Application Research Of Vision Guiding Technique In Assemble Robot Grasping Workpiece

Vision system applying to industrial robots has an important significance to improvethe intelligence level of industrial robots in the build process. Visual guidance is based oncomputer vision technology, use image information to locate the position of the targetobject in space, guide the direction for industrial robots to complete specific work. In thisprocess, the location accuracy of vision system has an important influence for guidingeffect.Guiding industrial robot based on vision to complete workpiece grasp is the supporttechnology of most of the assembly line, the goal of this paper is to guide the robot tocomplete the accurate hub workpiece grasp, and the key problems to be solved includeimage distortion rectification, hub workpiece recognition and accurate location. In order torectify the distortion of the images collected, first use Zhang Zhengyou calibration methodto calculate the internal and external camera parameters, the least squares method tocalculate the camera distortion coefficients, and then Newton iterative method for imagedistortion rectification. In view of the problem of low accuracy to deal directly with theimage itself, propose a recognition method based on design template. First, use the hubworkpiece design parameter information to generate design templates through the pixelmapping mode. Second, calculate the geometry feature of the boundary contour in designtemplate and hub workpiece image to combine into eigenvectors with translation, rotationand scaling invariance. Last, compare eigenvectors between hub workpiece and designtemplate to complete hub workpiece recognition based on the minimum similarityprinciples. The strong point of the method is to make full use of the accurate informationprovided by design parameters, therefore can effectively improve the accuracy of thefeature recognition. At the same time, it can be further formed a local alignment locationmethod based on the design template. This method takes the gas nozzle area of the designtemplate corresponding to the identified hub workpiece as a reference. Use differencematrix of the area in workpiece image and design template, through rotational and translational alignment, to obtain the final accurate location results of center coordinates ofthe workpiece and gas nozzle towards.This paper takes an experimental study on several hub workpiece images collected inwork environment. Experimental results show that this method can locate the key parts ofthe hub workpiece accurately. The correct recognition rate is above98%and erroraccuracy of position and angle is respectively within2pixels and0.1degrees. It cancomplete accurate workpiece location and meet the actual needs of the production line.