Desing And Implementation On Vision-Based Ball Location Control Using Industrial Robot

The machine vision applications in industry is more and more widely, From this background, we designed a visual sensing system used in location sensing based on ABB’s industrial robot, the system consists of a platform mounted on ABB robotic arm, camera, PC, robotic actuator, serial communications and related software. The ball is located on the platform, which is captured by the video and adjusted by ABB robotic arm to control the movement of the platform so that the ball can move to a predetermined position. The thesis is divided into three parts:get the ball position with image processing technology; design the control algorithms; accomplish control of the position of the ball with ABB robotic arm.In the presence of external light interference conditions to sample the target with a CMOS camera, take the image obtained gray, denoising, binarization, morphological processing, edge extraction and other operations, Hough transform algorithm used to identify the position of the ball center.After obtaining the position of the center of the ball, use the deviation of a predetermined position based on the current position, the P control algorithm is called for each manipulator joint movement.Each joint manipulator control command is sent to the ABB robot controller via RS-232. Motion control programs written in the RAPID language environment, real-time serial data is read to call the RAPID procedure to adjust the rotating or moving the joint, the control platform enables the ball reaches the predetermined position under the action of gravity.We built experimental hardware platform, then using MATLAB programming realize all algorithms in WIN7environment. At last the experiments prove that the design of the hardware platform is reasonable, the algorithm is effective.