Please wait a minute...
浙江大学学报(工学版)
浙江大学学报(工学版)  2022, Vol. 56 Issue (10): 2066-2076    DOI: 10.3785/j.issn.1008-973X.2022.10.018
土木工程、交通工程、海洋工程     
顾及空间关系的3D LiDAR铁路支持装置自动提取
杨元维1,2,3(),张跃1,高贤君1,马冰洁1,郭申奥1,许磊4
1. 长江大学 地球科学学院，湖北 武汉 430100
2. 湖南科技大学 测绘遥感信息工程湖南省重点实验室，湖南 湘潭 411201
3. 北京市测绘设计研究院 城市空间信息工程北京市重点实验室，北京 100045
4. 中国铁路设计集团有限公司，天津 300251
Automatic extraction of spatially relevant 3D LiDAR railway support facility
Yuan-wei YANG1,2,3(),Yue ZHANG1,Xian-jun GAO1,Bing-jie MA1,Shen-ao GUO1,Lei XU4
1. School of Geosciences, Yangtze University, Wuhan 430100, China
2. Hunan Provincial Key Laboratory of Geo-Information Engineering in Surveying, Mapping and Remote Sensing, Hunan University of Science and Technology, Xiangtan 411201, China
3. Beijing Key Laboratory of Urban Spatial Information Engineering, Beijing Institute of Surveying and Mapping, Beijing 100045, China
4. China Railway Design Corporation Limited Company, Tianjin 300251, China
 全文: PDF(2483 KB)   HTML
摘要：

为了提高不同类型支持装置在不同场景中的提取可靠性及精度，提出顾及空间关系的3D LiDAR支持装置自动提取算法. 通过对GNSS轨迹点抽稀处理获取关键轨迹点，对原始数据进行分层、分块处理，获取原始点云支柱，支持装置数据区域. 通过邻域搜索获取支柱中心点，依据其与轨迹点和支持装置搜索层的空间关系构建空间索引. 以该索引为驱动实现支持装置的初提取，获取含有接触线的支持装置，引入柱状搜索、参数化投影滤波的方式滤除接触线点云，实现提取结果的优化. 经过测试可知，该算法对提取支持装置的MIoU均超过93%，Dice系数均超过94%，可以兼顾多类型支持装置，具有较强的鲁棒性和应用价值.
关键词： 接触网检测;  空间索引;  三维激光点云;  支持装置提取;  点云邻域搜索    
Abstract:

An automatic extraction algorithm of 3D LiDAR support facilities considering the spatial relationship was proposed in order to improve the extraction reliability and accuracy of multi-type support facilities. The key trajectory points were obtained by thinning the GNSS trajectory points. Then the original data was layered and divided into blocks to obtain the original point cloud pillar and support device data area. The center point of the pillar was obtained through the neighborhood search, and a spatial index was constructed based on the spatial relationship between the trajectory point and the search layer of the support device. The index was used as the driving force to realize the initial extraction of the support device and obtain the support device containing the contact line. The column search and parametric projection filtering were introduced to filter out the contact line point cloud for optimization of extraction results. The test results show that the algorithm has more than 93% MIoU and 94% Dice coefficient for extracting support facilities, which can consider multiple types of support facilities and has strong robustness and application value.
Key words: detection of catenary    spatial index    3D LiDAR    support facility extraction    point cloud neighborhood search
收稿日期: 2022-01-13 出版日期: 2022-10-25
CLC:  U 225  
基金资助: 城市轨道交通数字化建设与测评技术国家工程实验室开放课题基金资助项目(2021ZH02)；湖南科技大学测绘遥感信息工程湖南省重点实验室开放基金资助项目(E22133)；城市空间信息工程北京市重点实验室经费资助项目(20210205)；海南省地球观测重点实验室开放基金资助项目(2020LDE001)；长江大学大学生创新创业训练计划资助项目(Dk2021013)
作者简介: 杨元维（1983—），男，副教授，从事接触网检测、三维点云数据处理与分析的研究. orcid.org/0000-0002-4221-4563. E?mail： yyw_08@163.com
服务　　
把本文推荐给朋友
加入引用管理器
E-mail Alert
作者相关文章　　
杨元维
张跃
高贤君
马冰洁
郭申奥
许磊

引用本文:

杨元维,张跃,高贤君,马冰洁,郭申奥,许磊. 顾及空间关系的3D LiDAR铁路支持装置自动提取[J]. 浙江大学学报(工学版), 2022, 56(10): 2066-2076.

Yuan-wei YANG,Yue ZHANG,Xian-jun GAO,Bing-jie MA,Shen-ao GUO,Lei XU. Automatic extraction of spatially relevant 3D LiDAR railway support facility. Journal of ZheJiang University (Engineering Science), 2022, 56(10): 2066-2076.

