Update files

.woodpecker/build-arm.yml

@@ -8,24 +8,9 @@ labels:
 steps:
   - name: build-push
     image: docker
-    environment:
-      REGISTRY_HOST:
-        from_secret: registry_host
-      REGISTRY_USER:
-        from_secret: registry_user
-      REGISTRY_TOKEN:
-        from_secret: registry_token
     commands:
-      - echo "$REGISTRY_TOKEN" | docker login "$REGISTRY_HOST" -u "$REGISTRY_USER" --password-stdin
+      - if [ "$CI_COMMIT_BRANCH" = "main" ]; then export TAG=arm; else export TAG=${CI_COMMIT_BRANCH}-arm; fi
-      - export TAG=${CI_COMMIT_BRANCH}-arm
+      - docker build -f Dockerfile.jetson -t localhost:5000/detections:$TAG .
-      - export BUILD_DATE=$(date -u +%Y-%m-%dT%H:%M:%SZ)
+      - docker push localhost:5000/detections:$TAG
-      - |
-        docker build -f Dockerfile.jetson \
-          --build-arg CI_COMMIT_SHA=$CI_COMMIT_SHA \
-          --label org.opencontainers.image.revision=$CI_COMMIT_SHA \
-          --label org.opencontainers.image.created=$BUILD_DATE \
-          --label org.opencontainers.image.source=$CI_REPO_URL \
-          -t $REGISTRY_HOST/azaion/detections:$TAG .
-      - docker push $REGISTRY_HOST/azaion/detections:$TAG
     volumes:
       - /var/run/docker.sock:/var/run/docker.sock

Dockerfile.jetson

@@ -1,6 +1,4 @@
 FROM nvcr.io/nvidia/l4t-jetpack:r36.2.0
-ARG CI_COMMIT_SHA=unknown
-ENV AZAION_REVISION=$CI_COMMIT_SHA
 
 RUN apt-get update && apt-get install -y \
     python3 python3-pip python3-dev gcc \

_docs/00_problem/acceptance_criteria.md

@@ -5,7 +5,7 @@
 - Detections with confidence below `probability_threshold` (default: 0.25) are filtered out.
 - Overlapping detections with containment ratio > `tracking_intersection_threshold` (default: 0.6) are deduplicated, keeping the higher-confidence detection.
 - Tile duplicate detections are identified when all bounding box coordinates differ by less than 0.01 (TILE_DUPLICATE_CONFIDENCE_THRESHOLD).
-- Physical size filtering: detections exceeding `max_object_size_meters` for their class (defined in classes.json, range 2–20 meters) are removed.
+- Physical size filtering: detections exceeding `max_object_size_meters` for their class (defined in classes.json, range 2–20 meters) are removed when ground sampling distance can be computed from camera settings.
 
 ## Video Processing
 
@@ -17,7 +17,7 @@
 
 - Images ≤ 1.5× model dimensions (1280×1280): processed as single frame.
 - Larger images: tiled based on ground sampling distance. Tile physical size: 25 meters (METERS_IN_TILE). Tile overlap: `big_image_tile_overlap_percent` (default: 20%).
-- GSD calculation: `sensor_width * altitude / (focal_length * image_width)`.
+- GSD calculation: `sensor_width * altitude / (focal_length * image_width)` when `altitude` is provided.
 
 ## API
 

_docs/00_problem/input_data/data_parameters.md

@@ -36,7 +36,7 @@ Media path is resolved from the Annotations service via `GET /api/media/{media_i
 | tracking_intersection_threshold | float | 0.6 | Overlap ratio for NMS deduplication |
 | model_batch_size | int | 8 | Inference batch size |
 | big_image_tile_overlap_percent | int | 20 | Tile overlap for large images (0-100%) |
-| altitude | float | 400 | Camera altitude in meters |
+| altitude | float | optional | Camera altitude in meters. When omitted, GSD-based size filtering and image tiling are skipped. |
 | focal_length | float | 24 | Camera focal length in mm |
 | sensor_width | float | 23.5 | Camera sensor width in mm |
 

