基于贪心选择及斜率探测扩充的轨面提取方法

doi:10.3778/j.issn.1673-9418.2008003

计算机科学与探索 ›› 2022, Vol. 16 ›› Issue (1): 205-216.DOI: 10.3778/j.issn.1673-9418.2008003

基于贪心选择及斜率探测扩充的轨面提取方法

曹义亲, 刘龙标⁺(), 何恬, 丁要男

华东交通大学软件学院,南昌 330013

收稿日期:2020-08-03 修回日期:2020-09-30 出版日期:2022-01-01 发布日期:2020-10-23
通讯作者: + E-mail: 569144312@qq.com
作者简介:曹义亲（1964—）,男,江西九江人,硕士,教授,硕士生导师,CCF会员,主要研究方向为图像处理、模式识别。
刘龙标（1995—）,男,江西赣州人,硕士研究生,主要研究方向为图像处理。
何恬（1996—）,女,江西南昌人,硕士研究生,主要研究方向为图像处理。
丁要男（1996—）,男,江西上饶人,硕士研究生,主要研究方向为图像处理。
基金资助:
国家自然科学基金(61663009);江西省科技支撑计划重点项目(20161BBE50081)

Method of Rail Surface Extraction Based on Greedy Selection and Slope Detection Expansion

CAO Yiqin, LIU Longbiao⁺(), HE Tian, DING Yaonan

School of Software, East China Jiaotong University, Nanchang 330013, China

Received:2020-08-03 Revised:2020-09-30 Online:2022-01-01 Published:2020-10-23
About author:CAO Yiqin, born in 1964, M.S., professor, M.S. supervisor, member of CCF. His research interests include image processing and pattern recognition.
LIU Longbiao, born in 1995, M.S. candidate. His research interest is image processing.
HE Tian, born in 1996, M.S. candidate. Her research interest is image processing.
DING Yaonan, born in 1996, M.S. candidate. His research interest is image processing.
Supported by:
National Natural Science Foundation of China(61663009);Key Project of Science and Technology Support Plan of Jiangxi Province(20161BBE50081)

摘要/Abstract

摘要：

传统的钢轨表面区域提取方法不同程度存在需要预先给定轨面宽度、假定轨面在轨道图像中央和手动选取边界等前提条件,且存在自适应性差,光照敏感,无法在轨头圆角处存在尘泥等噪声时将轨面完整提取等问题。针对上述问题,提出了一种基于YUV空间的贪心算法选择及斜率探测扩充的轨面区域提取方法。首先将RGB轨道图像转化到YUV空间,提取其V分量,减弱环境光照以及噪声的干扰;其次绘制V分量的灰度投影反转曲线,利用该曲线灰度均值和中值进行候选轨面区间划分;随后利用贪心算法求出划分后曲线中的最大子序和区间,进行轨面粗提取;最后利用斜率探测扩充法进行轨面精提取,在粗提取的边界两侧进行一定距离的斜率探测,用偏转角大于设定阈值的中间位置更新轨面边界。实验结果表明,该方法可以精准快速地提取轨面区域,平均精度达0.929 6,准确率达96.67%,平均时间为25.96 ms,具有一定实用价值。

关键词: 钢轨表面区域, YUV色彩空间, 灰度投影, 贪心算法, 斜率探测扩充

Abstract:

The traditional rail surface area extraction method has some preconditions, for instance, the width of rail surface should be given in advance, the rail surface should be assumed to be in the center of the rail image and the boundary should be selected manually. Besides, it also has some problems, such as poor self-adaptability, light sensitivity, and the inability to extract the rail surface completely when there are noises such as dust and mud at the rounded corners of the rail head. Aiming at the above problems, a method of extracting rail surface area based on YUV space greedy algorithm selection and slope detection expansion is proposed. Firstly, the RGB rail image is converted to YUV space, and its V component is extracted to reduce the interference of ambient light and noise. Secondly, the grayscale projection inversion curve of the V component is drawn, and the gray mean and median of the curve are used to divide the candidate orbital intervals. Then, greedy algorithm is used to calculate the interval of the maximum suborder sum in the divided curve for rough extraction of rail surface. At last, the slope detection expansion method is used to accurately extract the rail surface, the slope detection at a certain distance is carried out on both sides of the coarse-extracted boundary, and the rail surface boundary is updated with the middle position where the deflection angle is greater than the set threshold. Experimental results show that the proposed method can accurately and rapidly extract the rail surface area, with an average precision of 0.9296, an average accuracy of 96.67%, and an average time of 25.96 ms, which is of certain practical value.

Key words: rail surface area, YUV space, grayscale projection, greedy algorithm, slope detection expansion

中图分类号:

TP391.41

曹义亲, 刘龙标, 何恬, 丁要男. 基于贪心选择及斜率探测扩充的轨面提取方法[J]. 计算机科学与探索, 2022, 16(1): 205-216.