链接本文:

https://www.zjujournals.com/eng/CN/10.3785/j.issn.1008-973X.2022.10.018    或    https://www.zjujournals.com/eng/CN/Y2022/V56/I10/2066

图 1  点云数据采集装备
参数 数值
测线扫描频率 50 ~ 200 Hz
点扫描速率 最大为101. 6 万点/s
角度分辨率/角度精度 0. 0088°/0. 02° RMS
垂直视角 360°
测程 0. 3 ~ 119 m
线性误差 ≤1 mm
载体移动速度 25 m/h
表 1  Z+F Profiler 9012测量系统的具体参数
图 2  支持装置提取的框架图
图 3  轨迹点抽稀处理的示意图
图 4  仿射变换过程示意图
图 5  分层分块处理的示意图
图 6  支柱中心点提取的示意图
图 7  球形搜索与柱体搜索的示意图
图 8  粗裁剪的示意图
图 9  接触线滤除原理的示意图
n Dis/m Gsf t/s
5 20 1∶1 1030
6 24 5∶6 1010
7 28 5∶6 990
8 32 2∶3 990
9 36 2∶3 980
10 40 1∶2 980
11 44 1∶2 980
12 48 1∶2 970
13 52 1∶3 960
14 56 1∶3 960
15 60 1∶3 940
表 2  抽稀值的实验结果表
图 10  接触网纵断面目标物体分级图
图 11  支持装置的提取效果
图 12  测试精度与线性装置阈值折线的对比图
图 13  不同类型支持装置的提取结果对比图
图 14  提出算法与对比算法的精度对比图
支持装置类别 t/s
棘轮双支持装置 14
棘轮单支持装置 10
双支持装置 12
单支持装置 2
钢架支持装置 51
回路支持装置 2
表 3  多类型支持装置的耗时分析表
图 15  不同情形支柱分布的提取效果展示
本文实验 对比实验
支柱分布 IoU/% Dice/% MIoU/% 支柱分布 IoU/% Dice/% MIoU/%
相邻支柱分布，支柱1 96.99 95.76 97.43 相邻支柱分布，支柱1 87.21 93.15 78.68
相邻支柱分布，支柱2 98.47 97.83 97.43 相邻支柱分布，支柱2 70.14 82.45 78.68
非对称支柱分布，支柱1 95.67 97.79 96.05 非对称支柱分布，支柱1 87.92 93.55 76.67
非对称支柱分布，支柱2 97.89 98.93 96.05 非对称支柱分布，支柱2 79.68 88.65 76.67
非对称支柱分布，支柱3 94.59 97.22 96.05 非对称支柱分布，支柱3 62.42 75.63 76.67
对称支柱分布，支柱1 96.65 98.29 94.66 对称支柱分布，支柱1 85.78 92.34 70.81
对称支柱分布，支柱2 96.43 98.18 94.66 对称支柱分布，支柱2 83.91 91.24 70.81
对称支柱分布，支柱3 95.33 97.61 94.66 对称支柱分布，支柱3 55.45 71.34 70.81
对称支柱分布，支柱4 90.24 94.87 94.66 对称支柱分布，支柱4 58.09 73.49 70.81
表 4  支柱分布不同情况对比的测试结果
图 16  10公里真实铁路场景提取效果图(部分)
1 宋章, 张广泽, 蒋良文, 等 川藏铁路主要地质灾害特征及地质选线探析[J]. 铁道标准设计, 2016, 60 (1): 14- 19
SONG Zhang, ZHANG Guang-ze, JIANG Liang-wen, et al Analysis of the characteristics of major geological disasters and geological alignment of Sichuan-Tibet railway[J]. Railway Sta-ndard Design, 2016, 60 (1): 14- 19
2 AGGARWAL R K, JOHNS A T, JAYASINGHE J A S B An overview of the condition monitoring of overhead lines[J]. Electric Power Systems Research, 2000, 53 (1): 15- 22
3 牛英杰, 苏燕辰, 程敦诚, 等 高铁接触网U型抱箍螺母故障检测算法[J]. 浙江大学学报: 工学版, 2021, 55 (10): 1912- 1921
NIU Ying-jie, SU Yan-chen, CHENG Dun-cheng, et al High-speed rail contact network U-holding nut fault detection algorithm[J]. Journal of Zhejiang University: Engineering Science, 2021, 55 (10): 1912- 1921
4 张国山, 凌朝清, 王欣博, 等 接触线几何参数图像检测系统设计[J]. 天津工业大学学报, 2014, 33 (5): 57- 62
ZHANG Guo-shan, LING Chao-qing, WANG Xin-bo, et al Image detection system design for geometry parameters of contact line[J]. Journal of TianGong University, 2014, 33 (5): 57- 62
5 MOSTAFA A, SANDER O E Application of template matching for improving classification of urban railroad point clouds[J]. Sensors, 2016, 16 (12): 2112
doi: 10.3390/s16122112
6 韩志伟, 刘志刚, 张桂南, 等 非接触式弓网图像检测技术研究综述[J]. 铁道学报, 2013, 35 (6): 40- 47
HAN Zhi-wei, LIU Zhi-gang, ZHANG Gui-nan, et al Summary of research on non-contact pantograph[J]. Railway Society, 2013, 35 (6): 40- 47
7 LOU Y, TIAN Z, JIAN T, et al A fast algorithm for rail extraction using mobile laser scanning data[J]. Remote Sensing, 2018, 10 (12): 1998
8 WANG Y J, CHEN Q, ZHU Z Y, et al A survey of mobile laser scanning applications and key techniques over urban areas[J]. Remote Sensing, 2019, 11 (13): 1540
