(float(np.mean(current_result.ground_mask)))
+            candidate_fraction.append(float(np.mean(candidate_result.ground_mask)))
+            disagreement_fraction.append(
+                float(np.mean(current_result.ground_mask != candidate_result.ground_mask))
+            )
+            intersection = int(
+                np.count_nonzero(current_result.ground_mask & candidate_result.ground_mask)
+            )
+            union = int(np.count_nonzero(current_result.ground_mask | candidate_result.ground_mask))
+            algorithm_iou.append(float(intersection / union) if union else 1.0)
+
+        combined_points = np.concatenate(collected_points)
+        combined_current = np.concatenate(collected_current)
+        combined_current_assigned = np.concatenate(collected_current_assigned)
+        combined_candidate = np.concatenate(collected_candidate)
+        combined_candidate_assigned = np.concatenate(collected_candidate_assigned)
+        selected_source_point_count += source_point_count
+        selected = _uniform_indices(combined_points.shape[0], maximum_display_points)
+        displayed = {
+            "points_xyz_map": combined_points[selected].astype("<f4"),
+            "current_ground": combined_current[selected].astype("u1"),
+            "current_assigned": combined_current_assigned[selected].astype("u1"),
+            "candidate_ground": combined_candidate[selected].astype("u1"),
+            "candidate_assigned": combined_candidate_assigned[selected].astype("u1"),
+        }
+        window_arrays.append(displayed)
+        preview_row = int(
+            np.argmin(
+                np.abs(source_frame_indices.astype(np.int64) - spec.preview_source_frame_index)
+            )
+        )
+        window_reports.append(
+            {
+                "index": window_index,
+                "key": spec.key,
+                "label": spec.label,
+                "start_seconds": spec.start_seconds,
+                "end_seconds": spec.end_seconds,
+                "midpoint_seconds": (spec.start_seconds + spec.end_seconds) / 2,
+                "source_lidar_samples": int(source_rows.shape[0]),
+                "source_point_count": source_point_count,
+                "display_point_count": int(selected.shape[0]),
+                "source_frame_start": int(source_frame_indices[source_rows[0]]),
+                "source_frame_end": int(source_frame_indices[source_rows[-1]]),
+                "preview_source_frame_index": int(source_frame_indices[preview_row]),
+                "preview_session_seconds": float(times[preview_row]),
+            }
+        )
+
+    offsets = np.concatenate(
+        (
+            np.asarray([0], dtype="<i8"),
+            np.cumsum(
+                [item["points_xyz_map"].shape[0] for item in window_arrays],
+                dtype=np.int64,
+            ),
+        )
+    ).astype("<i8")
+    arrays: dict[str, npt.NDArray[Any]] = {
+        "window_offsets": offsets,
+        "points_xyz_map": np.concatenate([item["points_xyz_map"] for item in window_arrays]).astype(
+            "<f4"
+        ),
+        "current_ground": np.concatenate([item["current_ground"] for item in window_arrays]).astype(
+            "u1"
+        ),
+        "current_assigned": np.concatenate(
+            [item["current_assigned"] for item in window_arrays]
+        ).astype("u1"),
+        "candidate_ground": np.concatenate(
+            [item["candidate_ground"] for item in window_arrays]
+        ).astype("u1"),
+        "candidate_assigned": np.concatenate(
+            [item["candidate_assigned"] for item in window_arrays]
+        ).astype("u1"),
+    }
+    logical_content_sha256 = _logical_sha256(arrays)
+    identity = {
+        "schema_version": LIDAR_FIELD_REVIEW_SCHEMA,
+        "source_pack_id": source.pack_id,
+        "source_pack_identity_sha256": source.manifest["identity_sha256"],
+        "source_artifact_sha256": source.manifest["artifact"]["sha256"],