CAO Yiqin, LIU Longbiao, HE Tian, DING Yaonan. Method of Rail Surface Extraction Based on Greedy Selection and Slope Detection Expansion[J]. Journal of Frontiers of Computer Science and Technology, 2022, 16(1): 205-216.

图/表 14

图1 实际采集的轨道图像

Fig.1 Rail image collected in practical environment

图2 轨道图像灰度分布图

Fig.2 Gray scale distribution of rail image

图3 轨道图像的三种色彩空间探索

Fig.3 Three color spaces exploration of rail images

图4 轨面区域提取算法流程图

Fig.4 Flow chart of rail surface extraction algorithm

图5 轨道图像的灰度反转投影曲线

Fig.5 Grayscale inversion projection curve of rail image

图6 V分量的灰度反转投影搜索曲线

Fig.6 Grayscale inversion projection search curve of V component

图7 钢轨头部特征分析图

Fig.7 Analysis of rail head characteristics

图8 粗略边界两侧斜率探测分析

Fig.8 Slope detection analysis on both sides of rough boundary

图9 粗略边界右侧斜率探测精分析

Fig.9 Slope analysis on right side of rough boundary

图10 IoU示意图及其简化图

Fig.10 IoU schematic and simplified drawings

图11 IoU示意图及其简化图

Fig.11 IoU schematic and simplified drawings

表1 本文算法提取结果分析

Table 1 Algorithm extraction results analysis

Image type	IoUs	Accuracy/%	Time/ms
RCRH无干扰	0.934 7	100.00	24.17
RCRH有干扰	0.922 7	95.00	26.84
总体	0.926 9	96.67	25.95

表2 不同轨面提取算法性能对比

Table 2 Performance comparison of different rail surface extraction algorithms

算法	Rail_1, 真实边界：[922, 1 504]				Rail_2, 真实边界：[596, 1 282]
算法	预测边界	IoUs	有效性	耗时/ms	预测边界	IoUs	有效性	耗时/ms
HP^[6]	[455, 2 128]	0.347 9	No	47.2	[588, 1 800]	0.561 8	No	52.6
VP^[7]	[1, 311]	0.000 0	No	22.8	[1, 200]	0.000 0	No	18.5
TEBP^[8]	[878, 1 460]	0.859 4	Yes	14.2	[580, 1 266]	0.954 4	Yes	12.8
HPCG^[9]	[878, 1 460]	0.859 4	Yes	18.9	[580, 1 266]	0.954 4	Yes	16.3
CHMM^[11]	[878, 1 460]	0.859 4	Yes	16.3	[580, 1 266]	0.954 4	Yes	13.2
EHLS^[12]	[948, 1 415]	0.802 4	Yes	37.5	[606, 1 197]	0.859 0	Yes	38.4
EHSV^[13]	[917, 1 455]	0.908 0	Yes	28.1	[596, 1 825]	0.557 2	No	29.3
Proposed method	[942, 1 503]	0.963 9	Yes	21.4	[599, 1 282]	0.992 7	Yes	22.9