_docs/00_problem/input_data/expected_results/image_different_types_expected.csv

@@ -1 +1,3 @@
 center_x,center_y,width,height,label,confidence_min
+0.104737,0.375955,0.154999,0.144996,Truck,0.73
+0.576021,0.728599,0.1495,0.142923,Truck,0.83

_docs/00_problem/input_data/expected_results/image_large_expected.csv

@@ -1 +1,4 @@
 center_x,center_y,width,height,label,confidence_min
+0.214701,0.770089,0.274945,0.28434,ArmorVehicle,0.73
+0.588816,0.581807,0.269273,0.209632,ArmorVehicle,0.49
+0.314031,0.56274,0.039824,0.060367,MilitaryMan,0.32

_docs/00_problem/input_data/expected_results/image_small_expected.csv

@@ -1 +1,2 @@
 center_x,center_y,width,height,label,confidence_min
+0.465599,0.20807,0.137469,0.194655,ArmorVehicle,0.77

_docs/00_problem/input_data/expected_results/results_report.md

@@ -37,21 +37,18 @@ For videos, the additional field:
 ### Images
 
 | # | Input File | Description | Expected Result File | Expected Detection Count | Notes |
-|---|------------|-------------|---------------------|-------------------------|-------|
+|---|------------|-------------|---------------------|---|-------|
-| 1 | `image_small.jpg` | 1280×720 aerial, contains detectable objects | `image_small_expected.csv` | ? | Primary test image for single-frame detection |
+| 1 | `image_small.jpg`| 1280×720 aerial, contains detectable objects| `image_small_expected.csv`| 1 | Primary test image for single-frame detection|
-| 2 | `image_large.JPG` | 6252×4168 aerial, triggers GSD-based tiling | `image_large_expected.csv` | ? | Coordinates normalized to full image (not tile) |
+| 2 | `image_large.JPG`| 6252×4168 aerial, triggers GSD-based tiling| `image_large_expected.csv`| 3 | Coordinates normalized to full image (not tile)|
-| 3 | `image_dense01.jpg` | 1280×720 dense scene, many clustered objects | `image_dense01_expected.csv` | ? | Used for dedup and max-detection-cap tests |
+| 3 | `image_different_types.jpg`| 900×1600, varied object classes| `image_different_types_expected.csv`| 2 | Must contain multiple distinct class labels|
-| 4 | `image_dense02.jpg` | 1920×1080 dense scene variant | `image_dense02_expected.csv` | ? | Borderline tiling, dedup variant |
+| 4 | `image_empty_scene.jpg`| 1920×1080, no detectable objects| `image_empty_scene_expected.csv`| 0 | CSV has headers only — zero detections expected|
-| 5 | `image_different_types.jpg` | 900×1600, varied object classes | `image_different_types_expected.csv` | ? | Must contain multiple distinct class labels |
-| 6 | `image_empty_scene.jpg` | 1920×1080, no detectable objects | `image_empty_scene_expected.csv` | 0 | CSV has headers only — zero detections expected |
 
 ### Videos
 
 | # | Input File | Description | Expected Result File | Notes |
 |---|------------|-------------|---------------------|-------|
-| 7 | `video_test01.mp4` | Standard test video | `video_test01_expected.csv` | Primary async/SSE/video test. List key-frame detections. |
+| 5 | `video_test01.mp4` | Standard test video | `video_test01_expected.csv` | Primary async/SSE/video test. List key-frame detections. |
-| 8 | `video_1.mp4` | Video variant | `video_1_expected.csv` | Secondary local fixture for resilience and concurrent-style validation. |
+| 6 | `video_1_faststart.mp4` | Faststart video variant | `video_1_faststart_expected.csv` | Streaming compatibility variant. Separate long-video overflow fixture is not currently present in local fixtures. |
-| 9 | `video_1_faststart.mp4` | Faststart video variant | `video_1_faststart_expected.csv` | Streaming compatibility variant. Separate long-video overflow fixture is not currently present in local fixtures. |
 
