handical

Hand-eye extrinsic calibration of multiple fixed cameras and a robot arm.

A chessboard is mounted on the robot end-effector (ee). We move the ee around to different locations within the cameras view and assume that the ee pose in the robot's base frame is known via forward kinematic. (Future versions can address noises on those ee poses.) This script takes corner point measurements and compute the camera poses wrt the robot's base and also the chessboard's pose wrt the ee.

Installation

• Dependencies:

  • OpenCV2 (with python support)

        pip install opencv-python
    

  • PyYAML

        pip install pyyaml
    

  • gtsam: (tested with tag 4.0.2)
    • Download:

       git clone [email protected]:borglab/gtsam.git
       git checkout 4.0.2 -b 4.0.2
    

    • Build and install: you need to build gtsam with cython wrapper support. Please follow the instructions here for the dependencies and which compile flag to set.

• Build:

  • First, make sure your PATH env contains gtsam's install/lib folder. This is necessary for our cmake script to be able to find gtsam's cmake scripts.
  • Also, make sure your PYTHONPATH contains gtsam's Cython install path (normally in CMAKE_INSTALL_PREFIX/cython when you build gtsam). This is so that it can find gtsam Cython header: gtsam/gtsam.pxd.
  • Then just build and install the project using cmake.