sop-violation-visual-recognition-with-multi-camera-system

This Project was in 6 phases

The first phase deals with live streaming the feed from a triple camera setup.

The second phase deals with mask detection using a trained model. The model was trained using yolov5

The third phase deals with orthogonal top view projection on a top view

The fourth phase deals with projecting dot in the generated top view in place of the bounding box for "mask" and "no_mask"

The fifth phase deals with detecting sop violations

the sixth phase deals with generating three different heatmaps:

• a static heatmap for all detections.
• an animated heatmap for all detections.
• a heatmap for sop violations