基于人工神经网络的太阳能电池片表面质量检测系统  

Solar cell surface quality detection system based on artificial neural network

作　　者：伍李春 刘明周[1] 蒋倩男 葛茂根[1] 凌琳[1] 

机构地区：[1]合肥工业大学机械工程学院,安徽合肥230009

出　　处：《合肥工业大学学报（自然科学版）》2017年第9期1176-1180,1192,共6页Journal of Hefei University of Technology：Natural Science

基　　金：安徽省科技攻关计划资助项目(1604a0902182)

摘　　要：针对依赖人工进行太阳能电池片表面质量检测时效率和精度低的问题,文章提出了基于机器视觉以及人工神经网络的太阳能电池片表面质量检测方法。将表面缺陷分为外形缺陷、颜色缺陷、裂纹以及丝印线路缺陷4类,基于模板匹配检测外形缺陷,基于HIS空间下的颜色直方图检测颜色缺陷;针对细微性缺陷容易受噪声影响的特点,利用2类人工神经网络进行断栅检测,并对这2类神经网络进行比较。大量实验结果验证了上述方法能够准确、快速地检测出太阳能电池片表面缺陷。For the low efficiency and precision problem of solar cell surface test relying on manual la- bor, a method of the solar cell surface detection based on maehine vision and artificial neural network is raised. Surface defects are firstly divided into four categories, including appearance defects, color defects, cracks and defects of screen printing line. Then the appearance defects are detected based on template matching and the color detection is realized according to the image of color histogram on HIS space. Finally, for the characteristics of small defects which are easily affected by noise, two types of artificial neural networks are used to detect the broken gate and the two networks are compared. The experimental results show that the presented method can accurately and quickly detect the solar cell surface defects.

关 键 词：缺陷检测 机器视觉 人工神经网络 正则化径向基函数(RBF)网络 学习向量化(LVQ)网络 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]