Research On Tool Wear Detection System Based On Machine Vision

Metal cutting has a great influence on the development of the manufacturing industry,and the wear and tear of cutting tools are closely related to the quality of processing.There are too many subjective factors when testing tool wear by using artificial experience,so it is easy to cause the waste of resources.But the traditional cutting tool wear detection when machine goes down,although it can accurately detect tool wear,but also reduces the processing efficiency.In the process of single batch and small batch customization,special tools are sometimes needed,and tool wear can not be judged according to the traditional tool wear experience.Therefore,it is necessary to design an intelligent,intuitive and real-time tool wear detection system.This thesis mainly complete the following contents:The tool wear detection technology based on machine vision is studied,the tool wear image preprocessing algorithm is researched,and the method of image recognition tool wear measurement is designed.The dynamic acquisition scheme of the ball end milling cutter and the milling cutter side edge is designed respectively according to the characteristics of the ball end milling cutter and the milling cutter side edge and the rotating speed of the machine tool spindle,the feed,the diameter of the cutter and the collection field of CCD industrial camera.Carry out the test of milling tool wear on the machine in the designed tool wear detection system.Firstly the tool wear image is collected by the acquisition scheme designed in this paper,and image registration and image fusion research is carried out by using SIFT algorithm.Secondly The real-time multi milling tool wear images are spliced into clear tool wear images.Finally,image processing is done to identify tool wear amount,and automatic recognition experiments are carried out against milling cutter and ball end milling cutter respectively.The tool wear is identified and compared with the wear volume detected by the super depth of field.The type recognition of tool wear based on convolution neural network is studied.The training set of this paper is built by using the Image Net network structure,and the data set is trained by the feedback neural network.Finally,the accuracy of two network structures is compared with the pre training convolution neural network and the untrained convolution neural network.In this paper,the system is designed using Visual Studio,and CCD camera SDK MATLAB software development kit for jointly developing tool wear detection system,using Visual Studio development software interface,Visual Studio and camera CCD image acquisition part development kit combines development,MATLAB image processing and deep learning kit with Visual Studio 2010 hybrid programming writing tool wear abrasion quantity measurement and tool wear recognition part.