Intelligent Online System of Perforated Workpiece Detection
-
摘要: 针对工业生产中用传统方式对带孔工件检测的误差大、效率低等问题,提出了一种基于机器视觉的带孔工件在线智能检测系统。该系统应用了由粗到精的智能检测网络,采用了改进的霍夫变换算法及连通域标记算法,实现了对工件的孔洞完整性检测、工件制式合格性检测以及核心孔的孔径尺寸测量。通过实验测试,整个检测过程约10 s,最大测量偏差不超过0.5pixel,能准确完成对工件测量合格性的判断,达到了生产线上实时精密、非接触、稳定性高的智能化检测要求。Abstract: An intelligent online system based on machine vision for perforated workpiece detection is proposed in this paper to solve the large errors and low efficiency in traditional detection methods in industrial production. The system has designed a coarse to fine intelligent detection network. It completes the hole integrity detection of the workpiece, the inspection of the workpiece system conformity and the measurement of the aperture size of the core hole by using the improved Hough transform algorithm and connected domain labeling algorithm. The detection process takes about 10 seconds, and the maximum measurement deviation is less than 0.5 pixel. It can accurately determine the qualification of the workpiece and meets the requirements of real-time, precise, non-contact and high-stability intelligent detection in the production line.
-
Key words:
- machine vision /
- perforated workpiece /
- intelligent detection /
- Hough transform algorithm
-
表 1 工件制式测量结果(ε=0.8)
mm 测量次数 1 2 3 4 5 6 7 8 D-value 0.165 0.224 0.008 0.773 0.061 0.054 0.012 0.607 表 2 左孔测量结果pixel
参数 1 2 3 4 5 6 7 8 半径R 219.333 219.320 219.310 219.333 219.342 219.310 219.330 219.345 圆心X 652.5 652.5 652.0 652.5 652.5 653.0 652.5 652.5 圆心Y 1269.5 1269.0 1269.5 1269.5 1269.5 1270.0 1269.5 1269.5 Δ 0.0013 0.1080 0.1125 0.0013 0.6441 0.2167 0 0.0063 表 3 右孔测量结果
pixel 参数 1 2 3 4 5 6 7 8 半径R 219.667 219.660 219.659 219.665 219.680 219.640 219.667 219.665 圆心X 671.6 671.5 671.0 672.0 671.6 671.0 671.5 671.5 圆心Y 394.0 394.5 393.5 394.0 394.5 394.5 394.5 393.5 Δ 0.0029 0.2092 0.7061 0.1250 0.1191 0.239 0.1258 0.125 表 4 两种方法对核心孔的测量结果比较
pixel 方法 左孔 右孔 时间T/s X Y R X Y R 传统霍夫变换 653 1268 219 671 394 220 9.64 改进霍夫变换 652.5 1269.5 219.667 671.6 394 219.333 0.02 -
[1] Stroupe A W, Martin M C, Balch T. Distributed sensor fusion for object position estimation by multi-robot systems[C]//Proceedings of 2001 IEEE International Conference on Robotics and Automation. Seoul, South Korea: IEEE, 2001: 1092-1098 [2] 孙丁.新的二值图像八近邻边界跟踪和内外边标定算法[D].上海: 华东师范大学, 2011Sun D. A new contour tracing and inner and outer boundary labeling algorithm in eight-connected binary images[D]. Shanghai: East China Normal University, 2011(in Chinese) [3] Cuevas E, Zaldivar D, Pérez-Cisneros M, et al. Circle detection using discrete differential evolution optimization[J]. Pattern Analysis and Applications, 2011, 14(1):93-107 http://d.old.wanfangdata.com.cn/NSTLQK/NSTL_QKJJ0223146464/ [4] Cuevas E, Oliva D, Zaldivar D, et al. Circle detection using electro-magnetism optimization[J]. Information Sciences, 2012, 182(1):40-55 http://d.old.wanfangdata.com.cn/NSTLQK/NSTL_QKJJ0225807731/ [5] Chung K L, Huang Y H, Shen S M, et al. Efficient sampling strategy and refinement strategy for randomized circle detection[J]. Pattern Recognition, 2012, 45(1):252-263 doi: 10.1016/j.patcog.2011.07.004 [6] 舒龙庆, 曾垂力.一种基于圆的几何特性改进的圆检测随机算法[J].集成技术, 2015, 4(2):46-49 http://d.old.wanfangdata.com.cn/Periodical/jcjx201502006Shu L Q, Zeng C L. An improved randomized algorithm for circle detection based on geometric properties of the circle[J]. Journal of Integration Technology, 2015, 4(2):46-49(in Chinese) http://d.old.wanfangdata.com.cn/Periodical/jcjx201502006 [7] Seo S W, Kim M. Efficient architecture for circle detection using Hough transform[C]//Proceedings of 2015 International Conference on Information and Communication Technology Convergence. Jeju, South Korea: IEEE, 2015: 570-572 [8] Ye H S, Shang G C, Wang L N, et al. A new method based on Hough transform for quick line and circle detection[C]//Proceedings of the 8th International Conference on Biomedical Engineering and Informatics. Shenyang, China: IEEE, 2016: 52-56 [9] Djekoune A O, Messaoudi K, Amara K. Incremental circle Hough transform:an improved method for circle detection[J]. Optik, 2017, 133:17-31 doi: 10.1016/j.ijleo.2016.12.064 [10] Bukowska D M, Chew A L, Huynh E, et al. Semi-automated identification of cones in the human retina using circle Hough transform[J]. Biomedical Optics Express, 2015, 6(12):4676-4693 doi: 10.1364/BOE.6.004676 [11] 陈令刚.基于模板匹配的多圆识别算法[J].电脑知识与技术, 2017, 13(24):173-174, 177 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=dnzsyjs-itrzyksb201724080Chen L G. A algorithm for multiple-circle detection on images using template matching[J]. Computer Knowledge and Technology, 2017, 13(24):173-174, 177(in Chinese) http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=dnzsyjs-itrzyksb201724080 [12] 段黎明, 汪威, 张霞.改进的Hough变换实现圆检测[J].计算机集成制造系统, 2013, 19(9):2148-2152 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=jsjjczzxt201309006Duan L M, Wang W, Zhang X. Circle detection through improved Hough transform[J]. Computer Integrated Manufacturing System, 2013, 19(9):2148-2152(in Chinese) http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=jsjjczzxt201309006 [13] 林金龙, 石青云.用点Hough变换实现圆检测的方法[J].计算机工程, 2003, 29(11):17-18 doi: 10.3969/j.issn.1000-3428.2003.11.007Lin J L, Shi Q Y. Circle recognition through a point Hough transformation[J]. Computer Engineering, 2003, 29(11):17-18(in Chinese) doi: 10.3969/j.issn.1000-3428.2003.11.007 [14] 卜飞宇, 祝青, 王涛.一种改进的基于最大类间方差的二值化方法[J].电脑知识与技术, 2015, 11(5):188-189, 195 http://d.old.wanfangdata.com.cn/Periodical/dnzsyjs-itrzyksb201505088Bu F Y, Zhu Q, Wang T. An improved binarization method based on maximum variance between clusters[J]. Computer Knowledge and Technology, 2015, 11(5):188-189, 195(in Chinese) http://d.old.wanfangdata.com.cn/Periodical/dnzsyjs-itrzyksb201505088 [15] Wu M J, Song Z X, Li B P, et al. A method to detect circle based on Hough transform[C]//Proceedings of the 1st International Conference on Information Sciences, Machinery, Materials and Energy. Chongqing: Atlantis Press, 2015