基于3D点云语义地图表征的智能车定位

朱云涛; 李飞; 胡钊政; 吴华伟

doi:10.3963/j.jssn.1674-4861.2021.06.017

基于3D点云语义地图表征的智能车定位

doi: 10.3963/j.jssn.1674-4861.2021.06.017

朱云涛^{1, 2,},
李飞¹,
胡钊政^{1, 2, ,},
吴华伟³

1.
武汉理工大学智能交通系统研究中心武汉 430063
2.
武汉理工大学重庆研究院重庆 401120
3.
湖北文理学院汽车与交通工程学院湖北襄阳 441053

基金项目:

国家重点研发计划项目 2018YFB1600801

武汉市科技局项目 2020010601012165

武汉市科技局项目 2020010602011973

武汉市科技局项目 2020010602012003

重庆市自然科学基金项目 cstc2020jcyj-msxmX0978

详细信息

作者简介:
朱云涛(1997—), 硕士研究生. 研究方向: 计算机视觉、激光SLAM. E-mail: zyt941292303@whut.edu.cn

通讯作者:
胡钊政(1979—), 博士, 教授. 研究方向: 计算机视觉、智能车路协同. E-mail: zzhu@whut.edu.cn

中图分类号: U495
计量
- 文章访问数: 929
- HTML全文浏览量: 418
- PDF下载量: 30
- 被引次数: 0
出版历程
- 收稿日期: 2021-09-14
- 网络出版日期: 2022-01-12

A Localization Method for Intelligent Vehicles Based on Semantic Map Representation Extracted from 3D Cloud Points

ZHU Yuntao^{1, 2
,},
LI Fei¹,
HU Zhaozheng^{1, 2
, ,},
WU Huawei³

1.
Intelligent Transportation Systems Research Center, Wuhan University of Technology, Wuhan 430063, China
2.
Chongqing Research Institute, Wuhan University of Technology, Wuhan 430063, China
3.
School of Automotive and Traffic Engineering, Hubei University of Arts and Science, Xiangyang 441053, Hubei, China

摘要

摘要: 为提高智能车节点定位准确率, 研究了基于3D点云语义地图表征的智能车定位方法。该方法分为3个部分: ①基于三维激光点云的语义分割, 包括地面分割, 交通标志牌分割和杆状语义目标分割; ②面向智能车的点云语义地图表征, 利用分割的语义目标投影, 生成带权有向图, 语义路, 语义编码, 再以语义编码和高精度GPS的全局位置组成语义地图表征模型; ③基于语义表征模型的智能车定位, 包括基于GPS匹配的粗定位和基于语义编码渐进匹配的节点定位。实验在3种长度不同、复杂度不同的道路场景下进行, 节点定位准确率分别为98.5%, 97.6%和97.8%, 结果表明所提出的定位方法节点定位准确率高、鲁棒性强且适用于不同的道路场景。
- 智能交通 /
- 节点定位 /
- 地图表征 /
- 语义分割
Abstract: An intelligent vehicle localization method based on the semantic-map representation of 3D point clouds is proposed to improve the accuracy of node localization for intelligent vehicles. The method is divided into three parts. ① Semantic segmentation based on 3D laser point clouds includes segmentation of ground, traffic sign, and pole-shaped targets. ② Semantic-map representation for intelligent vehicles uses segmented targets to project. Directional projections with weight, semantic roads, and semantic coding are generated. The coding and global location from high-precision GPS make up the representation model. ③ Localization based on a semantic representation model includes coarse positioning from GPS matching and node localization from semantic coding matching. The experiments are performed in three road scenes with different lengths and complexities, and their localization accuracy is 98.5%, 97.6%, and 97.8%, respectively. The results show that the proposed method has high accuracy and strong robustness, suitable for different road scenes.
- intelligent transportation /
- node localization /
- map representation /
- semantic segmentation

HTML全文

图 1 系统流程图

Figure 1. Flow of the system

下载: 全尺寸图片幻灯片

图 2 俯仰角评估示意图

Figure 2. Schematic of pitch angle evaluation

下载: 全尺寸图片幻灯片

图 3 交通标志牌粗分割

Figure 3. Coarse segmentation of traffic signs

下载: 全尺寸图片幻灯片

图 4 交通标志牌精分割

Figure 4. Precise segmentation of traffic signs

下载: 全尺寸图片幻灯片

图 5 切片生长法分割

Figure 5. Segmentation of slice growth

下载: 全尺寸图片幻灯片

图 6 点云语义地图模型

Figure 6. Semantic map modeling from point clouds

下载: 全尺寸图片幻灯片

图 7 语义俯视投影图

Figure 7. Semantic overhead projection

下载: 全尺寸图片幻灯片

图 8 语义带权有向图

Figure 8. Semantically weighted digraph

下载: 全尺寸图片幻灯片

图 9 场景语义编码

Figure 9. Scene semantic coding

下载: 全尺寸图片幻灯片

图 10 GPS粗定位

Figure 10. GPS coarse positioning

下载: 全尺寸图片幻灯片

图 11 节点级定位

Figure 11. Node-level Localization

下载: 全尺寸图片幻灯片

图 12 实验平台和数据集路线

Figure 12. Experimental platform and data set route

下载: 全尺寸图片幻灯片

图 13 节点的地面分割结果

Figure 13. Ground segmentation result of a node

下载: 全尺寸图片幻灯片

图 14 节点的交通标志牌分割结果

Figure 14. Traffic-sign segmentation of a node

下载: 全尺寸图片幻灯片

图 15 节点的杆状语义分割结果

Figure 15. Pole-shaped object segmentation result of a node

下载: 全尺寸图片幻灯片

图 16 地图节点全局位置轨迹

Figure 16. Global positional trajectory of map nodes

下载: 全尺寸图片幻灯片

表 1 数据集1~3交通标志牌语义分割对比

Table 1. Comparative experiment results of data set 1~3

数据集	方法	TP_s/个	FP_s/个	FN_s/个	precision_s/%	complete_s/%
1	本文	25	1	1	96.2	96.2
1	文献[17]	24	1	2	96.0	92.3
2	本文	40	1	2	97.6	95.2
2	文献[17]	39	2	3	95.1	92.9
3	本文	54	4	3	92.9	94.7
3	文献[17]	52	5	5	91.2	91.2