+        "session_id": source.identity["session_id"],
+        "display_name": display_name,
+        "windows": [window.to_dict() for window in windows],
+        "window_count": len(windows),
+        "default_window_index": default_window_index,
+        "source_point_count": selected_source_point_count,
+        "display_point_count": int(arrays["points_xyz_map"].shape[0]),
+        "sampling": {
+            "method": "uniform-point-index-per-window",
+            "maximum_points_per_window": maximum_display_points,
+        },
+        "profile": profile.to_dict(),
+        "providers": {
+            "current": dict(current.identity),
+            "candidate": dict(patchwork.identity),
+        },
+        "logical_content_sha256": logical_content_sha256,
+        "producer_sha256": _sha256(Path(__file__).resolve(strict=True)),
+        "ground_truth": False,
+        "authority": {
+            "commands_enabled": False,
+            "navigation_or_safety_accepted": False,
+        },
+    }
+    identity_sha256 = hashlib.sha256(_canonical_json(identity)).hexdigest()
+    review_id = f"lidar-field-review-{identity_sha256}"
+    output_parent = output_root.expanduser().resolve()
+    output_parent.mkdir(mode=0o700, parents=True, exist_ok=True)
+    output = output_parent / review_id
+    if output.exists():
+        existing = LidarFieldReviewV1(output)
+        existing.close()
+        return output
+
+    report = {
+        "schema_version": LIDAR_FIELD_REVIEW_REPORT_SCHEMA,
+        "review_id": review_id,
+        "display_name": display_name,
+        "session_id": source.identity["session_id"],
+        "source_pack_id": source.pack_id,
+        "status": "diagnostic-only",
+        "source": {
+            "timeline_start_seconds": source.identity["timeline_start_seconds"],
+            "timeline_end_seconds": source.identity["timeline_end_seconds"],
+            "available_lidar_frames": source.identity["available_lidar_frames"],
+            "point_count": source.identity["point_count"],
+            "representation": "legacy-e10-vendor-map-with-pose",
+            "intensity_available": False,
+            "raw_scan_accepted": False,
+        },
+        "selection": {
+            "purpose": "operator-readable-central-urban-field-review",
+            "default_window_index": default_window_index,
+            "accumulation": "per-source-frame masks accumulated in map frame",
+            "sampling": identity["sampling"],
+        },
+        "windows": window_reports,
+        "metrics": {
+            "source_samples": len(current_latency),
+            "current": {
+                "provider": dict(current.identity),
+                "ground_fraction": _distribution(current_fraction),
+                "latency_ms": _distribution(current_latency),
+            },
+            "candidate": {
+                "provider": dict(patchwork.identity),
+                "ground_fraction": _distribution(candidate_fraction),
+                "latency_ms": _distribution(candidate_latency),
+            },
+            "comparison": {
+                "algorithm_to_algorithm_ground_iou": _distribution(algorithm_iou),
+                "ground_disagreement_fraction": _distribution(disagreement_fraction),
+                "is_accuracy_metric": False,
+            },
+        },
+        "decision": {
+            "status": "visual-review-only",
+            "production_promotion": False,
+            "reasons": [
+                "legacy E10 derivative does not retain intensity",
+                "source remains a vendor-mapped LIO product",
+                "independent ground labels are missing",
+            ],
+        },
+        "ground_truth": False,
+        "authority": identity["authority"],
+    }
+    staging = output_parent / f".{review_id}.{os.getpid()}.incomplete"
+    staging.mkdir(mode=0o700, exist_ok=False)
+    try:
+        arrays_path = staging / FIELD_REVIEW_ARRAYS_NAME
