基于红外图像的太阳能电池缺陷检测  

Detection of solar cell defects based on infrared image

作　　者：周维芳[1] 王志陶[2] 杨帆[2] 潘国锋[2] 

机构地区：[1]北华大学电气信息工程学院,吉林132000 [2]河北工业大学信息工程学院,天津300401

出　　处：《河北工业大学学报》2014年第1期8-14,共7页Journal of Hebei University of Technology

基　　金：国家科技重大专项课题(2009ZX02308-004)

摘　　要：针对目前太阳能电池缺陷检测中存在的缺陷检测种类单一且检测算法抗干扰能力较差等问题,提出了一种可在复杂背景的红外太阳能电池图像中检测并识别多种缺陷的方法.首先提出了一种快速的局部自适应阈值的二值化处理方法来提取缺陷信息和电极信息;其次做水平投影,定位电极位置,填充电极以消除其干扰;最后采用了一种根据缺陷轮廓信息特征进行分类识别的方法.经过大量实验验证,该方法可准确检测并识别红外太阳能电池图片中断线、隐裂、履带印等缺陷,检测准确率可达到99%以上,具有很好的适应性和准确性,大大提高了电池的检测效率.To solve the problems that single defect detection types and bad anti-interference ability of test algorithm in tests of infrared solar cells defections, a new method is proposed which can defect and identify various defections in com- plicated background of the solar cells pictures. Firstly, a method of local adaptive thresholding binarization is used to extract the messages of defections and electrodes. Secondly, horizontal projection can be used to get the positions of elec- trodes, then interferences can be eliminated by filling electrodes. Finally, a method which can classify according to de- fections contour information is used. A large number of experimental results demonstrate that our method can detect and identify the broken line fault, crack fault and trackprints fault accurately in the infrared image of solar cells, and the correct rate is over 99 %. The method with good adaptation and accuracy improves the solar cells defect efficiency greatly.

关 键 词：太阳能电池 缺陷检测 水平投影 轮廓 识别 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]