Revert the Vision subsystem & robot electronics to the baseline functionality we had during the season. Then, update the camera positions, drivetrain pose estimator, etc. until the pose estimates are accurate enough for us to use them.
PhotonPoseEstimator has a lot of PoseStrategies; maybe look into those?
I know Nomad (Build Blog Post) and the Patribots used vision this year if you need examples (though there are definitely more teams you can reference). I highly suggest that whoever does this issue gets familiar with the PhotonVision documentation and reads a few threads on Chief Delphi about this topic.
Read this post, since it talks about AprilTag changes for Crescendo.
Revert the Vision subsystem & robot electronics to the baseline functionality we had during the season. Then, update the camera positions, drivetrain pose estimator, etc. until the pose estimates are accurate enough for us to use them.
PhotonPoseEstimator has a lot of PoseStrategies; maybe look into those? I know Nomad (Build Blog Post) and the Patribots used vision this year if you need examples (though there are definitely more teams you can reference). I highly suggest that whoever does this issue gets familiar with the PhotonVision documentation and reads a few threads on Chief Delphi about this topic.
Read this post, since it talks about AprilTag changes for Crescendo.