ThreatDetector
AI ToolA real-time AI-powered surveillance system that detects weapons, tracks individuals, and triggers dynamic recording — turning passive camera feeds into an active security layer.
Overview
ThreatDetector processes live camera feeds through a YOLO-based object detection pipeline optimized for weapon identification. When a potential threat is detected, the system uses ByteTrack to maintain persistent identity tracking across frames, associates threats with specific individuals, and triggers dynamic recording with configurable pre- and post-event buffers.
The system supports multiple simultaneous camera feeds, provides a real-time web dashboard for monitoring, and maintains a searchable event history with full audit trails. It's designed to augment human security teams — not replace them — by ensuring threats are never missed in the feed.
Key Features
- YOLO11n weapon detection (optimized via ONNX)
- ByteTrack persistent person tracking
- Threat-to-person association
- Dynamic recording with event buffers
- Multi-camera simultaneous processing
- Real-time web monitoring dashboard
- Searchable event history
- User authentication and access control
Tech Stack
Architecture Highlights
Need AI-powered video analytics?
I build custom computer vision systems that turn camera feeds into actionable intelligence.
Start a Conversation