python package for working with RGB(D) cameras in the context of robotic manipulation.
And has following features:
- convenient access to the camera output by wrapping the API (currently only supports the ZED2i Camera)
- camera-agnostic code for detecting fiducial markers ( for extrinsics calibration or object localisation)
- camera-agnostic code for (re)projection of points back to the camera frame or the world frame, using prior knowledge on the objects; multi-view (TODO) or depth maps.
Using the Zed cameras requires installation of the SDK and the python wrapper.
See https://www.stereolabs.com/docs/app-development/python/install/#getting-started for more info.
If you already have the SDK, run the get_python_api script. If you don't have the SDK already, install the SDK which will also install the python API (if you choose to).
eef_pos_in_robot = position of the end-effector in the frame of the robot base joint. This corresponds to the transform FROM robot base TO end-effector.
marker_pos_in_cam = position of the aruco calibration pattern in the frame of the camera. This corresponds to the transform FROM the camera TO the calibration pattern.
| Abbreviation | Meaning |
|---|---|
| eef | End-effector |
| rotmat | Rotation matrix |
| pos | Position |
| rot | Rotation |
| hommat | Homogeneous matrix |