Files
geointel/docs/DEMO_SCENARIOS.md
T
Codex 6ea3586a3e
GeoIntel CI / docs-smoke (push) Has been cancelled
GeoIntel CI / contract-smoke (push) Has been cancelled
Initial GeoIntel V1 foundation
2026-06-16 23:36:32 +02:00

88 lines
2.2 KiB
Markdown

# GeoIntel Kempen — Demo Scenarios v1.0
Demo scenarios guide development toward portfolio-ready workflows.
## Demo 1 — Building Detection + GRB QA/QC
### Goal
Show the complete GeoAI loop: raster imagery → AI detections → geospatial output → official reference validation.
### Workflow
1. Create project: `Geel Building Detection QA`.
2. Select or draw area in Geel.
3. Upload a georeferenced aerial/orthophoto raster.
4. Fetch GRB buildings for the area or use OSM fallback.
5. Run object detection for buildings.
6. Convert detections to geospatial polygons.
7. Run QA/QC against GRB buildings.
8. Show precision, recall, F1, mean IoU.
9. Display TP/FP/FN on map.
10. Export detections and QA findings as GeoJSON/CSV.
### Recruiter value
Demonstrates Rasterio, tiling, YOLO/PyTorch, GeoPandas/Shapely, PostGIS, QA metrics and official geodata validation.
## Demo 2 — Vegetation Segmentation
### Goal
Show segmentation/mask-to-vector workflow.
### Workflow
1. Select area around Mol or Kasterlee.
2. Upload raster or use sample imagery.
3. Run segmentation or NDVI-based threshold.
4. Polygonize vegetation masks.
5. Calculate green area and fragmentation.
6. Export vegetation polygons.
## Demo 3 — Vector Change Detection
### Goal
Show change analysis before full raster AI change detection is implemented.
### Workflow
1. Select two building/reference layers for the same area.
2. Run vector change detection.
3. Classify added, removed and modified features.
4. Show change statistics and map overlays.
## Demo 4 — Raster Workbench
### Goal
Show professional raster handling.
### Workflow
1. Upload GeoTIFF.
2. Inspect metadata.
3. Clip to area.
4. Generate preview.
5. Generate tiles.
6. Export clipped raster.
## Demo 5 — Vector Workbench
### Goal
Show core GIS operations.
### Workflow
1. Upload GeoJSON/GPKG.
2. Inspect metadata.
3. Clip to area.
4. Buffer roads.
5. Intersect buffers with green/water/building layers.
6. Export result.
## Demo readiness checklist
- [ ] Demo datasets present or documented.
- [ ] No fake results.
- [ ] Every result is exportable.
- [ ] Every map layer is inspectable.
- [ ] Every long operation has job status.
- [ ] QA output can be explained in one minute.