+        np.savez_compressed(arrays_path, **arrays)  # type: ignore[arg-type]
+        report_path = staging / FIELD_REVIEW_REPORT_NAME
+        _write_json(report_path, report)
+        artifacts = [
+            _artifact("field-review", arrays_path, "application/x-npz"),
+            _artifact("field-review-report", report_path, "application/json"),
+        ]
+        for spec in windows:
+            preview_source = preview_paths[spec.key].expanduser().resolve(strict=True)
+            if not preview_source.is_file() or preview_source.is_symlink():
+                raise LidarGroundError("LiDAR field-review preview is invalid")
+            preview_target = staging / f"preview-{spec.key}.jpg"
+            shutil.copy2(preview_source, preview_target)
+            artifacts.append(
+                _artifact(
+                    f"preview-{spec.key}",
+                    preview_target,
+                    "image/jpeg",
+                )
+            )
+        manifest = {
+            "schema_version": LIDAR_FIELD_REVIEW_SCHEMA,
+            "review_id": review_id,
+            "identity_sha256": identity_sha256,
+            "identity": identity,
+            "created_at_utc": _utc_now(),
+            "classification": "private-derived-lidar-diagnostic",
+            "ground_truth": False,
+            "artifacts": artifacts,
+        }
+        _write_json(staging / FIELD_REVIEW_MANIFEST_NAME, manifest)
+        os.replace(staging, output)
+    except BaseException:
+        shutil.rmtree(staging, ignore_errors=True)
+        raise
+    validation = LidarFieldReviewV1(output)
+    validation.close()
+    return output
+
+
+def lidar_field_review_catalog_item(review: LidarFieldReviewV1) -> dict[str, object]:
+    report = review.report
+    metrics = _object(report["metrics"], "field-review metrics")
+    return {
+        "review_id": review.review_id,
+        "display_name": report["display_name"],
+        "session_id": report["session_id"],
+        "source_pack_id": report["source_pack_id"],
+        "status": report["status"],
+        "source": report["source"],
+        "selection": report["selection"],
+        "windows": report["windows"],
+        "metrics": metrics,
+        "decision": report["decision"],
+        "created_at_utc": review.manifest.get("created_at_utc"),
+        "ground_truth": False,
+        "authority": report["authority"],
+    }
+
+
+def _uniform_indices(count: int, maximum: int) -> npt.NDArray[np.int64]:
+    if count <= maximum:
+        return np.arange(count, dtype=np.int64)
+    return np.linspace(0, count - 1, maximum, dtype=np.int64)
+
+
+def _segmentation(
+    ground_mask: npt.NDArray[np.bool_],
+    assigned_mask: npt.NDArray[np.bool_],
+    point_count: int,
+    label: str,
+) -> None:
+    if (
+        ground_mask.shape != (point_count,)
+        or ground_mask.dtype != np.dtype("?")
+        or assigned_mask.shape != (point_count,)
+        or assigned_mask.dtype != np.dtype("?")
+        or np.any(ground_mask & ~assigned_mask)
+    ):
+        raise LidarGroundError(f"LiDAR field-review {label} mask is invalid")
+
+
+def _distribution(values: Sequence[float]) -> dict[str, float | int]:
+    array = np.asarray(values, dtype=np.float64)
+    if array.size == 0 or not np.isfinite(array).all():
+        raise LidarGroundError("LiDAR field-review distribution is invalid")
+    return {
+        "sample_count": int(array.shape[0]),
+        "minimum": float(np.min(array)),
+        "mean": float(np.mean(array)),
+        "p50": float(np.percentile(array, 50)),
+        "p95": float(np.percentile(array, 95)),
+        "maximum": float(np.max(array)),
+    }
+
+
+def _logical_sha256(arrays: Mapping[str, npt.NDArray[Any]]) -> str:
+    digest = hashlib.sha256()
+    for name in sorted(arrays):
+        array = np.ascontiguousarray(arrays[name])
+        digest.update(name.encode())
+        digest.update(array.dtype.str.encode())
