基于双高斯纹理滤波模板和极值点韦伯对比度的圆柱锂电池凹坑缺陷检测

Pit Defect Detection of Cylindrical Lithium Battery Based on Double Gaussian Texture Filtering Template and Extreme Point Weber Contrast

本文针对圆柱锂电池表面图像具有亮度不均匀、金属表面反射不均、氧化锈斑和高亮噪声点等问题,提出一种基于机器视觉的解决方案.采用定义的双高斯纹理滤波模板与图像进行卷积,提取图像每列的灰度分布曲线,采用定义的极值点韦伯对比度选择曲线上突变点的阈值,根据先验知识筛选出凹坑候选区域,利用区域特征和灰度特征排除非凹坑纹理.测试结果表明,所提方法的拒真率(false rejection rate)和认假率(false accept rate)分别为5.49%和5.38%,亮度不均匀和金属表面反射不均没有对凹坑检测产生影响.

A solution based on machine vision is proposed to solve the problems of surface image of cylindrical lithi⁃um battery,such as uneven brightness,uneven reflection on metal surface,oxidation rust spots and highlighting noise points.A defined double Gaussian texture filtering template was used to convolve with the image.The grayscale distribu⁃tion curve of each column of the image was extracted.The extraction threshold of the discontinuous point on the grayscale distribution curve was calculated by using the defined extreme point Weber contrast.The candidate pit regions were screened out according to the prior knowledge.The non-pit textures were excluded by using region features and gray value features.The test results indicate the false rejection rate(FRR)and false accept rate(FAR)are 5.49 percent and 5.38 percent respectively.And the uneven brightness and uneven reflection had no effect on pit detection.