AI Video Surveillance for Smart Agriculture
Real-time intrusion detection with YOLOv8 and DETR, OpenCV stream processing and MQTT messaging; inference optimized for edge deployment.
The pipeline processes video streams with OpenCV and uses YOLOv8 and DETR for intrusion detection. MQTT carries messages between the detection pipeline and connected components. Inference optimization targets edge deployment, where compute resources are limited.
Technologies
Python, YOLOv8, DETR, OpenCV, MQTT
How it works
- Video input
- YOLOv8 / DETR detection
- MQTT alerts
- Edge deployment