+        digest.update(_canonical_json(list(array.shape)))
+        digest.update(memoryview(array).cast("B"))
+    return digest.hexdigest()
+
+
+def _validate_artifacts(root: Path, value: object) -> dict[str, Path]:
+    artifacts = _list(value, "field-review artifacts")
+    resolved: dict[str, Path] = {}
+    for value in artifacts:
+        item = _object(value, "field-review artifact")
+        role = item.get("role")
+        relative = item.get("path")
+        if (
+            not isinstance(role, str)
+            or role in resolved
+            or not isinstance(relative, str)
+            or Path(relative).name != relative
+            or not isinstance(item.get("byte_length"), int)
+            or isinstance(item.get("byte_length"), bool)
+            or item["byte_length"] < 1
+            or _SHA256.fullmatch(str(item.get("sha256"))) is None
+        ):
+            raise LidarGroundError("LiDAR field-review artifact descriptor is invalid")
+        path = root / relative
+        if (
+            path.is_symlink()
+            or not path.is_file()
+            or path.stat().st_size != item["byte_length"]
+            or _sha256(path) != item["sha256"]
+        ):
+            raise LidarGroundError("LiDAR field-review artifact is invalid")
+        resolved[role] = path
+    if "field-review" not in resolved or "field-review-report" not in resolved:
+        raise LidarGroundError("LiDAR field-review artifacts are incomplete")
+    return resolved
+
+
+def _artifact(role: str, path: Path, media_type: str) -> dict[str, object]:
+    return {
+        "role": role,
+        "path": path.name,
+        "media_type": media_type,
+        "byte_length": path.stat().st_size,
+        "sha256": _sha256(path),
+    }
+
+
+def _read_json(path: Path) -> dict[str, Any]:
+    if path.is_symlink() or not path.is_file() or path.stat().st_size > 4_000_000:
+        raise LidarGroundError("LiDAR field-review JSON artifact is invalid")
+    try:
+        value = json.loads(path.read_text(encoding="utf-8"))
+    except (OSError, UnicodeError, json.JSONDecodeError) as exc:
+        raise LidarGroundError("LiDAR field-review JSON artifact is unreadable") from exc
+    return _object(value, "LiDAR field-review JSON")
+
+
+def _write_json(path: Path, value: Mapping[str, object]) -> None:
+    path.write_bytes(
+        json.dumps(
+            value,
+            ensure_ascii=False,
+            sort_keys=True,
+            indent=2,
+            allow_nan=False,
+        ).encode()
+        + b"\n"
+    )
+
+
+def _canonical_json(value: object) -> bytes:
+    return json.dumps(
+        value,
+        ensure_ascii=False,
+        sort_keys=True,
+        separators=(",", ":"),
+        allow_nan=False,
+    ).encode()
+
+
+def _object(value: object, label: str) -> dict[str, Any]:
+    if not isinstance(value, dict) or not all(isinstance(key, str) for key in value):
+        raise LidarGroundError(f"{label} must be an object")
+    return value
+
+
+def _list(value: object, label: str) -> list[Any]:
+    if not isinstance(value, list):
+        raise LidarGroundError(f"{label} must be a list")
+    return value
+
+
+def _nonnegative_int(value: object, label: str) -> int:
+    if not isinstance(value, int) or isinstance(value, bool) or value < 0:
+        raise LidarGroundError(f"{label} must be a non-negative integer")
+    return value
+
+
+def _sha256(path: Path) -> str:
+    digest = hashlib.sha256()
+    with path.open("rb") as stream:
+        while chunk := stream.read(1024 * 1024):
+            digest.update(chunk)
+    return digest.hexdigest()
+
+
+def _utc_now() -> str:
+    return datetime.now(UTC).isoformat(timespec="milliseconds").replace("+00:00", "Z")
diff --git a/src/k1link/web/lidar_api.py b/src/k1link/web/lidar_api.py
index 3d5b923..7d91522 100644
--- a/src/k1link/web/lidar_api.py
+++ b/src/k1link/web/lidar_api.py
@@ -6,13 +6,15 @@ from collections.abc import Callable
