Online Vision Measurement for Welding-quality Parameters of Solar Silicon Chips
-
摘要: 针对硅电池片高速串焊系统,根据焊接质量参数在线检测的技术指标对视觉测量系统的核心部件进行了选型计算;对视觉测量系统的内部参数进行了标定,完成了畸变校正与像素当量计算;设计了焊后硅片视觉检测流程,研究了图像预处理、图像掩膜ROI提取、改进双阈值Otsu图像分割技术;建立了硅电池片焊带偏移参数测量模型,提出了一种硅电池片焊接质量参数在线视觉测量方法。最后将该在线视觉测量方法应用于某硅电池片高速串焊生产线,其综合检测率和单帧检测时间满足焊接质量在线检测的精度和实时性要求,为提高太阳能电池组件的焊接质量和生产效率打下了良好的技术基础。Abstract: A vision measurement system is firstly designed by selecting its key components and calculating their performance parameters for the online detection requirements of welding quality for a high-speed serial welding system of silicon chip. Secondly, the intrinsic parameters of the vision measurement system are calibrated, which is then used for distortion correction and pixel-equivalent calculation. Thirdly, the vision detection process is designed for the silicon chip after welding, in which the techniques of image preprocessing, image-masking ROI extraction and improved double-threshold Otsu image segmentation are investigated, and a measurement model for welding-strip offset parameter is established. On this basis, an online vision measurement approach is developed for the welding-equality parameters of solar silicon chips. Finally, the present measurement approach is applied to a high-speed serial welding system of silicon chip. The comprehensive detection rate and measurement time per frame meet the accuracy and real-time requirements in the process of online measurement for welding equality, which lays a good technical foundation for improving welding equality and production efficiency of solar battery modules.
-
Key words:
- welding defect /
- online detection /
- vision measurement /
- defect recognition /
- image processing
-
表 1 串焊检测系统在线视觉测量技术指标
检测对象 电池串(单晶/多晶8-12片/串) 检测范围 125 mm×125 mm、156 mm×156 mm、78 mm×156 mm 检测节拍 2 s 总误检率 小于0.2% 漏检率 小于0.2% 片间距 检测范围1~5 mm, 误差在0.1 mm以内 焊带偏移 长度小于30 mm允许3处, 检测
精度0.1mm, 偏移量自由设置起焊点 同一片多条主栅起焊点位置偏差
小于0.5 mm, 误差0.2mm以内表 2 CCD摄像机标定结果
名称 标定结果 内参数矩阵 径向畸变 k1=0.0926, k2=-0.1754 标定误差 (kx, ky)=(3.36, 3.12)
(u0, v0)=(0.826, 0.715)
(k1, k2)=(0.003, 0.005)表 3 像素当量标定结果
编号 1 2 3 4 5 相邻圆心
距/pixel65.292 3 65.293 3 65.292 3 65.288 0 65.292 3 像素当
量/μm61.263 61.262 61.263 61.267 61.263 表 4 串检综合检测率
序号 nf nt nk/ng me/mo η 1 9 406 855 831/24 9/0 98.92% 2 10 341 940 925/15 6/1 92.68% 3 9 109 828 815/13 0/8 99.02% 4 9 934 903 885/18 1/9 93.43% 5 10 088 917 901/16 0/5 99.45% 表 5 单帧图像检测时间
定位分割 偏移 起焊点 片间距 总时间 < 100 ms 125 ms 24 ms 15 ms < 1 s -
[1] 刘苏宜, 王国荣, 钟继光.视觉系统在机器人焊接中的应用与展望[J].机械科学与技术, 2005, 24(11):1296-1300 doi: 10.3321/j.issn:1003-8728.2005.11.010Liu S Y, Wang G R, Zhong J G. Application and prospect of vision sensing system in robot welding[J]. Mechanical Science and Technology, 2005, 24(11):1296-1300(in Chinese) doi: 10.3321/j.issn:1003-8728.2005.11.010 [2] 张舞杰, 李迪, 叶峰.硅太阳能电池纹理缺陷检测[J].计算机应用, 2010, 30(10):2702-2704, 2708 http://d.old.wanfangdata.com.cn/Periodical/jsjyy201010038Zhang W J, Li D, Ye F. Texture defect inspection for silicon solar cell[J]. Journal of Computer Applications, 2010, 30(10):2702-2704, 2708(in Chinese) http://d.old.wanfangdata.com.cn/Periodical/jsjyy201010038 [3] 孙智权, 童钢, 赵不贿, 等.太阳电池自适应色系分类方法研究[J].太阳能学报, 2017, 38(6):1546-1552 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=tynxb201706013Sun Z Q, Tong G, Zhao B H, et al. Research of adaptive color classification method for solar cells[J]. Acta Energiae Solaris Sinica, 2017, 38(6):1546-1552(in Chinese) http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=tynxb201706013 [4] 高志良.太阳能电池片自动焊接中视觉检测关键技术的研究[D].天津: 河北工业大学, 2013Gao Z L. Key technologies research on the visual detection for automatic welding of solar cells[D]. Tianjin: Hebei University of Technology, 2013(in Chinese) [5] 李云峰, 李晟阳.基于梯度方向信息熵的印刷电路板缺陷检测[J].