 ## How to Fill
 

_docs/00_problem/input_data/expected_results/video_1_faststart_expected.csv

@@ -1 +1,81 @@
 time_sec,center_x,center_y,width,height,label,confidence_min
+0,0.289802,0.512695,0.077206,0.072666,ArmorVehicle,0.89
+10,0.753483,0.198423,0.057173,0.070441,ArmorVehicle,0.72
+108,0.621403,0.45839,0.026638,0.037245,Smoke,0.58
+110,0.637573,0.465766,0.027141,0.040499,Smoke,0.58
+112,0.610838,0.480663,0.0283,0.041891,Smoke,0.72
+114,0.60922,0.486313,0.026929,0.042859,Smoke,0.68
+116,0.615395,0.492105,0.028959,0.041361,Smoke,0.67
+118,0.612622,0.49208,0.027159,0.039772,Smoke,0.61
+120,0.488272,0.653743,0.028969,0.045788,Smoke,0.67
+122,0.276008,0.692814,0.061306,0.10955,Smoke,0.74
+127,0.5029,0.24006,0.116369,0.208533,Smoke,0.74
+130,0.350402,0.278989,0.122529,0.179202,Smoke,0.7
+132,0.525031,0.469289,0.103297,0.193989,Smoke,0.48
+134,0.475175,0.52287,0.099367,0.178105,Smoke,0.27
+136,0.042595,0.458038,0.073769,0.201478,Smoke,0.72
+138,0.398975,0.280201,0.132485,0.076752,Smoke,0.3
+14,0.956159,0.68494,0.075992,0.034521,Trenches,0.62
+149,0.115298,0.585385,0.037456,0.030006,Trenches,0.64
+149,0.149621,0.602923,0.039252,0.032263,Trenches,0.73
+149,0.534584,0.082995,0.03861,0.036615,Trenches,0.51
+151,0.363107,0.353294,0.036212,0.030509,Trenches,0.52
+151,0.395121,0.377565,0.03571,0.034257,Trenches,0.52
+153,0.540678,0.549795,0.04493,0.040794,Trenches,0.55
+155,0.124097,0.466852,0.074959,0.051802,Trenches,0.32
+163,0.687639,0.261656,0.116238,0.125195,TyreTracks,0.43
+168,0.378062,0.769232,0.032526,0.027289,Trenches,0.46
+174,0.435846,0.904734,0.03305,0.03245,Trenches,0.27
+176,0.942677,0.589411,0.024974,0.052446,Trenches,0.33
+181,0.612998,0.232146,0.08565,0.060314,Smoke,0.52
+191,0.390291,0.036179,0.054409,0.033613,Trenches,0.39
+193,0.413883,0.208317,0.053158,0.031181,Trenches,0.76
+193,0.430828,0.171384,0.030674,0.029638,Trenches,0.27
+193,0.484856,0.182248,0.04184,0.063353,Trenches,0.65
+193,0.491091,0.316695,0.031371,0.024474,Trenches,0.37
+195,0.556304,0.101112,0.046918,0.036026,Smoke,0.71
+197,0.519331,0.120967,0.044832,0.039692,Smoke,0.67
+199,0.176271,0.118214,0.047294,0.053538,Smoke,0.63
+2,0.653329,0.376178,0.078665,0.081241,ArmorVehicle,0.89
+202,0.171963,0.138687,0.043301,0.04702,Smoke,0.46
+214,0.413986,0.12009,0.050383,0.046863,Smoke,0.5
+219,0.364435,0.552301,0.028486,0.0314,ArmorVehicle,0.72
+240,0.502753,0.535129,0.036657,0.033246,Trenches,0.32
+242,0.019098,0.607661,0.038072,0.024671,Trenches,0.27
+242,0.02406,0.680742,0.044204,0.023859,Trenches,0.37
+242,0.027672,0.720891,0.033973,0.04163,Trenches,0.65
+242,0.041207,0.628713,0.021846,0.044712,Trenches,0.29
+242,0.053195,0.682519,0.027647,0.038862,Trenches,0.65
+242,0.84993,0.631653,0.046269,0.042584,Trenches,0.47
+244,0.017405,0.7017,0.034866,0.023287,Trenches,0.35
+244,0.018653,0.733875,0.03717,0.025044,Trenches,0.48
+244,0.032765,0.621231,0.064777,0.023041,Trenches,0.45
+244,0.05303,0.734179,0.034922,0.041688,Trenches,0.7
+244,0.070526,0.649712,0.023196,0.061101,Trenches,0.44
+244,0.076041,0.69735,0.028326,0.039375,Trenches,0.65
+246,0.025853,0.893559,0.051857,0.02183,Trenches,0.33
+246,0.055409,0.908165,0.020511,0.040618,Trenches,0.32
+252,0.735668,0.309579,0.026179,0.044193,Trenches,0.61
+252,0.760952,0.371567,0.026674,0.062061,Trenches,0.52
+282,0.566367,0.010329,0.036472,0.020469,Smoke,0.49
+284,0.341962,0.011659,0.03561,0.023243,Smoke,0.59
+287,0.423856,0.023553,0.03954,0.044283,Smoke,0.42
+290,0.397077,0.010032,0.037649,0.019987,Smoke,0.29
+292,0.415613,0.011793,0.036875,0.023537,Smoke,0.37
+294,0.385953,0.009416,0.037843,0.018655,Smoke,0.39
+30,0.343423,0.763509,0.098184,0.079214,TyreTracks,0.82
+32,0.438461,0.687812,0.09794,0.07654,TyreTracks,0.85
+4,0.197519,0.416036,0.067097,0.083232,ArmorVehicle,0.83
+41,0.700711,0.590761,0.10644,0.088566,TyreTracks,0.74
+47,0.830646,0.762364,0.108269,0.091804,TyreTracks,0.37
+47,0.698529,0.00825,0.022697,0.01672,Vehicle,0.34
+49,0.349817,0.663337,0.075551,0.061795,TyreTracks,0.77
+6,0.276864,0.60931,0.060284,0.066828,ArmorVehicle,0.8
+67,0.846456,0.527966,0.03052,0.058466,Trenches,0.32
+69,0.68478,0.514788,0.025761,0.053171,Trenches,0.46
+71,0.368416,0.460942,0.030577,0.050027,Trenches,0.78
+73,0.177122,0.455578,0.031931,0.048597,Trenches,0.36
+8,0.379446,0.434878,0.071124,0.077504,ArmorVehicle,0.55
+87,0.128406,0.138269,0.040148,0.032909,Smoke,0.56
+90,0.118965,0.089119,0.037023,0.033037,Smoke,0.32
+95,0.570426,0.373717,0.031503,0.021361,Trenches,0.48

