基于机器视觉的锂电池极耳焊接缺陷检测技术研究与分析

Research and Analysis of Welding Defect Detection Technology for Lithium Battery Lug Based on Machine Vision

针对锂电池极耳在焊接工艺中所产生的极耳翻折、焊点不足以及焊破等缺陷问题,提出机器视觉对极耳的焊接缺陷进行识别与检测。首先,利用球积分光源以及同轴光所组成的组合式光源对焊后的极耳进行均匀化图像采集,并提取相应的ROI;然后,采用中值滤波的方法对ROI进行平滑去噪,并使用分段线性灰度增强的方法对图像进行预处理,再利用最大类间方差法(OSTU)对ROI进行图像分割以及特征提取,有效地对极耳焊接缺陷进行了识别与检测;最后,对极耳焊接缺陷检测系统的检测效果进行了分析,检出率为98.77%,验证了视觉检测方法的可行性与合理性。

In view of the defects of welding process of lithium battery lug,such as lug fold,insufficient solder joint and weld break,machine vision was proposed to identify and detect the welding defects of the lug.Firstly,the combined light source composed of spherical integral light source and coaxial light source was used to collect the homogenized image of the electrode ear after welding,and the corresponding ROI was extracted;then,the method of median filtering was used to smooth and denoise the ROI,and the method of piecewise linear gray enhancement was used to preprocess the image,and the method of maximum inter class variance(OSTU)was used to segment and characterize the ROI.Finally,the detection effect of the welding defect detection system was analyzed,and the detection rate was 98.77%,which verified the feasibility and rationality of the visual detection method.