Research On Visual Detection Algorithm For Surface Defects Of Cell Based On Intelligent Learning

With the rapid development of the new energy automobile industry,the demand for lithium ion power batteries has gradually increased.As a core component of a lithium battery,the quality of the battery not only affects the service life of the lithium battery,but also causes safety problems.The traditional manual inspection has been unable to meet the testing requirements of modern industry.Combined with machine vision technology,the defect detection of the cell can be completed with high precision and high efficiency to achieve the level of modern industrial automation detection.Therefore,this paper proposes a machine vision system that automatically detects the surface defects of the cell.The main work is as follows:Firstly,hardware experimental platform for cell surface defect detection has been established,including the design of the cell imaging system and the selection of industrial cameras,lenses and light sources.It completes the acquisition on the surface of the cell.Secondly,According to the structural characteristics of the surface defects of the cell,the specific process of the defect detection algorithm has been introduced.According to the structural characteristics of the surface defects of the cell,the specific process of the defect detection algorithm has been introduced.This paper has used median filter and histogram equalization algorithm to Image preprocessing.An adaptive threshold segmentation method has been proposed for pre-processed images.Then,the morphological processing method has been used to obtain a defective area on the surface of the cell.It completes defect detection on the surface of the cell.Thirdly,the grayscale features and morphological features of the defect area of the cell have been extracted by studying extraction method of the image feature.According to the experimental results and feature evaluation criteria,six sets of feature data have been selected.To realize the classification of defects on the surface of cell,a classification method combining decision tree and support vector machine has been proposed by comparing K nearest neighbor method,support vector machine and decision tree.Finally,the software platform of the whole detection system has been built,which realized the functions of image acquisition and display,cell defect detection and classification,and display of experimental results.The accuracy of the entire inspection system can reach 95%,and the accuracy of metal leakage and decarburization defects can reach 95.44%.It proves the feasibility and high operational efficiency of the detection algorithm.