Experimental Path Tracking of a Cartesian Robot Using Image Processing

This study examines the implementation of image processing to track different paths for a Cartesian robot. Images were captured using a calibrated mobile phone camera and imported to MATLAB. These images include path images and a check-board image from the same position. The check-board image is used to extract the rotation and translation matrices and to calibrate the camera. In MATLAB, images were processed using several algorithms to obtain a binary image of the path. Image coordinates of the course were converted to the actual co-ordinates using the translation and rotation matrices mentioned. These coordinates are reduced and sorted to make a path and then sent to the robot Microcontroller. The Microcontroller moves the robot end-effector according to the coordinates to track the path. The results section investigated image processing accuracy and the robot’s path tracking for three linear, sinusoidal, and circular courses. The results show that the image processing errors reach a maximum of 1mm, and the robot tracking errors reach a maximum of 3mm; considering the conditions, the results are acceptable.