Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
OpenCV+YOLO+LLAVA powered video surveillance system
| Date | Stars |
|---|---|
| 2026-07-24 | 793 |
| 2026-07-25 | 793 |
| 2026-07-28 | 793 |
| 2026-07-30 | 793 |
| 2026-08-06 | 794 |
Today
+1 stars today
This week
+1 stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.13%/day
# Machina - Video Surveillance System A simple-to-use AI-powered video surveillance system that can detect and track objects in real-time from cameras, featuring advanced zoom controls, multiple resolution presets, persistent configuration, and smart frame optimization.  ## What does it do? Machina watches video feeds and automatically detects objects like people, cars, and other things you care about. It keeps track of each object and can record video or take screenshots when needed. **Key Features:** - Runs completely offline - no internet required - Records video and takes snapshots with one key press - Real-time object detection and tracking using YOLO models - Advanced zoom and pan controls with smart object tracking - Crowd detection and clustering algorithms - Instant replay system with 10-second rolling buffer - Multiple resolution presets (640x480 to 1920x1080) with keypad shortcuts - Fullscreen support with aspect ratio preservation - Military mode for enhanced detection accuracy - Smart frame skipping and performance optimization - Dynamic YOLO input size adjustment for optimal performance - Semi-transparent selection overlay with real-time drawing - **NEW: Persistent configuration saving** - remembers your settings between sessions - **NEW: Multi-instance webcam support** - automatically detects and configures multiple webcams - **NEW: Command-line stream selection** - specify stream source via --stream argument - **NEW: Audio visualization** - real-time audio waveform display (experimental) - **NEW: First-run detection** - automatic initial configuration ## Quick Start 1. **Install Python 3.8+** and make sure you have a webcam or camera stream 2. **Download and install:** ```bash git clone https://github.com/PsyChip/machina cd machina pip install -r requirements.txt ``` 3. **Torch library should be installed manually:** ```bash pip uninstall torch torchvision torchaudio pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118 ``` 4. **Configure your settings:** Edit the top section of `app.py` to customize: - YOLO model (default: yolo12n) - Video source (RTSP URL or webcam) - Performance settings and thresholds 5. **Run the program:** ```bash # For webcam (default) python app.py # For specific RTSP stream python app.py --stream "rtsp://your-camera-url" # For video file python app.py --stream "path/to/video.mp4" ``` ## How to Use ### Setting up your camera You can specify your video source in multiple ways: **Option 1: Command line argument (recommended)** ```bash python app.py --stream 0 # Webcam python app.py --stream "rtsp://camera-url" # IP Camera python app.py --stream "video.mp4" # Video file ``` **Option 2: Edit `app.py` configuration** - **Webcam:** Set `rtsp_stream = 0` - **IP Camera:** Set `rtsp_stream = "rtsp://your-camera-url"` - **Video File:** Set `rtsp_stream = "path/to/your/video.mp4"` ### Controls | Key | What it does | | ----------- | ----------------------------------------- | | `Q` | Quit | | `R` | Start/Stop recording | | `S` | Take screenshot | | `F` | Reset window to stream resolution | | `M` | Toggle military mode (enhanced detection) | | `Space` | Skip frames | | `Backspace` | Toggle replay mode | | `1` | Resize to 640x480 | | `2` | Resize to 800x600 | | `3` | Resize to 1024x768 | | `4` | Resize to 1280x800 | | `5` | Resize to 1920x1080 | | `6` | Reset to original stream size | | `Enter`
Excerpt of 8,861 characters
Read on GitHub
28
Would you bet a product on this? Bounded 0–100 and slow moving.
matched fp:56194e2fa975a51a, topic:yolo, topic:opencv, readme:object detection