_docs/00_problem/input_data/expected_results/video_test01_expected.csv

@@ -1 +1,5 @@
 time_sec,center_x,center_y,width,height,label,confidence_min
+0,0.289857,0.512138,0.07729,0.072891,ArmorVehicle,0.89
+2,0.653617,0.376064,0.078636,0.08118,ArmorVehicle,0.89
+4,0.197561,0.416208,0.068112,0.084079,ArmorVehicle,0.85
+6,0.27662,0.609538,0.059521,0.067212,ArmorVehicle,0.8

_docs/02_document/modules/ai_config.md

@@ -20,7 +20,7 @@ Data class holding all AI recognition configuration parameters, with factory met
 | `tracking_intersection_threshold` | double | 0.6 | IoU threshold for overlapping detection removal |
 | `model_batch_size` | int | 1 | Batch size for inference |
 | `big_image_tile_overlap_percent` | int | 20 | Tile overlap percentage for large image splitting |
-| `altitude` | double | 400 | Camera altitude in meters |
+| `altitude` | double? | optional | Camera altitude in meters. When missing, GSD-based filtering is disabled |
 | `focal_length` | double | 24 | Camera focal length in mm |
 | `sensor_width` | double | 23.5 | Camera sensor width in mm |
 
@@ -51,7 +51,7 @@ Data class holding all AI recognition configuration parameters, with factory met
 
 ## Configuration
 
-Camera/altitude parameters (`altitude`, `focal_length`, `sensor_width`) are used for ground sampling distance calculation in aerial image processing.
+Camera/altitude parameters (`altitude`, `focal_length`, `sensor_width`) are used for ground sampling distance calculation in aerial image processing. If `altitude` is missing, the service skips GSD-based size filtering and does not tile large images by physical size.
 
