Research On Vision Inspection Method Of Cylindrical Lithium Battery End Face Defects

In the manufacturing process,due to the lack of manufacturing level,defects such as pits and scratches may be left on the surface of the battery,and even electrolyte leakage may occur.Once such batteries enter the market,it will not only reduce the company’s image,but also cause a series of security risks.Therefore,it is necessary to detect and eliminate the defective batteries before they leave the factory.Some battery manufacturers still rely on workers’ naked eyes to detect,but the way of detecting cylindrical lithium battery end face defects by naked eyes has been unable to adapt to today’s rapid development of society.More and more attention has been paid to the new detection methods of machine vision,camera instead of human eye and computer instead of human brain.In this paper,machine vision is used to detect indentation,deformation,positive position deviation and leakage defects of cylindrical lithium battery.This paper introduces the causes and characteristics of each defect,and briefly introduces the imaging method of the defect.The battery area is divided,and the detection scheme of different areas is defined.For the metal surface indentation defect,the idea of image difference is used to highlight the defect features on the low resolution image,and then the gray distribution features of the image are analyzed.The defect area is extracted by the adaptive threshold method,and the region is screened by morphological features and location features,and finally the defect area is obtained.There are two kinds of metal surface deformation defects,one is severe deformation,the other is slight deformation.For severe deformation defects,they are directly detected by global fixed threshold and feature selection.For the slight deformation defect,the feature is described firstly.After the gray feature is clear,the gray curve information is extracted to achieve the purpose of recognition.For the defect of positive position offset,the circle of gasket at fixed position is extracted by Hough transform to obtain the axis coordinates of battery,and then the contour line of cover plate is obtained by edge detection to obtain the center coordinates of cover plate.The two coordinates are compared to determine whether the offset is or not.Two schemes for the detection of battery leakage defects are studied.They are the leakage detection method based on the adaptive threshold of the gray distribution in the hole and the leakage detection method based on the gray curve amplitude screening.The library test shows that the latter has better detection effect and can avoid the false detection problem of battery leakage detection due to image matching.