Trinetra AI is an AI-powered CCTV surveillance suite built with Python, OpenCV, YOLO-compatible Ultralytics models, FastAPI, and a React monitoring dashboard.
The working application is in:
SmartVision-AI-Surveillance-Suite/
- Live webcam and RTSP/IP camera support
- OpenCV video capture and frame processing
- YOLO object detection with green and red bounding boxes
- Highway surveillance speed-limit warning demo
- FastAPI REST APIs and WebSocket live streaming
- React dashboard with live camera view, alerts, settings, model controls, analytics, and event panels
- Motion-triggered event recording architecture with pre-event buffering
- Modular surveillance domains for highway, traffic, city security, retail, industrial safety, parking, railway, campus, home, and wildlife use cases
Open PowerShell from the repository root:
cd SmartVision-AI-Surveillance-Suite
copy .env.example .env
python -m venv .venv
.\.venv\Scripts\python.exe -m pip install --upgrade pip
.\.venv\Scripts\python.exe -m pip install -r requirements.txt
.\.venv\Scripts\uvicorn.exe api_gateway.main:app --host 127.0.0.1 --port 8000In another PowerShell terminal:
cd SmartVision-AI-Surveillance-Suite\frontend
npm install
npm run dev -- --host 127.0.0.1Open the dashboard:
http://127.0.0.1:5173
- Select
highway_surveillancefrom the module list. - Use the webcam selector in the Live Video panel to choose your camera.
- Keep
Detection ONenabled to see OpenCV/YOLO annotated frames. - Green boxes are normal detections.
- Red boxes indicate alert-worthy detections, such as speed-limit warnings in highway mode.
- Click Settings to view API, camera, and storage status.
- Click Model Controls to tune demo confidence and edge/recording switches.
- Click Test Alert or Test Event to populate the alert and event panels for presentation.
Full architecture, module layout, API routes, Docker setup, recording design, and extension notes are documented in:
SmartVision-AI-Surveillance-Suite/README.md