 ## External Integrations
 

_docs/02_document/tests/test-data.md

@@ -8,12 +8,9 @@
 | classes-json | `classes.json` (repo root) | 19 detection classes with Id, Name, Color, MaxSizeM | All tests | Volume mount to detections `/app/classes.json` | Container restart |
 | image-small | `input_data/image_small.jpg` | JPEG 1280×720 — below tiling threshold (1920×1920) | FT-P-01..03, 05, 07, 13..15, FT-N-03, 06, NFT-PERF-01..02, NFT-RES-01, 03, NFT-SEC-01, NFT-RES-LIM-01 | Volume mount to consumer `/media/` | N/A (read-only) |
 | image-large | `input_data/image_large.JPG` | JPEG 6252×4168 — above tiling threshold, triggers GSD tiling | FT-P-04, 16, NFT-PERF-03 | Volume mount to consumer `/media/` | N/A (read-only) |
-| image-dense-01 | `input_data/image_dense01.jpg` | JPEG 1280×720 — dense scene with many clustered objects | FT-P-06, NFT-RES-LIM-03 | Volume mount to consumer `/media/` | N/A (read-only) |
-| image-dense-02 | `input_data/image_dense02.jpg` | JPEG 1920×1080 — dense scene variant, borderline tiling | FT-P-06 (variant) | Volume mount to consumer `/media/` | N/A (read-only) |
 | image-different-types | `input_data/image_different_types.jpg` | JPEG 900×1600 — varied object classes for class variant tests | FT-P-13 | Volume mount to consumer `/media/` | N/A (read-only) |
 | image-empty-scene | `input_data/image_empty_scene.jpg` | JPEG 1920×1080 — clean scene with no detectable objects | Edge case (zero detections) | Volume mount to consumer `/media/` | N/A (read-only) |
 | video-test-01 | `input_data/video_test01.mp4` | MP4 video — standard async/SSE/video detection tests | FT-P-08..12, FT-N-04, 07, NFT-PERF-04, NFT-RES-02, NFT-SEC-03 | Volume mount to consumer `/media/` | N/A (read-only) |
-| video-1 | `input_data/video_1.mp4` | MP4 video — local variant for concurrent and resilience-style tests | NFT-RES-02 (variant), NFT-RES-04 | Volume mount to consumer `/media/` | N/A (read-only) |
 | video-1-faststart | `input_data/video_1_faststart.mp4` | MP4 video — faststart/local streaming variant | Streaming compatibility checks | Volume mount to consumer `/media/` | N/A (read-only) |
 | empty-image | Generated at build time | Zero-byte file | FT-N-01 | Generated in e2e/fixtures/ | N/A |
 | corrupt-image | Generated at build time | Random binary garbage (not valid image format) | FT-N-02 | Generated in e2e/fixtures/ | N/A |
@@ -31,12 +28,9 @@ Each test run starts with fresh containers (`docker compose down -v && docker co
 | azaion.onnx | `_docs/00_problem/input_data/azaion.onnx` | YOLO ONNX detection model | All detection tests |
 | image_small.jpg | `_docs/00_problem/input_data/image_small.jpg` | 1280×720 aerial image | Single-frame detection, health, negative, perf tests |
 | image_large.JPG | `_docs/00_problem/input_data/image_large.JPG` | 6252×4168 aerial image | Tiling tests |
-| image_dense01.jpg | `_docs/00_problem/input_data/image_dense01.jpg` | Dense scene 1280×720 | Dedup, detection cap tests |
-| image_dense02.jpg | `_docs/00_problem/input_data/image_dense02.jpg` | Dense scene 1920×1080 | Dedup variant |
 | image_different_types.jpg | `_docs/00_problem/input_data/image_different_types.jpg` | Varied classes 900×1600 | Class variant tests |
 | image_empty_scene.jpg | `_docs/00_problem/input_data/image_empty_scene.jpg` | Empty scene 1920×1080 | Zero-detection edge case |
 | video_test01.mp4 | `_docs/00_problem/input_data/video_test01.mp4` | Standard video | Async, SSE, video, perf tests |
-| video_1.mp4 | `_docs/00_problem/input_data/video_1.mp4` | Video variant | Resilience, concurrent tests |
 | video_1_faststart.mp4 | `_docs/00_problem/input_data/video_1_faststart.mp4` | Faststart video variant | Streaming compatibility checks |
 | classes.json | repo root `classes.json` | 19 detection classes | All tests |
 

