Design Of Automatic Gluing System For Shoe Sole Based On Machine Vision

As a kind of necessities for human life,the quality of shoes has always been concerned by people.And the gluing quality of shoes greatly affects its durability and aesthetics.The gluing of traditional footwear industry is manual,not only the quality of gluing is difficult to guarantee,the efficiency of gluing is low,and the body will be irreversibly damaged if exposed to the glue for a long time.Machine vision is widely used in industrial production,it can work in scenes that is harmful to humans,and more efficient.Therefore,this thesis mainly discusses the extraction of sole gluing trajectory based on machine vision technology,and the robot is used to replace the manual spraying operation,so as to improve the quality and speed of sole gluing and promote the development of automation in the footwear industry.The specific research contents mainly include:Firstly,a method of extracting 2D sole gluing trajectory based on monocular vision is studied.A median filtering method fused with wavelet transform is used to preprocess the image,and the improved maximum entropy segmentation algorithm and Sobel operator is used to extract the contour of the sole,so as to improve the segmentation accuracy.At the same time,in order to obtain the 2D gluing trajectory that can make robots run,the outer contour of the sole is offset inwards equidistantly,and the NUBRS curve is used for difference to make the trajectory smoother.Secondly,a method of extracting 3D sole gluing trajectory based on binocular vision is discussed.The matching cost calculation method combining BT cost and census transform is used to match the left and right images,and the four path aggregation strategy is adopted for cost aggregation to avoid the lack of accuracy of single matching cost;At the same time,the point cloud image obtained by the 3D reconstruction is projected to the 2D plane for offset processing,and reflect them back to the point cloud to obtain the 3D gluing trajectory of the sole.Then,the Nurbs curve is used to smooth the 3D gluing trajectory of the sole.Finally,a sole automatic gluing system based on machine vision is designed and its general framework is analyzed.The hardware configuration of each system of the gluing equipment is briefly introduced,and the software interface of the automatic gluing system of the sole is designed,and the program of extracting and optimizing the gluing trajectory of the sole is written,the feasibility of the system is verified by experiments.To sum up,the research in this thesis based on machine vision to extract the gluing trajectory of the sole has certain practical value for promoting the realization of full automation in the footwear industry.