Anti-Poaching

Detecting unauthorized vehicle activity in protected wildlife reserves, especially at night when poaching typically occurs.

Sensor: SAR (works through clouds and at night — critical for nocturnal poaching activity).

Pipeline:

• Collect: SAR strip over reserve boundary and access roads.
• Map: change detection against a baseline backscatter image. Flag new high-reflectivity points on known access routes and within the reserve perimeter (vehicles are strong SAR reflectors due to corner reflector geometry).
• Reduce: filter detections by time of day (nighttime = elevated suspicion), location (inside reserve boundary or on unauthorized access routes), and cluster size. Generate alert with coordinates, estimated vehicle count, and heading from track direction.

Alert payload: ~2 KB (detection coordinates, time, number of vehicles, heading estimate, confidence score).

Feasibility:

• Vehicle detection in SAR is well-established — vehicles are bright point targets against natural backgrounds.
• Person-scale detection is not feasible from LEO (requires < 5 m GSD; typical SAR is 10--20 m). The approach detects vehicles, not individuals.
• Nighttime filtering significantly reduces false positives from legitimate daytime traffic (rangers, tourists).
• Repeat-pass change detection catches new tracks in soft terrain (sand, mud) even after vehicles have left.
• Revisit frequency (90 min orbit period) limits real-time tracking but suffices for alerting rangers to investigate.