_docs/02_tasks/done/AZ-138_test_infrastructure.md

@@ -27,8 +27,6 @@ e2e/
 ├── fixtures/
 │   ├── image_small.jpg          (1280×720 JPEG, aerial, detectable objects)
 │   ├── image_large.JPG          (6252×4168 JPEG, triggers tiling)
-│   ├── image_dense01.jpg        (1280×720 JPEG, dense scene, clustered objects)
-│   ├── image_dense02.jpg        (1920×1080 JPEG, dense scene variant)
 │   ├── image_different_types.jpg (900×1600 JPEG, varied object classes)
 │   ├── image_empty_scene.jpg    (1920×1080 JPEG, no detectable objects)
 │   ├── video_short01.mp4        (short MP4 with moving objects)
@@ -130,8 +128,6 @@ Two Docker Compose profiles:
 | `reset_mocks` | function (autouse) | Calls `POST /mock/reset` on both mocks before each test |
 | `image_small` | session | Reads `image_small.jpg` from `/media/` volume |
 | `image_large` | session | Reads `image_large.JPG` from `/media/` volume |
-| `image_dense` | session | Reads `image_dense01.jpg` from `/media/` volume |
-| `image_dense_02` | session | Reads `image_dense02.jpg` from `/media/` volume |
 | `image_different_types` | session | Reads `image_different_types.jpg` from `/media/` volume |
 | `image_empty_scene` | session | Reads `image_empty_scene.jpg` from `/media/` volume |
 | `video_short_path` | session | Path to `video_short01.mp4` on `/media/` volume |
@@ -150,8 +146,6 @@ Two Docker Compose profiles:
 | classes.json | repo root `classes.json` | JSON (19 objects with Id, Name, Color, MaxSizeM) | All tests (volume mount to detections) |
 | image_small.jpg | `input_data/image_small.jpg` | JPEG 1280×720 | Health, single image, filtering, negative, performance tests |
 | image_large.JPG | `input_data/image_large.JPG` | JPEG 6252×4168 | Tiling tests, performance tests |
-| image_dense01.jpg | `input_data/image_dense01.jpg` | JPEG 1280×720 dense scene | Dedup tests, detection cap tests |
-| image_dense02.jpg | `input_data/image_dense02.jpg` | JPEG 1920×1080 dense scene | Dedup variant |
 | image_different_types.jpg | `input_data/image_different_types.jpg` | JPEG 900×1600 varied classes | Weather mode class variant tests |
 | image_empty_scene.jpg | `input_data/image_empty_scene.jpg` | JPEG 1920×1080 empty | Zero-detection edge case |
 | video_short01.mp4 | `input_data/video_short01.mp4` | MP4 short video | Async, SSE, video processing tests |

e2e/conftest.py

@@ -228,17 +228,6 @@ def image_small():
 def image_large():
     return _read_media("image_large.JPG")
 
