PhotonVision — FRC Computer Vision
Website: https://photonvision.org/ Purpose: Free, fast, easy computer vision for FRC
What It Is
PhotonVision is the largest FOSS FRC vision project. It runs on a coprocessor (Raspberry Pi 3/4, OrangePi 3, etc.) and handles:
- AprilTag detection and pose estimation
- Object detection (ML-based)
- Multi-camera support
- Driver mode integration
Why We Use It
Team 2890 runs PhotonVision on Mothman for AprilTag detection. It’s our primary vision system.
Key Capabilities
| Feature | Description |
|---|---|
| AprilTag Tracking | Out of the box FRC target support |
| Multi-Tag Pose | Fuse all available data for peak accuracy |
| Camera Calibration | Per-camera intrinsics maximize homography accuracy |
| ML Object Detection | Hardware-accelerated inferencing for gamepieces |
| Driver Mode | Same camera for driving and vision |
| Multi-Camera | Run as many cameras as hardware supports |
Getting Started
- Download PhotonVision image for your coprocessor
- Flash to SD card
- Connect to network — start tracking in minutes
- Configure via web dashboard
Connection to YAGSL
PhotonVision works alongside YAGSL (swerve). YAGSL handles drive control, PhotonVision handles vision targets.
2026 Champs Talk
PhotonVision 2026 presentation covers:
- Rubik Pi 3 hardware
- Object Detection with PhotonVision
- Slide deck coming soon
In The Curriculum
This topic belongs in:
- Programming Track → Level 2 (after basic Java/WPILib)
- Vision Module (Apriltag tracking, camera setup)