doi: 10.3390/rs11131540
9 GUAN H, LI J, YU Y, et al Automated road information extraction from mobile laser scanning data[J]. IEEE Transactions on Intelligent Transportation Systems, 2015, 16 (1): 194- 205
doi: 10.1109/TITS.2014.2328589
10 李小路, 曾晶晶, 王皓, 等 三维扫描激光雷达系统设计及实时成像技术[J]. 红外与激光工程, 2019, 48 (5): 35- 42
LI Xiao-lu, ZENG Jing-jing, WANG Hao, et al Design and real-time imaging technology of three-dimensional scanning LiDAR[J]. Infrared and Laser Engineering, 2019, 48 (5): 35- 42
11 刘文强, 刘志刚, 张桂南, 等 基于摄像机标定与卡尔曼滤波的接触网几何参数检测值修正[J]. 铁道学报, 2014, 36 (9): 28- 33
LIU Wen-qiang, LIU Zhi-gang, ZHANG Gui-nan, et al Correction of detected values of catenary geometric parameters based on camera calibration and Kalman filtering[J]. Journal of the China Railway Society, 2014, 36 (9): 28- 33
12 陈国翠, 顾桂梅, 余晓宁, 等 融合PHOG和 BOW-SURF特征的接触网绝缘子缺陷检测方法[J]. 小型微型计算机系统, 2021, 42 (1): 172- 177
CHEN Guo-cui, GU Gui-mei, YU Xiao-ning, et al Method for defect detection of catenary insulators by combining the features of PHOG and BOW-SURF[J]. Journal of Chinese Computer Systems, 2021, 42 (1): 172- 177
13 郭保青, 余祖俊, 张楠, 等 铁路场景三维点云 分割与分类识别算法[J]. 仪器仪表学报, 2017, (9): 2103- 2111
GUO Bao-qing, YU Zu-jun, ZHANG Nan, et al 3D point cloud segmentation, classification and recognition algorithm of railway scene[J]. Chinese Journal of Scientific Instrument, 2017, (9): 2103- 2111
doi: 10.3969/j.issn.0254-3087.2017.09.002
14 PASTUCHA E Catenary system detection, localization and classification using mobile scanning data[J]. Remote Sensing, 2016, 8 (10): 801
15 LAMAS D, SOILÁN M, GRANDÍO J, et al Automatic point cloud semantic segmentation of complex railway environments[J]. Remote Sensing, 2021, 13 (12): 2332
16 JUNG J, CHEN L, SOHN G, et al Multi-range conditional random field for classifying railway electrification system objects using mobile laser scanning data[J]. Remote Sensing, 2016, 8 (12): 1008
doi: 10.3390/rs8121008
17 YUAN Y, CHEN X, WANG J Object-contextual representations for semantic segmentation[J]. European Conference on Computer Vision, 2020, (11): 173- 190
18 CHEN L, XU C, LIN S, et al A deep learning-based method for overhead contact system component recognition using mobile 2D LiDAR[J]. Sensors, 2020, 20 (4): 2224
doi: 10.1109/JSEN.2019.2949146
19 LIN S, XU C, CHEN L, et al LiDAR point cloud recognition of overhead catenary system with deep learning[J]. Sensors, 2020, 20 (4): 2212
20 QI C R, SU H, MO K, et al. PointNet: deep learning on point sets for 3D classification and segmentation [C]// 2017 IEEE Conference on Computer Vision and Pattern Recognition. Hawaii: IEEE, 2017: 77-85.
21 费立凡, 何津, 马晨燕, 等 3维Douglas-Peucker算法及其在DEM自动综合中的应用研究[J]. 测绘学报, 2006, 35 (3): 278- 284
FEI Li-fan, HE Jin, MA Chen-yan, et al Three dimensional Douglas-Peucker algorithm and the study of its application to automated generalization of DEM[J]. Acta Geodaetica et Cartographica Sinica, 2006, 35 (3): 278- 284
doi: 10.3321/j.issn:1001-1595.2006.03.016
[1] 谭文垦, 王长虹, 石忆邵. 基于XML的数字地下空间索引QR树研究[J]. J4, 2009, 43(09): 1615-1620.