Research On Visual Sorting Method And Apllication Based On Template Matching

With the wide application of intelligent equipment in i ndustrial production,the machine vision has received much attention as an importan t branch of intelligent equipment.The domestic visual softwares are poor in speed,robustness and so on.Although many demestic enterprises and research instituti ons are doing related research,but the domestic market has been occupied by foreign visual softwares in a large proportion.According to our requirements in deal ing with practical problems of engineering,we have developed a visual sorting software independently and successfully applied it to engineering tasks.The main research contents are as follows:Firstly,a template matching method based on shape is studied,the similarity measure is using the dot product operation of the gradient vector of the target object' edge and the template' edge.By using greedy algorithm,the matching speed of the algorithm is greatly improved,and then the complexity of the algorithm is reduced by reducing the image resolution by using the image Pyramid.Through two methods,the matching speed of the algorithm is improved.In addition,the cross correlation method and the template matching method based on Hausdorff distance and modified hough transform are introduced.Secondly,for different objects,it can determine the minimum and maximum threshold automatically.The number of layers in the image and the translation step and the rotation step of the tempalte are determined auto matically,making the algorithm be more intelligent.In matching,through experiments,selecting the filtering method of building the image Pyramid.And solved the problem of finding the optimal targets.The visual sorting software is developed under the VS 2013 development environment by using C++ programming language.Then,design experiments to verify the performance of the proposed algorithm.It is divided into two sets of experiments,respectively,to test the robustness of the algorithm and the time used for the algorithm.In the ro bustness test,we mainly test the performance of algorithm against occlusion,confusion and nonlinear illumination change.The algorithm based on shape contrasts with the cross correlation method and the template matching method based on Hausdorff distance and modified hough transform,the rusult indicates that the algorithm based on shape is better.Finally,the visual sorting system of the Delta robot is built,and the calibration of the camera and the calibration between the robot and the belt are completed,ensuring the accuracy of the sorting sustem.Through the dynamic target tracking algorithm based on image sequence,the target recognition and acquisition are avoided.Through the visual software developed by ourselves,a lot of application s are finished successfully in the laboratory stage,and the packaging project of ham sausage has been successfully applied in practical production,greatly improve production efficiency.