- Refactored camera handling to support multiple cameras via environment variables.
- Added API endpoints for camera selection and retrieval of available cameras.
- Updated stream processing to dynamically switch between selected cameras.
- Integrated device selection for YOLO inference to utilize CUDA if available.
- Implemented a new camera_stream.py file for real-time video processing.
- Integrated YOLO model for detecting people in the video stream.
- Added functionality for grouping detected individuals based on proximity.
- Implemented alert system for detecting groups over a specified duration.
- Included state management for stream status and alerts.
- Utilized Flask for serving video stream and alert information.
- Configured environment variables for RTSP stream access.