AI-Based On-Board Maritime Object Detection for Earth Observation Payload Data Reduction on Versal Embedded Hardware
The study developed an AI system for satellites to automatically select which maritime objects are downlinked based on onboard detection results. The system processes satellite imagery through correction, tiling, and vessel detection using a quantized model on embedded hardware. It demonstrates that different data reduction modes (like metadata only or image crops) absorb detection errors differently, allowing the user to judge data volume reduction based on actual error trade-offs.