锻压技术

基于双打光模板匹配的冲压件表面缺陷检测

英文标题：Surface defect detection of stamping parts based on double light pattern matching

作者：李松周亚同张忠伟池越韩春颖

分类号：TP391

出版年,卷(期):页码：2018,43(11):137-145

摘要：

针对冲压件表面缺陷人工检测效率低、误检率高以及劳动强度大等问题，提出了一种基于双打光模板匹配的冲压件表面缺陷检测方法。首先针对不同的表面缺陷类型，根据照明效果采用底部环形光源与顶部面光源两种不同的打光方式。接着先后采用边缘提取、旋转定位、填充、模板匹配及缺陷提取等方法对采集的冲压件图像进行处理。在线检测应用表明，划痕、月牙磕碰缺陷采用低角度环形光源的打光方式和锈迹缺陷采用高角度面光源的打光方式具有良好的缺陷提取效果。整个表面缺陷图像处理方法可以快速、有效地检测出冲压件的表面划痕、月牙磕碰、锈迹等缺陷。检测精度最高可达0.134 mm2，检测速度最快可达1.196 s。该方法在检测速度和精度上能满足工厂检测要求，大幅提高生产效率，降低生产成本。

For the problems of low manual inspection, high false detection rate and labor intensity in the surface defects of stamping parts, based on double-shot template matching, a surface defect detection method of stamped parts was proposed. According to the different surface defect types, two different lighting patterns, such as a bottom ring light source and a top surface light source, were adopted based on the lighting effect, and the collected stamping parts images were processed by edge extraction, rotation positioning, filling, template matching, and defect extraction. The application of online inspection shows that for the scratching and crescent bump defects, good defect extraction effect is obtained by low-angle ring light source lighting, and for the rust defects, high-angle surface light source lighting method is applied. Therefore, the entire surface defect image processing method can quickly and effectively detect defects such as surface scratches, crescent moon bumps, and rust marks on the stamped parts. Detection accuracy is up to 0.134 mm2, and detection speed is up to 1.196 s. Furthermore, this method can meet the requirements of factory inspection in terms of detection speed and accuracy to greatly increase production efficiency and reduce production costs.

基金项目：

河北省引进留学人员资助项目（CL201707）;教育部人文社会科学研究规划基金（15YJA630108）

李松（1994-），男，硕士研究生,E-mail：984646263@qq.com;通讯作者：周亚同（1973-），男，博士，教授，博士生导师,E-mail:zyt@hebut.edu.cn

参考文献：

[1]张国福,沈洪艳.机器视觉技术在工业检测中的应用综述[J].电子技术与软件工程,2013,(22):111.

Zhang G F, Shen H Y. Application of machine vision technology in industrial inspection[J]. Electronic Technology and Software Engineering, 2013,(22):111.

[2]刘振宇,李中生,冯柏润.机器视觉在工业生产线上的应用实现[J].微型机与应用,2013,32(17):27-30.

Liu Z Y, Li Z S, Feng B R. Realization of machine vision in industrial production line [J].Microcomputers and Its Applications,2013, 32(17):27-30.

[3]Chen Z X, Huang R, Zhang Z S, et al. New compensation method for surface roughness inspection by machine vision in different ambient light[J]. Journal of Southeast University, 2009, 39(1):136-140.

[4]徐信.冲压件表面缺陷图像检测系统的研究与开发[D].广州:广东工业大学,2016.

Xu X. Research and Development of Stamping Surface Defect Image Detection System[D]. Guangzhou: Guangdong University of Technology, 2016.

[5]黄柳青. 基于机器视觉的冲压件缺陷检测系统研究[D].广州:广东工业大学,2012.

Huang L Q. Research on Stamping Parts Defect Detection System Based on Machine Vision[D]. Guangzhou: Guangdong University of Technology,2012.

[6]陈宝文,姜军.基于计算机视觉的冲压件表面缺陷检测方法[J].科技信息,2014,(10):2-3,5.

Chen B W, Jiang J. Surface defect inspection method of stamping parts based on computer vision[J]. Science and Technology Information, 2014,(10): 2-3,5.

[7]陈广峰,管观洋,魏鑫.基于机器视觉的冲压件表面缺陷在线检测研究[J].激光与光电子学进展,2018,55(1):335-341.

Chen G F, Guan G Y, Wei X. Online stamping parts surface defects detection based on machine vision[J]. Laser and Optoelectronics Progress, 2018, 55(1): 335-341.

[8]庞章炯,张鹰,谢佩君,等.机器视觉在晶振外壳缺陷检测中的应用[J].制造技术与机床,2007,(3):87-91.

Pang Z J, Zhang Y, Xie P J, et al. Application of machine vision inspection system for crystal oscillator shell quality detection[J]. Manufacturing Technology and Machine Tool, 2007,(3): 87-91.

[9]陈海永,李泽楠, 孙鹤旭,等. 异形冲压件轮廓视觉检测系统设计[J].光学精密工程,2016,24(10):229-236.

Chen H Y,Li Z N,Sun H X, et al. Visual detection system design for profile of special-shaped stamping parts[J]. Editorial Office of Optics and Precision Engineering, 2016, 24(10):229-236.

[10]Aaron R. Rababaah, Yotan Demi-Ejegi. Automatic visual inspection system for stamped sheet metals (AVIS3M)[J]. IEEE International Conference on Computer Science, 2012, 2(9): 661-665.

[11]Wu X, Gao G. LED light design method for high contrast and uniform illumination imaging in machine vision[J].Applied Optics, 2018, 57(7): 1694.

[12]Dyer S A, He X. Least-squares fitting of data by polynomials[J]. Instrumentation and Measurement Magazine IEEE, 2002, 4(4): 46-51.

[13]张彪,邹哲,陈书界,等.基于仿射变换与Levenberg-Marquardt算法的织物图像配准[J].光学学报,2017,(1):351-359.

Zhang B, Zou Z, Chen S J, et al. Fabric image registration based on affine transformation and Levenberg-Marquardt algorithm[J]. Acta Optica Sinica, 2017,(1): 351-359.

[14]时佳悦,张蕊萍,董海鹰,等. 基于灰度变化率的低对比度CT图像分割研究[J]. 兰州交通大学学报,2017,36(3):57-62.

Shi J Y, Zhang R P, Dong H Y, et al. A defect segmentation algorithm for low contrast CT images based on the change rate of gray level[J]. Journal of Lanzhou Jiaotong University, 2017, 36(3): 57-62.

[15]左敏,曾广平,涂序彦,等.基于等价对的图像连通域标记算法[J].计算机仿真,2011,28(1):14-16.

Zuo M, Zeng G P, Tu X Y, et al. A connected domain labeling algorithm based on equivalence pair in binary image[J]. Computer Simulation, 2011, 28(1):14-16.

服务与反馈：

【本网站尚未开通全文下载服务】【加入收藏】