-
-@pytest.fixture(scope="session")
-def image_dense():
-    return _read_media("image_dense01.jpg")
-
-
-@pytest.fixture(scope="session")
-def image_dense_02():
-    return _read_media("image_dense02.jpg")
-
-
 @pytest.fixture(scope="session")
 def image_different_types():
     return _read_media("image_different_types.jpg")

src/ai_config.pxd

@@ -12,6 +12,7 @@ cdef class AIRecognitionConfig:
 
     cdef public int model_batch_size
 
+    cdef public bint has_altitude
     cdef public double altitude
     cdef public double focal_length
     cdef public double sensor_width

src/ai_config.pyx

@@ -25,7 +25,8 @@ cdef class AIRecognitionConfig:
 
         self.big_image_tile_overlap_percent = big_image_tile_overlap_percent
 
-        self.altitude = altitude
+        self.has_altitude = altitude is not None
+        self.altitude = 0.0 if altitude is None else float(altitude)
         self.focal_length = focal_length
         self.sensor_width = sensor_width
 
@@ -36,7 +37,7 @@ cdef class AIRecognitionConfig:
                 f'frame_period_recognition : {self.frame_period_recognition}, '
                 f'big_image_tile_overlap_percent: {self.big_image_tile_overlap_percent}, '
                 f'model_batch_size: {self.model_batch_size}, '
-                f'altitude: {self.altitude}, '
+                f'altitude: {self.altitude if self.has_altitude else None}, '
                 f'focal_length: {self.focal_length}, '
                 f'sensor_width: {self.sensor_width}'
                 )
@@ -56,7 +57,7 @@ cdef class AIRecognitionConfig:
 
             data.get("big_image_tile_overlap_percent", 20),
 
-            data.get("altitude", 400),
+            data.get("altitude", None),
             data.get("focal_length", 24),
             data.get("sensor_width", 23.5)
         )

src/annotation.pxd

@@ -7,6 +7,6 @@ cdef class Detection:
 cdef class Annotation:
     cdef public str name
     cdef public str original_media_name
-    cdef long time
+    cdef public long time
     cdef public list[Detection] detections
     cdef public bytes image

src/inference.pyx

@@ -311,13 +311,22 @@ cdef class Inference:
         cdef double ground_sampling_distance
         cdef int model_h, model_w
         cdef int img_h, img_w
+        cdef bint has_gsd
         model_h, model_w = self.engine.get_input_shape()
         img_h, img_w, _ = frame.shape
-        ground_sampling_distance = ai_config.sensor_width * ai_config.altitude / (ai_config.focal_length * img_w)
+        has_gsd = ai_config.has_altitude and ai_config.focal_length > 0 and ai_config.sensor_width > 0 and img_w > 0
-        constants_inf.log(<str>f'ground sampling distance: {ground_sampling_distance}')
+        ground_sampling_distance = 0.0
+        if has_gsd:
+            ground_sampling_distance = ai_config.sensor_width * ai_config.altitude / (ai_config.focal_length * img_w)
+            constants_inf.log(<str>f'ground sampling distance: {ground_sampling_distance}')
+        else:
+            constants_inf.log(<str>'ground sampling distance: skipped (altitude unavailable)')
         if img_h <= 1.5 * model_h and img_w <= 1.5 * model_w:
             all_frame_data.append((frame, original_media_name, f'{original_media_name}_000000', ground_sampling_distance))
         else:
+            if not has_gsd:
+                all_frame_data.append((frame, original_media_name, f'{original_media_name}_000000', ground_sampling_distance))
+                return
             tile_size = int(constants_inf.METERS_IN_TILE / ground_sampling_distance)
             constants_inf.log(<str> f'calc tile size: {tile_size}')
             res = self.split_to_tiles(frame, original_media_name, tile_size, ai_config.big_image_tile_overlap_percent)
@@ -410,6 +419,12 @@ cdef class Inference:
         if annotation.detections:
             constants_inf.log(<str> f'Initial ann: {annotation}')
 