 from pathlib import Path
 from typing import Any, Final
 
-from fastapi import APIRouter, HTTPException, Query
+from fastapi import APIRouter, HTTPException, Query, Response
 
 from k1link.compute import (
+    LidarFieldReviewV1,
     LidarGroundBenchmarkV1,
     LidarGroundError,
     LidarReplayError,
     LidarReplayPackV2,
+    lidar_field_review_catalog_item,
     lidar_ground_benchmark_catalog_item,
     lidar_ground_frame_detail,
     lidar_pack_catalog_item,
@@ -21,8 +23,10 @@ from k1link.compute import (
 
 LIDAR_CATALOG_SCHEMA: Final = "missioncore.lidar-replay-pack-catalog/v1"
 LIDAR_GROUND_CATALOG_SCHEMA: Final = "missioncore.lidar-ground-benchmark-catalog/v1"
+LIDAR_FIELD_REVIEW_CATALOG_SCHEMA: Final = "missioncore.lidar-field-review-catalog/v1"
 _PACK_ID = re.compile(r"^lidar-replay-pack-[a-f0-9]{64}$")
 _BENCHMARK_ID = re.compile(r"^ground-benchmark-[a-f0-9]{64}$")
+_FIELD_REVIEW_ID = re.compile(r"^lidar-field-review-[a-f0-9]{64}$")
 RootProvider = Callable[[], Path | None]
 
 
@@ -36,10 +40,16 @@ def configured_lidar_ground_root() -> Path | None:
     return Path(value).expanduser().absolute() if value else None
 
 
+def configured_lidar_field_review_root() -> Path | None:
+    value = os.environ.get("MISSIONCORE_LIDAR_FIELD_REVIEW_ROOT", "").strip()
+    return Path(value).expanduser().absolute() if value else None
+
+
 def build_lidar_router(
     *,
     root_provider: RootProvider = configured_lidar_replay_root,
     ground_root_provider: RootProvider = configured_lidar_ground_root,
+    field_review_root_provider: RootProvider = configured_lidar_field_review_root,
 ) -> APIRouter:
     router = APIRouter(prefix="/api/v1/lidar", tags=["lidar"])
 
@@ -269,4 +279,139 @@ def build_lidar_router(
                 detail="LiDAR ground frame не прошёл проверку целостности",
             ) from exc
 
+    @router.get("/field-reviews")
+    def list_lidar_field_reviews(
+        limit: int = Query(default=10, ge=1, le=50),
+    ) -> dict[str, Any]:
+        root = field_review_root_provider()
+        if root is None or not root.is_dir():
+            return {
+                "schema_version": LIDAR_FIELD_REVIEW_CATALOG_SCHEMA,
+                "configured": root is not None,
+                "items": [],
+                "valid_total": 0,
+                "invalid_total": 0,
+                "access": "read-only",
+            }
+        items: list[dict[str, object]] = []
+        invalid_total = 0
+        candidates = sorted(
+            (
+                candidate
+                for candidate in root.iterdir()
+                if candidate.is_dir() and _FIELD_REVIEW_ID.fullmatch(candidate.name) is not None
+            ),
+            key=lambda candidate: candidate.stat().st_mtime_ns,
+            reverse=True,
+        )
+        for candidate in candidates:
+            try:
+                review = LidarFieldReviewV1(candidate)
+                try:
+                    items.append(lidar_field_review_catalog_item(review))
+                finally:
+                    review.close()
+            except (LidarGroundError, OSError):
+                invalid_total += 1
+        return {
+            "schema_version": LIDAR_FIELD_REVIEW_CATALOG_SCHEMA,
+            "configured": True,
+            "items": items[:limit],
+            "valid_total": len(items),
+            "invalid_total": invalid_total,
+            "access": "read-only",
+        }
+
+    @router.get("/field-reviews/{review_id}/windows/{window_index}")
+    def get_lidar_field_review_window(
+        review_id: str,
+        window_index: int,
+    ) -> dict[str, object]:
+        if _FIELD_REVIEW_ID.fullmatch(review_id) is None or window_index < 0:
+            raise HTTPException(
+                status_code=404,
+                detail="LiDAR field-review window не найден",
+            )
+        root = field_review_root_provider()
+        if root is None or not root.is_dir():
+            raise HTTPException(