中国机械工程, 2017, 28(6):695-701 doi: 10.3969/j.issn.1004-132X.2017.06.010Li Y F, Li S Y. Defect detection for PCBs based on gradient direction information entropy[J]. China Mechanical Engineering, 2017, 28(6):695-701(in Chinese) doi: 10.3969/j.issn.1004-132X.2017.06.010 [6] Ghidoni S, Antonello M, Nanni L, et al. A thermographic visual inspection system for crack detection in metal parts exploiting a robotic workcell[J]. Robotics and Autonomous Systems, 2015, 74:351-359 doi: 10.1016/j.robot.2015.07.020 [7] 周太明, 周详, 蔡伟新.光源原理与设计[M].2版.上海:复旦大学出版社, 2006:8-10Zhou T M, Zhou X, Cai W X. Principle and design of light source[M]. 2nd ed. Shanghai:Fudan University Press, 2006:8-10(in Chinese) [8] Zhang Z Y. A flexible new technique for camera calibration[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2000, 22(11):1330-1334 doi: 10.1109/34.888718 [9] 刘亚强, 陈文艺.桶形畸变图像的一种校正方法[J].西安邮电学院学报, 2012, 17(2):27-31, 36 doi: 10.3969/j.issn.1007-3264.2012.02.007Liu Y Q, Chen W Y. A correction method of barrel distortion image[J]. Journal of Xi'an University of Posts and Telecommunications, 2012, 17(2):27-31, 36(in Chinese) doi: 10.3969/j.issn.1007-3264.2012.02.007 [10] 范彦斌, 李西兵, 沈自林.基于亚像素的边界检测与提取算法及误差分析[J].机械科学与技术, 2005, 24(2):189-191, 211 doi: 10.3321/j.issn:1003-8728.2005.02.018Fan Y B, Li X B, Shen Z L. The algorithm and accuracy analysis of edge detection and extraction based on the sub-pixel in digital image[J]. Mechanical Science and Technology, 2005, 24(2):189-191, 211(in Chinese) doi: 10.3321/j.issn:1003-8728.2005.02.018 [11] 陈怀东, 曹宗杰, 王久洪, 等.图像增强技术在超声成像检测中的应用[J].机械科学与技术, 2002, 21(S1):69-71 http://www.cnki.com.cn/Article/CJFDTotal-JXKX2002S1028.htmChen H D, Cao Z J, Wang J H, et al. Application of image enhancement technology to ultrasonic test[J]. Mechanical Science and Technology, 2002, 21(S1):69-71(in Chinese) http://www.cnki.com.cn/Article/CJFDTotal-JXKX2002S1028.htm [12] Ogura T, Hida K, Masuzuka T, et al. An automated ROI setting method using NEUROSTAT on cerebral blood flow SPECT images[J]. Annals of Nuclear Medicine, 2009, 23(1):33-41 doi: 10.1007/s12149-008-0203-7 [13] 汤楠, 穆向阳, 胡旭东.应用形态学运算的焊缝X射线图像处理研究[J].机械科学与技术, 2006, 25(9):1070-1072, 1119 doi: 10.3321/j.issn:1003-8728.2006.09.018Tang N, Mu X Y, Hu X D. A study of welding seam x-ray image processing using morphology[J]. Mechanical Science and Technology, 2006, 25(9):1070-1072, 1119(in Chinese) doi: 10.3321/j.issn:1003-8728.2006.09.018 [14] 赵慧阳.基于机器视觉的太阳能电池片表面缺陷检测的研究[D].秦皇岛: 燕山大学, 2011Zhao H Y. Research on surface defect inspection of solar cell based on machine vision[D]. Qinhuangdao: Yanshan University, 2011(in Chinese) [15] 武星, 张颖, 李林慧, 等.复杂光照条件下视觉导引AGV路径提取方法[J].农业机械学报, 2017, 48(10):15-24 doi: 10.6041/j.issn.1000-1298.2017.10.002Wu X, Zhang Y, Li L H, et al. Path extraction method of vision-guided AGV under complex illumination conditions[J]. Transactions of the Chinese Society for Agricultural Machinery, 2017, 48(10):15-24(in Chinese) doi: 10.6041/j.issn.1000-1298.2017.10.002 [16] Otsu N. A threshold selection method from gray-level histograms[J]. IEEE Transactions on Systems, Man, and Cybernetics, 1979, 9(1):62-66 doi: 10.1109/TSMC.1979.4310076 [17] 申铉京, 刘翔, 陈海鹏.基于多阈值Otsu准则的阈值分割快速计算[J].电子与信息学报, 2017, 39(1):144-149 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=dzkxxk201701021Shen X J, Liu X, Chen H P. Fast calculation of threshold based on multi-threshold Otsu criterion[J]. Journal of Electronics & Information Technology, 2017, 39(1):144-149(in Chinese) http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=dzkxxk201701021 [18] Vincent L. Morphological transformations of binary images with arbitrary structuring elements[J]. Signal Processing, 1991, 22(1):3-23 http://cn.bing.com/academic/profile?id=f15a2a42c0518437545b634ae8391a0c&encoded=0&v=paper_preview&mkt=zh-cn [19] 陈军, 邝泳聪, 欧阳高飞, 等.硅太阳能电池制备过程的质量控制策略[J].机械设计与制造, 2011(9):252-254 doi: 10.3969/j.issn.1001-3997.2011.09.093Chen J, Kuang Y C, Ouyang G F, et al. The quality control strategy of silicon solar cells production[J]. Machinery Design & Manufacture, 2011(9):252-254(in Chinese) doi: 10.3969/j.issn.1001-3997.2011.09.093