Which way is the best way to detect objects in PhotonVision?

We are currently working on our object detection system in PhotonVision and we have tried it on a note. We’ve tried different methods to detect it - through color and through shape (in photonvision UI - colored shape and object detection).

Which method has your team been using? What are the pros and cons to each one?

Thank you in advance :slight_smile:

Some teams are using an Orange Pi or other RK3588/RK3588S powered coprocessor to use PhotonVision’s AI object detection support.

We are working on bringing AI inference support for more generic hardware to PhotonVision using a Python shim script and gRPC.