+                status_code=503,
+                detail="LiDAR field-review storage не настроен",
+            )
+        candidate = root / review_id
+        if not candidate.is_dir():
+            raise HTTPException(
+                status_code=404,
+                detail="LiDAR field review не найден",
+            )
+        try:
+            review = LidarFieldReviewV1(candidate)
+            try:
+                return review.window_detail(window_index)
+            finally:
+                review.close()
+        except IndexError as exc:
+            raise HTTPException(
+                status_code=404,
+                detail="LiDAR field-review window не найден",
+            ) from exc
+        except (LidarGroundError, OSError) as exc:
+            raise HTTPException(
+                status_code=409,
+                detail="LiDAR field review не прошёл проверку целостности",
+            ) from exc
+
+    @router.get("/field-reviews/{review_id}/windows/{window_index}/preview")
+    def get_lidar_field_review_preview(
+        review_id: str,
+        window_index: int,
+    ) -> Response:
+        if _FIELD_REVIEW_ID.fullmatch(review_id) is None or window_index < 0:
+            raise HTTPException(
+                status_code=404,
+                detail="LiDAR field-review preview не найден",
+            )
+        root = field_review_root_provider()
+        if root is None or not root.is_dir():
+            raise HTTPException(
+                status_code=503,
+                detail="LiDAR field-review storage не настроен",
+            )
+        candidate = root / review_id
+        if not candidate.is_dir():
+            raise HTTPException(
+                status_code=404,
+                detail="LiDAR field review не найден",
+            )
+        try:
+            review = LidarFieldReviewV1(candidate)
+            try:
+                windows = review.report.get("windows")
+                if not isinstance(windows, list) or not 0 <= window_index < len(windows):
+                    raise IndexError(window_index)
+                window = windows[window_index]
+                if not isinstance(window, dict) or not isinstance(window.get("key"), str):
+                    raise LidarGroundError("LiDAR field-review preview key is invalid")
+                preview = review.preview_paths.get(window["key"])
+                if preview is None:
+                    raise LidarGroundError("LiDAR field-review preview is missing")
+                content = preview.read_bytes()
+            finally:
+                review.close()
+        except IndexError as exc:
+            raise HTTPException(
+                status_code=404,
+                detail="LiDAR field-review preview не найден",
+            ) from exc
+        except (LidarGroundError, OSError) as exc:
+            raise HTTPException(
+                status_code=409,
+                detail="LiDAR field-review preview не прошёл проверку",
+            ) from exc
+        return Response(
+            content=content,
+            media_type="image/jpeg",
+            headers={"Cache-Control": "private, max-age=31536000, immutable"},
+        )
+
     return router
diff --git a/tests/test_lidar_field_review.py b/tests/test_lidar_field_review.py
new file mode 100644
index 0000000..fc710e6
--- /dev/null
+++ b/tests/test_lidar_field_review.py
@@ -0,0 +1,238 @@
+from __future__ import annotations
+
+import hashlib
+import json
+from pathlib import Path
+
+import numpy as np
+import pytest
+from fastapi import APIRouter
+from fastapi.routing import APIRoute
+
+from k1link.compute import (
+    E10_LIDAR_PACK_SCHEMA,
+    E10LidarFieldSource,
+    FieldReviewWindowSpec,
+    GroundBenchmarkProfile,
+    GroundSegmentation,
+    LidarFieldReviewV1,
+    LidarGroundError,
+    build_lidar_field_review,
+)
+from k1link.web.lidar_api import build_lidar_router
+
+
+class _Candidate:
