Design On Control System Of Parallel Robot Based On Visual Processing

In industrial enviroment,the location of workpiece often need to be adjusted.If this monotonous and repetitive job is done by normal workers,it will be inefficient and lead to exhausted.However,vision robot can operate rapidly and accurately for this kind of working situation.Taking B&R parallel robot system as the study subject,the operation of vision robot's sorting process was simplified as the demo of picking and placing of geometries which using parallel robot.This study subject designed image processing program and path planning algorithm which including digital image acquisition,digital image processing,angle calculation with route planning and TCP/IP communication.Installing the camera at the top side of inner wall of emergency door.Getting the image of workbench through image capture software and saved the image on computer.The image needed pre-process such as image enhancement and projection transformation to acquire a better view.The difference between the geometries of outer ring and the holes of inner ring was discovered through segmentation.A shape recognition method was created to define the shape about the geometries of outer ring and the holes of inner ring and then calculated the coordinates of centroid.A color recognition method was created to define the color of the geometries of outer ring.Using corner detection algorithm to obtain corner coordinates of the geometries of outer ring and calculated the angle needed to adjust when the geometries of outer ring were moving to the holes of inner ring.The sequence of geometries during picking and placing process was based on ant colony algorithm in order to achieve a shortest circuit.After data resetting,the result of angle calculation with route planning will become to TCP frame.Building TCP/IP communication software on computer and sent TCP frame to B&R PCC.When VNC Viewer received packet data successfully,it will send starting signal to parallel robot.The robot then picked up the geometries on the outer ring and placed them into the holes in the inner ring to accomplish the moving of geometries.All programs and algorithms were tested on B&R parallel robot system.From the tests results,the accuracy rate of color recognition was up to 100% and the accuracy rate of shape recognition was as least 94%,meanwhile,the precision and rapidity of placement was satisfactory.