Online Eye-Hand Calibration for Robotic Arms

Worked for an early stage startup ‘Voaige’ to come up with an Online eye-hand calibration methodology suitable for their application and wrote the software for the same.

Robotic Pick and Place applications requires the robot to know the relative pose of the camera with respect to itself and thereby map the position and orientation of the objects from the image to the real world. This calibration process is aptly called ‘Eye-in-Hand Calibration’ (since the camera is mounted at the end-effector of the robotic arm). 

The conventional method of camera calibration requires that we know the position of key-points in the global frame of reference and use this information to calculate the camera position. This method is tedious and can prove to be severely restrictive (requiring a template with points mapped in the global frame and manual guidance for the robotic arm to reach each individual point). This method can be replaced with an online calibration process where the robot identifies a set of key points in the camera image, consequently carries out a set of predetermined motions and finally identifies the same set of key points after each motion and recalculates camera relative position using this information.