Persist raster tile CRS metadata

CHANGELOG.md

@@ -7,12 +7,13 @@
 
 # Changelog
 
-## Sprint 123 YOLO class normalization and real-data inference fix (2026-07-07)
+## Sprint 123 YOLO class and tile CRS normalization (2026-07-07)
 
 - Fixed configured-YOLO class filtering so model labels such as `Building` match operator/domain filters such as `building`.
 - Persisted configured-YOLO class names as canonical lowercase values while preserving the original model label in detection provenance.
 - Added regression coverage for the mixed-case YOLO class route that caused the Geel real-data smoke to persist zero detections.
 - Confirmed through direct Tower inference that the active local building model returns raw detections on the prepared Geel orthophoto tile; the remaining work is threshold/QA calibration rather than model availability.
+- Fixed raster tile manifest CRS propagation so generated tile manifests include source CRS metadata required to convert YOLO pixel boxes to WGS84 Detection GeoJSON coordinates.
 
 ## Sprint 122 Real operator data availability and raster metadata fix (2026-07-07)
 

backend/README.md

@@ -369,7 +369,9 @@ detections against persisted `vector_features`, persists QA/QC rows and exports
 the detection GeoJSON. It never seeds demo detections, enables fixture mode,
 fetches live providers or downloads model weights. Configured-YOLO model class
 labels are normalized to lowercase for filtering and persisted detections while
-the original model label is retained in detection provenance. Current V1 upload
+the original model label is retained in detection provenance. Raster tile
+manifests generated for AI handoff include source CRS metadata so pixel-space
+model outputs can be transformed to WGS84 GeoJSON coordinates. Current V1 upload
 support is limited to GeoTIFF-style rasters and GeoJSON/JSON reference vectors.
 
 ### Run backend

backend/app/services/raster_operations_service.py

@@ -939,8 +939,13 @@ class RasterOperationsService:
 
         manifest_tiles: list[dict[str, Any]] = []
         tile_paths: list[str] = []
+        source_crs: str | None = None
 
         with rasterio.open(dataset.storage_path) as source:
+            raw_source_crs = getattr(source, "crs", None)
+            source_crs = raw_source_crs.to_string() if hasattr(raw_source_crs, "to_string") else (
+                str(raw_source_crs) if raw_source_crs else dataset.crs
+            )
             source_count = getattr(source, "count", 0)
             if not source_count:
                 source_count = 1
@@ -981,6 +986,7 @@ class RasterOperationsService:
                             "pixel_window": [int(xoff), int(yoff), int(tile_width), int(tile_height)],
                             "bounds": RasterOperationsService._window_bounds_to_list(bounds),
                             "transform": [float(item) for item in transform.to_gdal()],
+                            "crs": source_crs,
                             "index": tile_index,
                         },
                     )
@@ -994,10 +1000,14 @@ class RasterOperationsService:
         except AppError:
             source_metadata = {"bounds": [0.0, 0.0, 0.0, 0.0]}
         bounds = source_metadata.get("bounds", [0.0, 0.0, 0.0, 0.0])
+        manifest_crs = source_crs or source_metadata.get("crs") or dataset.crs
         manifest_payload = {
             "tile_set_id": tile_set_id,
             "source_dataset_id": str(dataset.id),
             "source_raster_id": str(dataset.id),
+            "crs": manifest_crs,
+            "source_crs": manifest_crs,
+            "dataset_crs": dataset.crs,
             "bounds": [float(value) for value in bounds],
             "tile_size": int(tile_size),
             "overlap": int(overlap),

backend/tests/test_raster_operations_service.py

@@ -606,6 +606,10 @@ def test_raster_tile_returns_manifest_payload(monkeypatch, tmp_path) -> None:
                 float(window.yoff + window.height),
             )
 
