基于机器视觉的鞋面特征点自动识别改进方法

doi:10.13475/j.fzxb.20180401607

摘要/Abstract

摘要：

针对目前人工识别鞋面特征点方法实时性差,效率低,成本高的问题,提出一种基于机器视觉的鞋面特征点自动识别改进方法。首先,采用改进中值滤波法对采集图像进行预处理消除噪声干扰;其次,运用提出的自适应阈值分割法提取特征点关键区域;最后通过图像形态学处理和计算最小外接圆完成特征点的自动识别。为验证该方法的可靠性,在光强变化和非常规条件下对大量鞋面样本进行分组实验,并与传统一维和二维Otsu算法的检测结果进行对比。结果表明,该方法在多种复杂环境下具有更好的识别精度和鲁棒性,识别成功率在93%以上,且检测时间不超过0.5 s,可满足工业生产中的精度和实时性需求。

关键词: 鞋面识别, 机器视觉, 改进中值滤波, 自适应阈值分割法

Abstract:

Focusing on the problems of poor real-time,low efficiency and high cost of artificial recognition in vamps feature points,an improved method was proposed to automatically recognize the feature points of vamps by machine vision technology. Firstly,an improved median filter was used for preprocessing the grabbed images to eliminate noise interference. Secondly,by using the proposed adaptive threshold segmentation method, key regions of feature points were extracted. Finally,by morphological image processing and calculating the minimum circumscribed circle, the automatic identification of feature points was completed. In order to verify the reliability of the proposed method,group experiments were carried out on a large number of vamps samples under the condition of light intensity change and clutter,and the results were compared with the conventional one-dimensional and two-dimensional Otsu algorithm. The experimental results show that this method has better recognition accuracy and robustness in a variety of complex environments, the recognition success rate is above 93%, and the detection time is shorter than 0.5 s,which meets the demand of precision and real-time in industrial production.

Key words: vamp recognition, machine vision, improved median filter, adaptive threshold segmentation

中图分类号:

TS101

徐洋, 朱治潮, 盛晓伟, 余智祺, 孙以泽. 基于机器视觉的鞋面特征点自动识别改进方法[J]. 纺织学报, 2019, 40(03): 168-174.

XU Yang, ZHU Zhichao, SHENG Xiaowei, YU Zhiqi, SUN Yize. Improvement recognition method of vamp's feature points based on machine vision[J]. Journal of Textile Research, 2019, 40(03): 168-174.

图/表 12

图1

图2

图3

图4

图5

表1

表2

图6

图7

图8

表3

图9

参考文献 12

[1]	李慧德. 基于机器视觉的在线尺寸测量方法研究[D]. 天津:天津理工大学, 2014: 1-4.
	Ll Huide. Research of the online dimensional measu-rement based on machine vision[D]. Tianjin: Tianjin University of Technology, 2014: 1-4.
[2]	李文羽, 程隆棣. 基于机器视觉和图像处理的织物疵点检测研究新进展[J]. 纺织学报, 2014,35(3):158-164.
	LI Wenyu, CHENG Longdi. New progress of fabric defect detection based on computer vision and image processing[J]. Journal of Textile Research, 2014,35(3):158-164.
[3]	张开玉, 梁凤梅. 基于改进SURF的图像配准关键算法研究[J]. 科学技术与工程, 2013,13(10):2875-2879.
	ZHANG Kaiyu, LIANG Fengmei. Research on the key algorithm for image matching based on improved SURF[J]. Science Technology and Engineering, 2013,13(10):2875-2879.
[4]	SINHA S N, FRAHM J M, POLLEFEYS M, et al. Feature tracking and matching in video using programm able graphics hardware E[J]. Machine Vision & Applications, 2011,22(1):207-217.
[5]	章毓晋. 图象分割评价技术分类和比较[J]. 中国图象图形学报, 1996,1(2):151-158.
	ZHANG Yujin. A classification and comparison of evalution techniques for image segmentation[J]. Journal of Image and Graphic, 1996,1(2):151-158.
[6]	张连宽, PAUL Weckler, 肖德琴. 作物叶面图像自动分割方法[J]. 江苏大学学报(自然科学版), 2016,37(6):663-669.
	ZHANG Liankuan, PAUL Weckler, XIAO Deqin. Automtic leaf surface region segmentation from crop image[J]. Journal of Jiangsu University(Natural Science Edition), 2016,37(6):663-669.
[7]	李伟涛, 彭道黎, 吴见. 基于改进边缘分割算法的幼苗信息提取[J]. 农业机械学报, 2014,45(4):259-263.
	LI Weitao, PENG Daoli, WU Jian. Extraction of seedlings information based on improved edge segmentation algorithm[J]. Transactions of The Chinese Society for Agricultural Machinery, 2014,45(4):259-263.
[8]	OTSU N. A threshold selection method from gray-level histograms[J]. IEEE Transactions on Systems Man & Cybernetics, 2007,9(1):62-66.
[9]	CHEN Q, ZHAO L, LU J, et al. Modified two-dimensional Otsu image segmentation algorithm and fast realisation[J]. Iet Image Processing, 2012,6(4):426-433. doi: 10.1049/iet-ipr.2010.0078
[10]	赵于前, 杨元, 王琨. 基于模糊集理论的迭代多值化图像分割[J]. 光电子·激光, 2009,20(10):1403-1409.
	ZHAO Yuqian, YANG Yuan, WANG Kun. Iterative multi-level image segmentation based on fuzzy set theory[J]. Journal of Optoelectronics Laser, 2009,20(10):1403-1409.
[11]	KO S J, LEE S J. Center weighted median filter and their applications to image enhancement[J]. IEEE Transactions on Circuits and System, 1991,38(1):984-993. doi: 10.1109/31.83870
[12]	张丽, 陈志强, 高文焕, 等. 均值加速的快速中值滤波算法[J]. 清华大学学报(自然科学版), 2004,44(9):1157-1159.
	ZHANG Li, CHEN Zhiqiang, GAO Wenhuan, et al. Mean-based fast median filter[J]. Journal of Tsinghua University (Natural Science Edition), 2004,44(9):1157-1159.

