Research On Key Technologies Of Online Detection System For Positive And Negative Defects Of Power Lithium Battery

In the process of lithium battery production,differnt kinds of defects may appear in battery‘s positive and negative pole,resulting in great security risks.In actual production,artificial detection is used to detect the lithium battery.This method is prone to error detection and is inefficient.In this paper,18650 power lithium batteries were used as the research object.An image preprocessing algorithm for positive and negative electrodes of lithium battery was proposed.Different classification algorithms were proposed for different characteristics of positive and negative pole images.A positive and negative pole online detection system for lithium battery was built.The main work and research results are as follows:(1)The types of defects were analyzed.According to the requirements of lithium battery defect detection,the overall framework of detection algorithm was designed.According to the actual demand,the appropriate light source type and lighting mode were selected,and the hardware requirements of the industrial camera were determined.A visual system composed of industrial cameras and industrial light sources was set up,and the lithium battery images were collected.(2)In order to improve the quality of image,a suitable image preprocessing algorithm was proposed,which mainly includes Gauss denoising algorithm,improved homomorphic filtering algorithm,Sobel edge detection algorithm and adaptive threshold algorithm.The experimental results showed that after processing,the noise was reduced,the brightness and contrast were improved,and the edge information and the defect features were significantly enhanced.(3)The various features of the image were analyzed,and four kinds of data were selected as the characteristic quantity.SVM was used to classify images,and cross validation was used to optimize classifier parameters.This classification method was compared with the other two methods.The experimental results showed that: SVM based classification algorithm has better detection effect for lithium battery imges.(4)The hardware and software of the detection system was designed and built.Three detection methods were verified on the experimental platform.The experiment showed that the algorithm has good effect by comparing with traditional algorithm and artificial detection.The accuracy,safety and detection speed were improved.