+    class FakeCRS:
+        def to_string(self):
+            return "EPSG:31370"
+
     class FakeSource:
         width = 10
         height = 10
@@ -614,6 +618,7 @@ def test_raster_tile_returns_manifest_payload(monkeypatch, tmp_path) -> None:
             self.profile = {"width": self.width, "height": self.height, "count": 1, "dtype": "uint8", "transform": None}
             self.transform = None
             self.nodata = 0
+            self.crs = FakeCRS()
 
         def read(self, *args, **kwargs):
             return FakeArray()
@@ -670,7 +675,10 @@ def test_raster_tile_returns_manifest_payload(monkeypatch, tmp_path) -> None:
     assert payload["manifest"]["overlap"] == 1
     assert payload["manifest"]["source_dataset_id"] == str(dataset_id)
     assert payload["manifest"]["source_raster_id"] == str(dataset_id)
+    assert payload["manifest"]["crs"] == "EPSG:31370"
+    assert payload["manifest"]["source_crs"] == "EPSG:31370"
     assert payload["manifest"]["count"] == payload["count"]
+    assert payload["manifest"]["tiles"][0]["crs"] == "EPSG:31370"
     assert payload["manifest"]["tiles"][0]["bounds"] == [0.0, 0.0, 4.0, 4.0]
     assert payload["manifest"]["ai_inference"] is False
     assert payload["manifest"]["tile_server"] is None

docs/AI_PIPELINES.md

@@ -187,7 +187,7 @@ Elke tile moet opslaan:
 - pixel window
 - geospatial bounds
 - transform
-- CRS
+- CRS, and the manifest must also carry source CRS metadata
 - tile size
 - overlap
 

docs/CODEX_EXECUTION_LOG.md

@@ -1,24 +1,42 @@
-## Sprint 123 YOLO class normalization and real-data inference fix (2026-07-07)
+## Sprint 123 YOLO class and tile CRS normalization (2026-07-07)
 
 Changed:
 - Investigated the Geel real-data smoke that persisted zero detections despite the configured building model being available.
 - Confirmed on Tower that `/app/models/yolov8n-building-segmentation.pt` reports model class `Building` and returns 4 raw detections at confidence `0.5` on the same real Geel tile manifest.
 - Fixed configured-YOLO detection persistence so model class names are compared case-insensitively against `class_filter`, persisted as canonical lowercase domain classes, and preserve the original model class name in `properties_json.model_class_name`.
-- Added regression coverage in `backend/tests/test_sprint8b_yolo_foundation.py`.
+- Found a second live GIS correctness issue: generated tile manifests carried Lambert bounds/transforms but no CRS, so detection GeoJSON could expose EPSG:31370 coordinates as if they were EPSG:4326.
+- Fixed raster tile manifest generation to include `crs`, `source_crs` and `dataset_crs` on the manifest and `crs` on each tile entry when the source raster CRS is known.
+- Added regression coverage in `backend/tests/test_sprint8b_yolo_foundation.py` and `backend/tests/test_raster_operations_service.py`.
 
