detect-objects

You are helping the user run AI object detection on geospatial imagery using geoai.

Input: $@

Follow these steps in order.

Step 1 -- Parse arguments

Extract:

• $0 as the model name: buildings, cars, ships, solar-panels, parking-lots, agriculture, or grounded-sam
• $1 as the input raster path
• --text PROMPT for GroundedSAM text-prompted segmentation (required when model is grounded-sam)
• --output FILE for the output vector file (default: ./<model>_detections.gpkg)

If the model name is not recognized, list the available models and ask the user to pick one.

Model mapping:

Argument	GeoAI Class
buildings	geoai.BuildingFootprintExtractor
cars	geoai.CarDetector
ships	geoai.ShipDetector
solar-panels	geoai.SolarPanelDetector
parking-lots	geoai.ParkingSplotDetector
agriculture	geoai.AgricultureFieldDelineator
grounded-sam	geoai.GroundedSAM

Step 2 -- Check GPU availability

python3 -c "
import torch
if torch.cuda.is_available():
    print(f'GPU: {torch.cuda.get_device_name(0)}')
    print(f'CUDA: {torch.version.cuda}')
    print(f'Memory: {torch.cuda.get_device_properties(0).total_mem / 1e9:.1f} GB')
else:
    print('GPU: not available (CPU mode)')
    print('Warning: inference will be significantly slower without a GPU')
"

If no GPU is available, warn the user but continue.

Step 3 -- Resolve the input file

If $1 looks like an absolute path, use it directly. Otherwise:

find "$PWD" -name "$1" -not -path '*/.git/*' 2>/dev/null

If no file specified and state exists, check for recently inspected/downloaded files:

STATE_DIR=""
test -f .geoai-skills/state.json && STATE_DIR=".geoai-skills"
PROJECT_ROOT="$(git rev-parse --show-toplevel 2>/dev/null || echo "$PWD")"
PROJECT_ID="$(echo "$PROJECT_ROOT" | tr '/' '-')"
test -f "$HOME/.geoai-skills/$PROJECT_ID/state.json" && STATE_DIR="$HOME/.geoai-skills/$PROJECT_ID"