GitHub - alebal123bal/khadas_yolov8n_multithread: Real-time YOLOv8n UAV detection at the sensors 46 FPS ceiling, in ~140 MB of RAM

▤ Summary

This project delivers a real-time UAV detection pipeline built around YOLOv8n on Rockchip RK3588S hardware. It uses hardware-accelerated capture, resize, inference, and display to push throughput to the camera ceiling while keeping memory usage low. The system splits detection, tracking, temporal analysis, and LLM summarization into separate Unix-domain-socket-connected processes. It also includes build, cross-compile, and deployment instructions for running on RK3588S boards such as the Khadas Edge2. The repository targets educational and research use rather than production deployment.

▦ Linked Companies