 Validation:
 - RED: `python -m pytest backend/tests/test_sprint8b_yolo_foundation.py::test_yolo_class_filter_is_case_insensitive_and_persists_canonical_class -q` failed with `detection_count=0` because `Building` did not match `building`.
 - `python -m pytest backend/tests/test_sprint8b_yolo_foundation.py::test_yolo_class_filter_is_case_insensitive_and_persists_canonical_class -q` passed.
 - `python -m pytest backend/tests/test_sprint8b_yolo_foundation.py backend/tests/test_model_asset_catalog.py backend/tests/test_sprint121_real_data_detection_qa_smoke.py backend/tests/test_sprint122_raster_upload_metadata_mapping.py -q` passed: 20 tests.
+- RED: `python -m pytest backend/tests/test_raster_operations_service.py::test_raster_tile_returns_manifest_payload -q` failed because the tile manifest had no `crs`.
+- `python -m pytest backend/tests/test_raster_operations_service.py::test_raster_tile_returns_manifest_payload -q` passed.
+- `python -m compileall backend/app` passed.
+- `bash scripts/run_readiness_check.sh` passed: 387 backend tests, Alembic head check, frontend typecheck/build and shell syntax checks.
+- Tower deploy from commit `71c2cd9` passed with `GEOINTEL_INSTALL_AI=true`.
+- Deploy-time live migration smoke passed; PostGIS reported `3.6 USE_GEOS=1 USE_PROJ=1 USE_STATS=1` and Alembic head was `202606120900`.
+- Deploy-time browser runtime verification passed for `http://192.168.10.150:1202`.
+- Real-data smoke after the class-normalization deploy passed and persisted 4 detections:
+  - project: `cb80638d-dbef-48ac-b19c-cec7c3efc96e`
+  - raster dataset: `ae0ff76d-70c0-404f-b777-54d14517179a`
+  - reference dataset: `8ac01b4f-bd6a-4d6a-b625-a0950ae0f3eb`
+  - analysis run: `7ba34274-411d-45e3-8f54-c37baec598b1`
+  - quality check: `66907e6a-9ed7-4b0c-976f-5ad1ba9b8b7a`
+  - detection export: `516d37a3-2305-48a2-a3dd-b56f70eb055e`
+  - persisted detections used canonical `class_name=building` and preserved `model_class_name=Building`.
 
 Open:
-- Full readiness, Tower deploy and repeated live real-data smoke still need to be run for this pass.
+- Redeploy the tile-CRS manifest fix and rerun the real-data smoke to verify Detection GeoJSON coordinates are in EPSG:4326.
 
 Limitations:
-- This fixes class routing and persistence, not model quality. Thresholds, false positives and reference IoU quality still need calibration on larger local orthophoto samples.
+- This fixes class routing, persistence and future tile manifest CRS propagation. Existing tile manifests generated before this fix remain missing CRS and should be regenerated before AI runs.
+- Thresholds, false positives and reference IoU quality still need calibration on larger local orthophoto samples.
 
 Next recommended pass:
-- Redeploy to Tower, rerun the Geel real-data smoke, then inspect persisted detections and QA metrics to choose practical confidence/IoU defaults.
+- Redeploy the tile-CRS fix, rerun the Geel real-data smoke, confirm WGS84 Detection GeoJSON coordinates and then calibrate practical confidence/IoU defaults.
 
 ## Sprint 122 Real operator data availability and raster metadata fix (2026-07-07)
 

docs/TODO.md

@@ -94,6 +94,7 @@ This file now starts with the current implementation status. Older preparation/b
 - [x] Add operator-provided real raster/reference detection + QA workflow smoke.
 - [x] Validate the configured building model on a real georeferenced Kempen orthophoto/GeoTIFF with persisted reference vectors and QA/QC metrics.
 - [x] Fix configured-YOLO mixed-case class labels so `Building` model output matches `building` domain filters.
+- [x] Persist CRS metadata in raster tile manifests so AI detections can be transformed to WGS84 GeoJSON correctly.
 - [ ] Calibrate confidence, IoU and model selection against persisted Geel detections and additional local orthophoto/reference samples.
 
 ## Sprint 8 status

scripts/README.md

@@ -182,7 +182,9 @@ against persisted `vector_features`, and exports the detection run as GeoJSON.
 It does not seed demo data, enable fixture detections, fetch external data or
 download model weights. Configured-YOLO model class labels are normalized to
 lowercase for filtering and persisted detections, while the original model label
-is retained in detection provenance. A zero detection count is accepted
+is retained in detection provenance. Raster tile manifests generated by the
+workflow include source CRS metadata so persisted detection GeoJSON coordinates
+can be transformed to WGS84. A zero detection count is accepted
 operationally only when the selected model genuinely returns no usable
 detections after class filtering; it must be interpreted as model/data quality
 evidence rather than as a successful building extraction result.