[1]郑文斌、蔡志山、俞斌、陈炜、赵毅武.汽车玻璃黑边区域划痕的视觉检测技术[J].福建工程学院学报,2020,18(01):63-68.[doi:10.3969/j.issn.1672-4348.2020.01.012]
　ZHENG Wenbin,CAI Zhishan,YU Bin,et al.Visual detection technology for scratches in frit band of automobile glass[J].Journal of FuJian University of Technology,2020,18(01):63-68.[doi:10.3969/j.issn.1672-4348.2020.01.012]
点击复制

汽车玻璃黑边区域划痕的视觉检测技术()
分享到：

《福建工程学院学报》[ISSN:2097-3853/CN:35-1351/Z]

卷:
第18卷
期数:
2020年01期
页码:
63-68
栏目:
出版日期:
2020-02-25

文章信息/Info

Title:
Visual detection technology for scratches in frit band of automobile glass
作者:
郑文斌、蔡志山、俞斌、陈炜、赵毅武
福建工程学院信息科学与工程学院
Author(s):
ZHENG Wenbin;  CAI Zhishan;  YU Bin;  CHEN Wei;  ZHAO Yiwu
School of Information Science and Engineering, Fujian University of Technology
关键词:
汽车玻璃;  黑边区域;  划痕检测;  图像处理;  视觉检测
Keywords:
automobile glass;  frit band;  scratch detection;  image processing;  visual detection
分类号:
TP391
DOI:
10.3969/j.issn.1672-4348.2020.01.012
文献标志码:
A
摘要:
针对汽车玻璃黑边区域划痕的自动检测需求，以汽车角窗玻璃为对象，进行划痕视觉检测方法的研究。首先采用灰度值拉伸来凸显划痕，通过维纳滤波降噪和Laplacian算子的锐化加强划痕细节，然后利用区域生长分割法进行玻璃黑边区域分割，采用形态学方法实现灰尘等干扰信息的去除以及断裂划痕的连接，最后用8领域连通区域标记法进行划痕的识别和标记。测试结果表明，该算法适用于不同型号的汽车玻璃黑边区域划痕的检测，检测的准确率达到99%，划痕标记的准确率达到97%，检测速度快，符合在线检测的要求。
Abstract:
According to the needs of automatic detection of scratches in the frit band of automobile glass, the study of visual detection for scratches was carried out with the car’s rear window glass as the object. First, the method of gray value stretching was used to highlight the scratches, Wiener filtering was used to reduce the noise of the requested image, and the scratch details were strengthened by Laplacian operator sharpening. Secondly, the frit band of the automobile glass was segmented from the requested image by using the regional growth segmentation method, and the morphological method was used to remove the interference information and connect the fractured scratches. Finally, the 8-domain connected region-marking method was used to identify and mark the scratch. Experimental results show that the algorithm can be applied to the detection of scratches on the frit band of different types of automobile glass. The accuracy of the detection is 99%, and the accuracy of the scratch marking is 97%. The detection is fast and meets the requirements of online detection.

参考文献/References:

[1] 宋迪. 基于机器视觉的手机隔板划痕检测研究[D]. 湘潭: 湘潭大学, 2014.[2] 李晨, 杨甬英, 熊浩亮, 等. 基于滤波差分的双阈值弱划痕提取算法[J]. 强激光与粒子束, 2015, 27(7): 103-110.[3] 吴珊, 陈斌, 杨溪远, 等. 卡片划痕检测关键算法设计[J]. 计算机应用, 2017,37(S2): 171-174.[4] 王静颖. 基于机器视觉的玻璃缺陷检测系统的研究[D]. 天津: 天津大学, 2014.[5] 武跃华. 机器视觉划痕检测技术及应用研究[D]. 广州: 广东工业大学, 2011.[6] 王昌书. 基于机器视觉的手机曲面玻璃典型缺陷检测研究[D]. 广州: 华南理工大学, 2019.[7] 张铮, 倪红霞, 苑春苗. 精通Matlab数字图像处理与识别[M]. 北京: 人民邮电出版社, 2013.[8] 张乐, 项安. 区域生长算法的改进及其在异物检测中的应用[J]. 无损检测, 2010, 32(7): 490-492.[9] 范群贞, 吴浩, 林真. 基于形态学和区域生长法的医学图像分割[J]. 机电技术, 2019(3): 26-29.[10] 赵文宏, 彭超, 陈红星. 基于形态学玻璃屏幕表面划痕检测方法研究[J]. 浙江工业大学学报, 2016, 44(3): 279-282.[11] TAO X, XU D, ZHANG Z, et al. Weak scratch detection and defect classification methods for a large-aperture optical element[J]. Optics Communications, 2017, 387: 390-400.

更新日期/Last Update: 2019-02-25