openmv.io
OpenMV boards run MicroPython firmware for on-camera machine vision — no host PC required for many pipelines.
Source
What you can run on-device
- Face detection (BlazeFace), hand landmarks, YOLOv8 person tracking
- AprilTag pose, QR and 1D barcodes, color blob tracking
- Custom TFLite models (e.g. from Roboflow)
Quick start
- Install OpenMV IDE (Windows/macOS/Linux)
- Connect the cam over USB (blue heartbeat LED when ready)
- Connect in IDE → run example scripts from the docs
Core libraries
| Module | Role |
|---|---|
csi | Camera sensor config |
image | Blobs, features, drawing |
ml | TFLite inference + postprocessors |
machine | GPIO, I²C, SPI, UART |
network / microdot | WiFi and tiny HTTP server |
Board-specific pinouts: N6, AE3, H7, RT1062, Nicla Vision, Portenta, etc.