+    @property
+    def identity(self) -> dict[str, object]:
+        return {
+            "provider_id": "test-patchwork/v1",
+            "source_commit": "c" * 40,
+        }
+
+    def segment(self, xyzi: np.ndarray) -> GroundSegmentation:
+        ground = xyzi[:, 0] <= 0.25
+        assigned = np.ones(xyzi.shape[0], dtype=np.bool_)
+        return GroundSegmentation(ground, assigned, 0.25)
+
+
+def _canonical_json(value: object) -> bytes:
+    return json.dumps(
+        value,
+        ensure_ascii=False,
+        sort_keys=True,
+        separators=(",", ":"),
+        allow_nan=False,
+    ).encode()
+
+
+def _sha256(path: Path) -> str:
+    return hashlib.sha256(path.read_bytes()).hexdigest()
+
+
+def _source_pack(root: Path) -> Path:
+    base_points = np.asarray(
+        [
+            [-1.0, -1.0, 0.00],
+            [-0.5, -1.0, 0.02],
+            [0.0, -1.0, 0.01],
+            [0.5, -1.0, 0.03],
+            [1.0, -1.0, 0.00],
+            [-1.0, 0.0, 0.01],
+            [-0.5, 0.0, 0.02],
+            [0.0, 0.0, 0.04],
+            [0.5, 0.0, 0.35],
+            [1.0, 0.0, 0.70],
+        ],
+        dtype=np.float32,
+    )
+    points = np.concatenate(
+        [base_points + np.asarray([index * 0.1, index, 0], dtype=np.float32) for index in range(4)]
+    ).astype("<f4")
+    arrays = {
+        "frame_indices": np.arange(4, dtype="<i8"),
+        "source_frame_indices": np.asarray([100, 110, 120, 130], dtype="<i8"),
+        "session_seconds": np.asarray([1.0, 2.0, 3.0, 4.0], dtype="<f8"),
+        "sample_available": np.ones(4, dtype="?"),
+        "cloud_offsets": np.asarray([0, 10, 20, 30, 40], dtype="<i8"),
+        "cloud_points_map": points,
+        "pose_positions_map": np.zeros((4, 3), dtype="<f8"),
+        "pose_quaternions_map_from_lidar": np.tile(
+            np.asarray([0.0, 0.0, 0.0, 1.0], dtype="<f8"),
+            (4, 1),
+        ),
+        "lidar_camera_delta_ms": np.zeros(4, dtype="<f8"),
+        "pose_point_delta_ms": np.zeros(4, dtype="<f8"),
+        "intrinsic_fx_fy_cx_cy": np.asarray(
+            [100.0, 100.0, 50.0, 50.0],
+            dtype="<f8",
+        ),
+        "distortion_kb4": np.zeros(4, dtype="<f8"),
+        "t_camera_from_lidar": np.eye(4, dtype="<f8"),
+    }
+    identity = {
+        "schema_version": E10_LIDAR_PACK_SCHEMA,
+        "session_id": "synthetic-ravnoves00",
+        "frame_count": 4,
+        "available_lidar_frames": 4,
+        "point_count": 40,
+        "timeline_start_seconds": 1.0,
+        "timeline_end_seconds": 4.0,
+    }
+    identity_sha256 = hashlib.sha256(_canonical_json(identity)).hexdigest()
+    pack = root / f"e10-lidar-pack-{identity_sha256}"
+    pack.mkdir(parents=True)
+    arrays_path = pack / "lidar-pack.npz"
+    np.savez_compressed(arrays_path, **arrays)  # type: ignore[arg-type]
+    manifest = {
+        "schema_version": E10_LIDAR_PACK_SCHEMA,
+        "pack_id": pack.name,
+        "identity_sha256": identity_sha256,
+        "identity": identity,
+        "artifact": {
+            "path": arrays_path.name,
+            "media_type": "application/x-npz",
+            "byte_length": arrays_path.stat().st_size,
+            "sha256": _sha256(arrays_path),
+        },
+    }
+    (pack / "manifest.json").write_bytes(json.dumps(manifest, sort_keys=True).encode())
+    return pack
+
+
+def _endpoint(router: APIRouter, path: str) -> object:
+    for route in router.routes:
+        if (
+            isinstance(route, APIRoute)
+            and route.path == path
+            and route.methods is not None
+            and "GET" in route.methods
+        ):
+            return route.endpoint
+    raise AssertionError(f"GET {path} route is missing")
+
+
+def test_field_review_is_accumulated_path_free_visual_evidence(
+    tmp_path: Path,
+) -> None:
+    source = E10LidarFieldSource(_source_pack(tmp_path / "source"))
+    windows = (
+        FieldReviewWindowSpec("street-a", "Улица A", 0.5, 2.1, 100),
