Research On On-line Detection Method Of Cylindrical Lithium Battery Surface Defects

Cylindrical lithium battery is a kind of lithium battery,its process is mature,high degree of automation,product quality is stable,relatively low cost,so the market share is very high,the application range is relatively wide.In the production process,the battery surface will inevitably appear a variety of defects,affecting the quality of the product and the assembly of the battery pack.At present,the surface defects of batteries mainly rely on manual testing,while the manual detection efficiency is low,the leakage rate is high,so it is difficult to meet the needs of automated production.Therefore,according to the requirements of the cooperative enterprise,this paper designs the circular surface defect detection system of lithium battery,and realizes the location of bar code,QR code and the detection of pit defects.This paper introduces the types of circular surface defects and the overall design scheme of the detection system,analyzes the causes of defects,the key points and difficulties of the system.According to the requirements of the testing equipment design hardware imaging system;on the basis of the specific detection needs to select the appropriate camera,lens and light source;through the experiment to determine the parameters of the imaging equipment;in the light of the shallow pit imaging experimental results of the light source space location design;in accordance with the two-dimensional code recognition results to determine the camera’s best photo number;in line with the photo scheme,the trigger device and the external trigger circuit of the camera are designed.According to the demand of the site,this paper designs the man-machine interface of the surface defect detection system of lithium battery: the design of Image acquisition module and algorithm detection module is realized through multithreaded architecture,the information interaction between defect detection system and motion control system is realized through Socket network communication,and the result summary and sending of multi-station are completed.The image preprocessing part introduces the method of obtaining grayscale image,the location of lithium battery and the method of extracting characters.Barcode and QR code location are realized by the morphological method of custom structure element.A pit defect detection method based on density detection,and a pit detection method based on Gaussian gradient convolution and grayscale characteristics are proposed.By comparing and analyzing the detection effect of the above two methods,the defect detection of pits is realized.The system designed in this paper has been debugged in the industrial field,and the detection time of a single battery is within 300 ms,which can basically meet the needs of the field.The detection rates of barcodes and QR codes are 98.08% and 97.97%,respectively,and the main reasons for inspection are the wear and breakage of barcodes and QR codes to varying degrees.The leakage rate of pit defects is 10.47%,and there are two main reasons for the leakage: first of all,the characteristics of the pits are not obvious;secondly,the pits is located at the bottom edge,resulting in the lack of characteristic information.The error detection rate of pit is 11.04%,the main reason of error detection is uneven distribution of image grayscale or regional edge drum bag.