Track fastener detection method based on decision tree classification and region growth

doi:10.13474/j.cnki.11-2246.2022.0257

Abstract

Abstract: Due to the large amount of data and small size of rail fasteners, the detection of rail fastener diseases is heavy and inefficient. Therefore, a fast rail fastener detection technology is urgently needed. The rapid development of line structure laser technology has the characteristics of high resolution, high precision and high response. In this paper, the line structure laser measurement technology is used to quickly obtain the fine 3D point cloud of rail fasteners, and the decision tree classification and regional growing algorithm are used to analyze the 3D point cloud of the fine fasteners, so as to realize the fast detection of fastener diseases and the calculation of fastener geometric parameters. Using this technology to detect the rail transit fasteners of Guangzhou Line 18, the detection results show that the method has high detection efficiency and accurate detection results.

Key words: track fasteners, line structure laser, 3D point cloud, decision tree, regional growing algorithm

CLC Number:

GAO Hong, WANG Yong, TANG Chao, WANG Xiaojing. Track fastener detection method based on decision tree classification and region growth[J]. Bulletin of Surveying and Mapping, 2022, 0(9): 18-22.

References

[1] DE RUVO P, DISTANTE A, STELLA E, et al. A GPU-based vision system for real time detection of fastening elements in railway inspection[C]//Proceedings of the 16th IEEE International Conference on Image Processing. Cairo, Egypt:IEEE,2019:2333-2336.
[2] RESENDIZ E, HART J M, AHUJA N. Automated visual inspection of railroad tracks[J]. IEEE Transactions on Intelligent Transportation Systems, 2013, 14(2):751-760.
[3] YANG Jinfeng, TAO Wei, LIU Manhua, et al. An efficient direction field-based method for the detection of fasteners on high-speed railways[J]. Sensors (Basel, Switzerland), 2011, 11(8):7364-7381.
[4] 刘甲甲, 李柏林, 罗建桥, 等. 融合PHOG和MSLBP特征的铁路扣件检测算法[J]. 西南交通大学学报, 2015, 50(2):256-263.
[5] MAGICHAN.激光位移传感器原理的两种测量方式[EB/OL]. (2022-04-11)[2022-06-04]. https://www.elecfans.com/news.
[6] 刘妮娜, 王永录. 轨道扣件智能检测系统的设计及应用[J]. 高速铁路技术, 2016, 7(4):55-59.
[7] 李玥华, 周京博, 刘利剑. 线结构光测量技术研究进展[J]. 河北科技大学学报, 2018, 39(2):115-124.
[8] 宋大虎, 李忠科, 程春霞. 基于线结构光的三维坐标测量技术研究[J]. 计算机工程, 2012, 38(22):291-293.
[9] KEYENCE.超高速轮廓测量仪LJ-V7000系列[EB/OL]. (2019-03-11)[2022-06-04]. https://www.keyence.com.cn/products/measure.
[10] 王乐. 线结构光钢轨轮廓全断面测量技术研究[D]. 北京:中国铁道科学研究院, 2021.
[11] 侯海倩, 唐超, 赵丽凤, 等. 轨道扣件病害自动化检测研究及应用[J]. 测绘通报, 2021(8):88-92.
[12] 吴梦, 陈建政, 林森. 基于图像处理的轨道扣件状态检测[J]. 信息与电脑(理论版), 2013(2):46-47.
[13] 崔昊. 基于结构光点云的轨道扣件服役状态检测关键技术研究[D]. 武汉:武汉大学, 2019.
[14] 倪费杰, 林群煦, 闫笑颜, 等. 基于Mask R-CNN算法在轨道扣件缺陷检测中的应用[J]. 机械工程师, 2021(4):20-23.
[15] 朱澎程. 基于纹理特征的快速弹条状态识别的实现[D]. 上海:上海工程技术大学, 2020.
[16] 宁业衍, 方宇, 杨皓, 等. 基于距离判断函数的钢轨扣件分割方法[J]. 智能计算机与应用, 2021, 11(11):64-67.
[17] 熊智敏. 城市轨道服役状态检测系统关键技术研究[D]. 武汉:武汉大学, 2017.
[18] 郑丹阳, 李立明, 孙睿, 等. 基于深度边缘特征的轨道扣件状态检测方法[J]. 铁道建筑, 2021, 61(4):138-142.