+        FieldReviewWindowSpec("street-b", "Улица B", 2.5, 4.1, 120),
+    )
+    previews = {}
+    for window in windows:
+        preview = tmp_path / f"{window.key}.jpg"
+        preview.write_bytes(b"\xff\xd8synthetic-jpeg\xff\xd9")
+        previews[window.key] = preview
+    profile = GroundBenchmarkProfile(
+        profile_id="synthetic-ravnoves00-field-review/v1",
+        patchwork_sensor_height_proxy_m=1.27,
+        patchwork_map_vertical_origin_offset_m=1.27,
+        patchwork_height_evidence="operator-estimated",
+    )
+    try:
+        output = build_lidar_field_review(
+            source,
+            tmp_path / "reviews",
+            patchwork=_Candidate(),
+            profile=profile,
+            preview_paths=previews,
+            windows=windows,
+            default_window_index=1,
+            maximum_display_points=12,
+        )
+    finally:
+        source.close()
+
+    review = LidarFieldReviewV1(output)
+    try:
+        detail = review.window_detail(1)
+        assert review.report["status"] == "diagnostic-only"
+        assert review.report["source"]["intensity_available"] is False
+        assert review.report["selection"]["default_window_index"] == 1
+        assert len(review.report["windows"]) == 2
+        assert review.report["windows"][0]["source_lidar_samples"] == 2
+        assert detail["point_count"] == 12
+        intensity = detail["intensity"]
+        authority = detail["authority"]
+        assert isinstance(intensity, dict)
+        assert isinstance(authority, dict)
+        assert intensity["available"] is False
+        assert detail["ground_truth"] is False
+        assert authority["commands_enabled"] is False
+        assert str(tmp_path) not in repr(detail)
+    finally:
+        review.close()
+
+    router = build_lidar_router(
+        root_provider=lambda: None,
+        ground_root_provider=lambda: None,
+        field_review_root_provider=lambda: output.parent,
+    )
+    catalog_route = _endpoint(router, "/api/v1/lidar/field-reviews")
+    window_route = _endpoint(
+        router,
+        "/api/v1/lidar/field-reviews/{review_id}/windows/{window_index}",
+    )
+    preview_route = _endpoint(
+        router,
+        "/api/v1/lidar/field-reviews/{review_id}/windows/{window_index}/preview",
+    )
+    catalog = catalog_route(limit=10)  # type: ignore[operator]
+    window = window_route(review_id=output.name, window_index=0)  # type: ignore[operator]
+    preview = preview_route(review_id=output.name, window_index=0)  # type: ignore[operator]
+
+    assert catalog["valid_total"] == 1
+    assert catalog["items"][0]["display_name"]
+    assert window["preview_url"].endswith("/windows/0/preview")
+    assert preview.media_type == "image/jpeg"
+    assert preview.body.startswith(b"\xff\xd8")
+    assert str(tmp_path) not in repr({"catalog": catalog, "window": window})
+
+
+def test_field_review_reader_rejects_tampered_identity(tmp_path: Path) -> None:
+    source = E10LidarFieldSource(_source_pack(tmp_path / "source"))
+    window = FieldReviewWindowSpec("street-a", "Улица A", 0.5, 4.1, 100)
+    preview = tmp_path / "street-a.jpg"
+    preview.write_bytes(b"\xff\xd8synthetic-jpeg\xff\xd9")
+    try:
+        output = build_lidar_field_review(
+            source,
+            tmp_path / "reviews",
+            patchwork=_Candidate(),
+            profile=GroundBenchmarkProfile(),
+            preview_paths={window.key: preview},
+            windows=(window,),
+            default_window_index=0,
+            maximum_display_points=12,
+        )
+    finally:
+        source.close()
+
+    manifest_path = output / "manifest.json"
+    manifest = json.loads(manifest_path.read_text(encoding="utf-8"))
+    manifest["identity"]["display_name"] = "tampered"
+    manifest_path.write_text(json.dumps(manifest), encoding="utf-8")
+    with pytest.raises(LidarGroundError, match="identity"):
+        LidarFieldReviewV1(output)