Stereo digital image correlation (3D-DIC) for non-contact measurement and refinement of delta robot arm displacement and jerk

This study proposes using the stereo digital image correlation (3D-DIC) method to measure the delta robot arm displacement and jerk to verify the robot arm end-effector position and improve the absolute accuracy. The combination of the cameras and robotic arm is eye-to-hand. The study's contribution is that the camera is a sensor for non-contact and full-field measurement. The non-contact sensor avoids changing the object's motion state due to contact. The two economic cameras used in this study cost about 500 USD. We obtain the external and internal parameters between the two cameras by calibrating the camera pinhole model and the chessboard pattern. A spherical feature is set up on the delta robotic arm end-effector. This study used the Hough circle transform and 3D-DIC to trace the spherical center. The end-effector displacement is compared by 3D-DIC and the laser displacement meter. The acceleration and jerk are compared by 3D-DIC and the accelerometer. This study ultimately verified that the displacement error is less than 2%. The peak values of acceleration and jerk are close to the results measured by the accelerometer. Applying this technique to 3D-printed products can reduce defects caused by excessive speed.