软包装锂离子电池表面划痕的检测方法  

Detection method for surface scratches of pouch Li-ion battery

作　　者：谷峰 王正家[1,2,3] 昝傲 权佳怡 GU Feng;WANG Zheng-jia;ZAN Ao;QUAN Jia-yi(School of Mechanical Engineering,Hubei University of Technology,Wuhan,Hubei 430068,China;Hubei Key Lab of Modern Manufacture Quality Engineering,Wuhan,Hubei 430068,China;School of Detroit Green Technology,Hubei University of Technology,Wuhan,Hubei 430068,China)

机构地区：[1]湖北工业大学机械工程学院,湖北武汉430068 [2]现代制造质量工程湖北省重点实验室,湖北武汉430068 [3]湖北工业大学底特律绿色工业学院,湖北武汉430068

出　　处：《电池》2023年第5期514-518,共5页Battery Bimonthly

基　　金：国家自然科学基金(51275158)。

摘　　要：表面缺陷是判断软包装锂离子电池质量的重要指标之一。针对软包装锂离子电池表面凹凸不平且存在反光,导致划痕自动化检测检出率低的问题,提出一种表面划痕检测方法。采用电荷耦合元件(CCD)相机及远心镜头,从3个方向拍照,通过光度立体法融合图像,生成电池表面反照率图像;通过二阶高斯卷积进行图像平滑,并提取划痕边缘特征信息;采用非线性灰度变换增强划痕细节;通过阈值分割对划痕进行提取。划痕的检测准确率为97.5%,满足检测精度需求。The surface defects were one of the important indicators to judge the quality of pouch Li-ion battery.Aiming at the uneven and reflective surface of pouch Li-ion battery,resulting in low detection rate when automatic detection of scratches,a scratch detection method on the surface was proposed.A charge coupled device(CCD)camera and telecentric lens were used to take pictures from three directions,the images were fused by photometric stereoscopic method to generate a battery surface albedo image.The image was smoothed by second-order Gaussian convolution,the scratch edge feature information was extracted.The scratch details were enhanced by nonlinear grayscale transformation.The scratches were extracted by threshold segmentation.The detection accuracy of scratches was 97.5%,which met the requirements of detection accuracy.

关 键 词：划痕检测 反照率图像 非线性灰度变换 软包装锂离子电池 表面划痕 光度立体法 高斯卷积 

分 类 号：TM912.9[电气工程—电力电子与电力传动]