相关文章 15

[1]	朱世根, 杨宏贤, 白云峰, 丁浩, 朱巧莲. 长条状细薄带钩零件变形自动检测系统[J]. 纺织学报, 2020, 41(10): 158-163.
[2]	张建新, 李琦. 基于机器视觉的筒子纱密度在线检测系统[J]. 纺织学报, 2020, 41(06): 141-146.
[3]	路浩, 陈原. 基于机器视觉的碳纤维预浸料表面缺陷检测方法[J]. 纺织学报, 2020, 41(04): 51-57.
[4]	王文胜, 李天剑, 冉宇辰, 卢影, 黄民. 筒子纱纱笼纱杆的定位检测方法[J]. 纺织学报, 2020, 41(03): 160-167.
[5]	金守峰, 林强强, 马秋瑞, 张浩. 基于BP 神经网络的织物表面绒毛质量的检测方法[J]. 纺织学报, 2020, 41(02): 69-76.
[6]	景军锋, 张君扬, 张缓缓, 苏泽斌. 梯度空间下的丝饼表面缺陷检测[J]. 纺织学报, 2020, 41(02): 44-51.
[7]	孙卫红, 阮棉奖, 邵铁锋, 梁曼. 基于机器视觉的生丝抱合性能检测方法[J]. 纺织学报, 2019, 40(08): 164-168.
[8]	景军锋, 张星星. 基于机器视觉的玻璃纤维管纱毛羽检测[J]. 纺织学报, 2019, 40(05): 157-162.
[9]	景军锋, 郭根. 基于机器视觉的丝饼毛羽检测[J]. 纺织学报, 2019, 40(01): 147-152.
[10]	牟新刚蔡逸超周晓陈国良. 基于机器视觉的筒子纱缺陷在线检测系统[J]. 纺织学报, 2018, 39(01): 139-145.
[11]	杨松林马帅丁朝鹏范红丽薛欢欢. 应用机器视觉的织物表面绒毛率测试系统[J]. 纺织学报, 2017, 38(06): 118-123.
[12]	蒋双歌王蕾王静刘建立高卫东. 织物折皱回复的各向异性表征[J]. 纺织学报, 2016, 37(06): 42-47.
[13]	李文羽程隆棣. 基于机器视觉和图像处理的织物疵点检测研究新进展[J]. 纺织学报, 2014, 35(3): 158-0.
[14]	王蕾, 刘建立, 高卫东. 基于机器视觉的织物洗涤后尺寸稳定性评价[J]. 纺织学报, 2012, 33(7): 43-47.
[15]	霍铖宇华强卢达范龙保李新徐刚. 基于虚拟仪器的针织机械用高速电磁阀测试系统[J]. 纺织学报, 2011, 32(10): 134-0.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed

中值滤波窗口大小/像素	滤波方法	鞋面数量/ 张	平均滤波时间/s
3	传统中值滤波	200	0.065
3	本文改进中值滤波	200	0.062
5	传统中值滤波	200	0.088
5	本文改进中值滤波	200	0.071

光照强度/ lx	阈值算法	鞋面数量/ 张	识别成功/ 张	识别成功率/%	平均识别时间/s
	一维Otsu	200	35	17.5	0.18
30	二维Otsu	200	190	95.0	0.32
	本文算法	200	188	94.0	0.17
	一维Otsu	200	146	73.0	0.18
35	二维Otsu	200	192	96.0	0.35
	本文算法	200	189	94.5	0.21
	一维Otsu	200	158	79.0	0.20
40	二维Otsu	200	197	98.5	0.35
	本文算法	200	195	97.5	0.19

混乱条件	鞋面数量/张	识别成功数量/张	识别成功率/%	平均识别时间/s
倾斜30%以内	200	190	93	0.22
正反颠倒	200	185	93	0.21
添加0.05椒盐噪声	200	190	95	0.18