图12 多种轨面提取算法对比结果

Fig.12 Comparison results of various rail surface extraction algorithms

参考文献 20

[1]	张辉, 宋雅男, 王耀南, 等. 钢轨缺陷无损检测与评估技术综述[J]. 仪器仪表学报, 2019, 40(2):11-25.
	ZHANG H, SONG Y N, WANG Y N, et al. Review of rail defect non-destructive testing and evaluation[J]. Chinese Journal of Scientific Instrument, 2019, 40(2):11-25.
[2]	LI Q Y, ZHONG Z D, LIANG Z P, et al. Rail inspection meets big data: methods and trends[C]// Proceedings of the 18th International Conference on Network-Based Information Systems, Taiwan, China, Sep 2-4, 2015. Washington: IEEE Computer Society, 2015: 302-308.
[3]	ZHANG H, JIN X T, JONATHAN Q M, et al. Automatic visual detection system of railway surface defects with curvature filter and improved Gaussian mixture model[J]. IEEE Transactions on Instrumentation and Measurement, 2018, 67(7):1593-1608. DOI URL
[4]	金侠挺, 王耀南, 张辉, 等. 基于贝叶斯CNN和注意力网络的钢轨表面缺陷检测系统[J]. 自动化学报, 2019, 45(12):2312-2327.
	JIN X T, WANG Y N, ZHANG H, et al. DeepRail: automatic visual detection system for railway surface defect using Bayesian CNN and attention network[J]. Acta Automatica Sinica, 2019, 45(12):2312-2327.
[5]	YU H M, LI Q Y, TAN Y Q, et al. A coarse-to-fine model for rail surface defect detection[J]. IEEE Transactions on Instrumentation & Measurement, 2019, 68(3):656-666.
[6]	唐湘娜, 王耀南. 铁轨表面缺陷的视觉检测与识别算法[J]. 计算机工程, 2013, 39(3):25-30.
	TANG X N, WANG Y N. Visual inspection and classification algorithm of rail surface defect[J]. Computer Engineering, 2013, 39(3):25-30.
[7]	王耀南, 尹逊帅, 贺振东, 等. 钢轨表面图像冗余信息的模糊匹配算法[J]. 湖南大学学报(自然科学版), 2016, 43(4):75-80.
	WANG Y N, YIN X S, HE Z D, et al. Algorithm of fuzzy matching for redundancies of rail surface images[J]. Journal of Hunan University (Natural Sciences), 2016, 43(4):75-80.
[8]	LI Q Y, REN S W. A real-time visual inspection system for discrete surface defects of rail heads[J]. IEEE Transactions on Instrumentation & Measurement, 2012, 61(8):2189-2199.
[9]	WU Y P, LI Q Y, JIA L M. Research on rail surface defect detection method based on UAV images[C]// Proceedings of the 2018 Prognostics and System Health Management Con-ference, Chongqing, Oct 26-28, 2018. Piscataway: IEEE, 2018: 553-558.
[10]	周咏. 基于图像处理的钢轨表面缺陷识别研究[D]. 兰州: 兰州交通大学, 2018.
	ZHOU Y. Research on the defect recognition for rail surface based on image processing[D]. Lanzhou: Lanzhou Jiaotong University, 2018.
[11]	贺振东, 王耀南, 刘洁, 等. 基于背景差分的高铁钢轨表面缺陷图像分割[J]. 仪器仪表学报, 2016, 37(3):640-649.
	HE Z D, WANG Y N, LIU J, et al. Background differencing-based high-speed rail surface defect image segmentation[J]. Chinese Journal of Scientific Instrument, 2016, 37(3):640-649.
[12]	MIN Y Z, XIAO B Y, DANG J W, et al. Real time detection system for rail surface defects based on machine vision[J]. EURASIP Journal on Image and Video Processing, 2018, 1:3.
[13]	顾桂梅, 李晓梅, 常海涛, 等. 基于HSV空间的钢轨表面区域快速提取算法[J]. 云南大学学报(自然科学版), 2019, 41(4):707-717.
	GU G M, LI X M, CHANG H T, et al. Fast extraction agori-thm for rail surface region based on HSV space[J]. Journal of Yunnan University (Natural Sciences Edition), 2019, 41(4):707-717.
[14]	PREMA C E, VINSLEY S S, SURESH S. Multi feature analysis of smoke in YUV color space for early forest fire detection[J]. Fire Technology, 2016, 52(5):1319-1342. DOI URL
[15]	LIU J H, ZHONG X. An object tracking method based on mean shift algorithm with HSV color space and texture features[J]. Cluster Computing, 2019, 22(3):6079-6090. DOI URL
[16]	SUN Y H, MU Y, QIN F, et al. Deer body adaptive threshold segmentation algorithm based on color space[J]. Computers, Materials and Continua, 2020, 64(2):1317-1328. DOI URL
[17]	YANG Y G, ZOU L, ZHOU Y H, et al. Visually meaningful encryption for color images by using Qi hyper-chaotic system and singular value decomposition in YCbCr color space[J]. International Journal for Light and Electron Optics, 2020, 213:164422. DOI URL
[18]	WEI X J, YE P X. Efficiency of orthogonal super greedy algorithm under the restricted isometry property[J]. Journal of Inequalities and Applications, 2019(1):124.
[19]	郅惠博, 李洪涛, 张含露. 国内外铁路用钢轨标准对比分析研究[J]. 科技视界, 2019, 32:1-6.
	ZHI H B, LI H T, ZHANG H L. Comparative analysis and research on railway rail standards at home and abroad[J]. Science & Technology Vision, 2019, 32:1-6.
[20]	GIRSHICK R B. Fast R-CNN[C]// Proceedings of the 2015 IEEE International Conference on Computer Vision, Santiago, Dec 7-13, 2015. Washington: IEEE Computer Society, 2015: 1440-1448.

基于贪心选择及斜率探测扩充的轨面提取方法

Method of Rail Surface Extraction Based on Greedy Selection and Slope Detection Expansion

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 14

参考文献 20

相关文章 3

编辑推荐

Metrics

[1]	张子辰, 岳昆, 祁志卫, 段亮. 时序知识图谱的增量构建[J]. 计算机科学与探索, 2022, 16(3): 598-607.
[2]	庞源, 武继刚, 陈龙, 姚棉阳. 边缘计算中多设备多任务的能耗均衡优化算法[J]. 计算机科学与探索, 2022, 16(2): 480-488.
[3]	孙焕良，富珊珊，刘俊岭，于戈，许鸿斐. 社会网络中弱关系团队形成问题研究[J]. 计算机科学与探索, 2016, 10(6): 773-785.