+        if ground_sampling_distance <= 0:
+            if not annotation.detections:
+                return <bint>False
+            constants_inf.log(<str>'Skipping physical-size filtering (ground sampling distance unavailable)')
+            return <bint>True
+
         cdef list[Detection] valid_detections = []
         for det in annotation.detections:
             m_w = det.w * img_w * ground_sampling_distance

src/main.py

@@ -15,6 +15,7 @@ import cv2
 import jwt as pyjwt
 import numpy as np
 import requests as http_requests
+from loguru import logger
 from fastapi import Body, Depends, FastAPI, File, Form, HTTPException, Request, UploadFile
 from fastapi.responses import Response, StreamingResponse
 from fastapi.security import HTTPAuthorizationCredentials, HTTPBearer
@@ -163,7 +164,7 @@ class AIConfigDto(BaseModel):
     tracking_intersection_threshold: float = 0.6
     model_batch_size: int = 8
     big_image_tile_overlap_percent: int = 20
-    altitude: float = 400
+    altitude: Optional[float] = None
     focal_length: float = 24
     sensor_width: float = 23.5
 
@@ -270,7 +271,8 @@ def _post_media_record(payload: dict, bearer: str) -> bool:
             timeout=30,
         )
         return r.status_code in (200, 201)
-    except Exception:
+    except Exception as exc:
+        logger.warning(f"Failed to create media record in annotations service: {exc}")
         return False
 
 
@@ -284,7 +286,8 @@ def _put_media_status(media_id: str, media_status: int, bearer: str) -> bool:
             timeout=30,
         )
         return r.status_code in (200, 204)
-    except Exception:
+    except Exception as exc:
+        logger.warning(f"Failed to update media status in annotations service for {media_id}: {exc}")
         return False
 
 
@@ -332,10 +335,13 @@ def _post_annotation_to_service(token_mgr: TokenManager, media_id: str,
     try:
         token = token_mgr.get_valid_token()
         image_b64 = base64.b64encode(annotation.image).decode() if annotation.image else None
+        total_seconds = int(annotation.time // 1000) if annotation.time else 0
+        hours, remainder = divmod(total_seconds, 3600)
+        minutes, seconds = divmod(remainder, 60)
         payload = {
             "mediaId": media_id,
             "source": 0,
-            "videoTime": f"00:00:{annotation.time // 1000:02d}" if annotation.time else "00:00:00",
+            "videoTime": f"{hours:02d}:{minutes:02d}:{seconds:02d}",
             "detections": [d.model_dump() for d in dtos],
         }
         if image_b64:
@@ -346,8 +352,10 @@ def _post_annotation_to_service(token_mgr: TokenManager, media_id: str,
             headers={"Authorization": f"Bearer {token}"},
             timeout=30,
         )
-    except Exception:
+    except Exception as exc:
-        pass
+        logger.warning(
+            f"Failed to post annotation to annotations service for media {media_id}: {exc}"
+        )
 
 
 def _cleanup_channel(channel_id: str):

tests/test_ai_config_from_dict.py

@@ -5,6 +5,15 @@ def test_ai_config_from_dict_defaults():
     assert cfg.model_batch_size == 8
     assert cfg.frame_period_recognition == 4
     assert cfg.frame_recognition_seconds == 2
+    assert cfg.has_altitude is False
+
+
+def test_ai_config_from_dict_altitude_override_sets_flag():
+    from inference import ai_config_from_dict
+
+    cfg = ai_config_from_dict({"altitude": 400})
+    assert cfg.has_altitude is True
+    assert cfg.altitude == 400
 
 
 def test_ai_config_from_dict_overrides():

tests/test_az174_db_driven_config.py

@@ -118,6 +118,23 @@ def test_resolve_media_for_detect_override_wins():
     assert "paths" not in cfg
 
 
+def test_resolve_media_for_detect_omits_altitude_when_not_provided():
+    # Arrange
+    import main
+
+    tm = main.TokenManager(_access_jwt(), "")
+    mock_ann = MagicMock()
+    mock_ann.fetch_user_ai_settings.return_value = {"probabilityThreshold": 0.2}
+    mock_ann.fetch_media_path.return_value = "/m/v.mp4"
+    with patch("main.annotations_client", mock_ann):
+        # Act
+        cfg, path = main._resolve_media_for_detect("vid-2", tm, None)
+    # Assert
+    assert "altitude" not in cfg
+    assert cfg["probability_threshold"] == 0.2
+    assert path == "/m/v.mp4"
+
+
 def test_resolve_media_for_detect_raises_when_no_media_path():
     # Arrange
     import main