You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Being able to use multiple cameras to detect apriltags would make early match localization quicker/easier and would allow fusing of apriltags throughout the entire match when neither of the 2 zeds are facing towards the base. We should be able to use one of our plain webcams to do this since apriltags don't require depth cameras
The text was updated successfully, but these errors were encountered:
Im pretty sure nvidia has support for this with isaac ros 3.1 (they added multi cam support for nvblox), but if not we could frankenstein it and stitch multiple camera feeds into one with OpenCV or something then feed that to nvblox. Determining the pixel that the tag is found in we could replicate apriltag detection but this is not ideal
Jonathan's comment is the way we would want this to be done. Run one instance of isaac_ros_apriltag per camera we want searching for apriltags, and since each of these instances return TFs from the camera frame to the tag frame, it should just straight up work honestly with the Apriltag math we are already doing since I wrote it to rely on the ROS TF tree. Although, some minor modifications might be needed.
Being able to use multiple cameras to detect apriltags would make early match localization quicker/easier and would allow fusing of apriltags throughout the entire match when neither of the 2 zeds are facing towards the base. We should be able to use one of our plain webcams to do this since apriltags don't require depth cameras
The text was updated successfully, but these errors were encountered: