Research On Motion Control Of Robot Based On Machine Vision

After more than50years of development, the robot has already been widelyapplied in every industry fields and become the important standard of industrymodernization. However, most industrial robots work in the “teaching-recurrence”mode at present. Robots can only complete point-to-point movements in the workingmode, so the flexibility of production line is unsatisfactory. Robots have the feature ofprocessing a large amount of information, high accuracy of detection, wide rangedetection and so on. The machine vision applied to industrial robots, robots cancomplete the target detection and judgment independently to improve the ability toadapt to the external environment.The paper studies on the visual industrial robot motion control on the basis ofREBOT-V-6R articulated industrial robot systems. The original robot control systemhas simple features, and can only easily crawl and identify single objects. This paperimproves the robot on the method of image processing, target location and motioncontrol by the study of visual image processing and motion control. The maincontents are as follows:Firstly, Researching the image processing theories and improving the imageprocessing algorithm of the robot by comparing and analyzing the effect of differenttreatment methods. And improving the image quality by pre-processing, usingthresholding method to do image segmentation, and then finish edge detection andcontour extraction for the image. Calculate the center of mass by calculating theminimum enclosing rectangle of the contour on the base of image contours.Secondly， Using D-H method for industrial robot kinematics modeling andanalying the robot’s kinematics, building the robot forward and inverse kinematicsequations and providing a theoretical for improving the robot motion control program.completting the design of a robot moving action accroding to the kinematic analysis.Thirdly, Doing the targeting experiments and identify the location of themulti-target, then get the coordinates of the centroid of each target object; accordingto coordinate datas on the basis of experiments conducted crawl handling experiments,all objects continuously moving target identification to specified location. The resultsshow that targeting is accurate, handling and rawling actions are so smoothly, and theexperiment achieved the desired goal.