下载: 导出CSV

表 2 数据集1~3杆状语义分割对比

Table 2. Comparative experiment results of data set 1~3

数据集	方法	TP_p/个	FP_p/个	FN_p/个	precision_p/%	complete_p/%
1	本文	127	3	5	97.7	96.2
1	文献[18]	120	4	12	96.8	90.1
2	本文	202	5	11	97.6	94.8
2	文献[18]	187	2	26	95.9	87.8
3	本文	310	11	16	96.6	96.2
3	文献[18]	283	17	43	94.3	86.8

下载: 导出CSV

表 3 数据集1~3定位精度对比

Table 3. Comparative experiment results of data set 1~3

数据集	方法	正确个数	错误个数	准确率/%
1	本文	578	9	98.5
1	文献[19]	431	156	73.4
2	本文	969	24	97.6
2	文献[19]	669	324	67.4
3	本文	1504	34	97.8
3	文献[19]	1101	437	71.6

下载: 导出CSV

参考文献(19)

[1]	张帆, 胡钊政, 陈佳良, 等. 基于Wi-Fi指纹与视觉融合的室内交通定位[J]. 交通信息与安全, 2019, 37(3): 61-69+100. doi: 10.3963/j.issn.1674-4861.2019.03.008 ZHANG Fan, HU Zhaozheng, CHEN Jialiang, et al. Indoor traffic positioning based on Wifi fingerprint and vision fusion[J]. Journal of Transport Information and Safety, 2019, 37(3): 61-69+100. (in Chinese). doi: 10.3963/j.issn.1674-4861.2019.03.008
[2]	WOLCOTT R W, EUSTICE R M. Robust LiDAR localization using multiresolution Gaussian mixture maps for autonomous driving[J]. The International Journal of Robotics Research, 2017, 36(3): 292-319. doi: 10.1177/0278364917696568
[3]	MUR-ARTAL R, TARDOS J D. ORB-SLAM2: An open-source SLAM system for monocular stereo and RGB-D cameras[J]. IEEE Transactions on Robotics, 2017, 33(5): 1255-1262. doi: 10.1109/TRO.2017.2705103
[4]	杨东凯, 寇艳红, 吴今培, 等. 智能交通系统中的地图匹配定位方法[J]. 交通运输系统工程与信息, 2003(3): 38-43. doi: 10.3969/j.issn.1009-6744.2003.03.008 YANG Dongkai, KOU Yanhong, WU Jinpei, et al. A map-matching method in intelligent transport systems[J]. Journal of Transportation Systems Engineering and Information Technology, 2003, 3(3): 38-43. (in Chinese). doi: 10.3969/j.issn.1009-6744.2003.03.008
[5]	ZHANG Ji, SINGH S. LOAM: LiDAR odometry and mapping in real-time[C]. Robotics: Science and Systems Conference, Berkeley, California, USA: University of California, Berkeley, 2014.
[6]	KOIDE K. A Portable 3D LIDAR-based system for long-term and wide-area people behavior measurement[J]. International Journal of Advanced Robotic Systems, 2019, 16(2): 1-13. http://www.researchgate.net/publication/331224140_A_Portable_3D_LIDAR-based_System_for_Long-term_and_Wide-area_People_Behavior_Measurement
[7]	THRUN S, MONTEMERLO M, Dahlkamp H, et al. Stanley: The robot that won the DARPA grand challenge[J]. Journal of Field Robotics, 2006, 23(9): 661-692. doi: 10.1002/rob.20147
[8]	ZERMAS D, IZZAT I, PAPANIKOLOPOULOS N. Fast segmentation of 3D point clouds: A paradigm on LiDAR data for autonomous vehicle applications[C]. International Conference on Robotics and Automation(ICRA), Singapore: IEEE, 2017.
[9]	彭泽民, 叶青. 自然环境下交通标志牌的检测与识别[J]. 电子测试, 2020(1): 45-47+92. https://www.cnki.com.cn/Article/CJFDTOTAL-WDZC202001015.htm PENG Zemin, YE Qing. Detection and identification of traffic signs in natural environment[J]. Electronic Test, 2020(1): 45-47+92. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-WDZC202001015.htm
[10]	李游. 基于车载激光扫描数据的城市街道信息提取技术研究[D]. 武汉: 武汉大学, 2017. LI You. Research on street information detection from mobile laser scanning data in urban areas[D]. Wuhan: Wuhan University, 2017. (in Chinese).
[11]	KUZNETSOVA K, SAMSONOVICH A V. Semantic-map-base dapproach to designing an insight problem solving assistant[J]. Procedia Computer Science, 2018(123): 258-264. http://www.sciencedirect.com/science/article/pii/S1877050918300413/pdf?md5=139ba648e1b132f94671dd265cc50796&pid=1-s2.0-S1877050918300413-main.pdf
[12]	DUBE R, DUGAS D, STUMM E, et al. SegMatch: Segment based place recognition in 3D point clouds[C]. International Conference on Robotics and Automation(ICRA), Singapore: IEEE, 2017.
[13]	LIU Yu, PETILLOT Y, LANE D, et al. Global localization with object-level semantics and topology[C]. International Conference on Robotics and Automation(ICRA), Montreal, QC, Canada: IEEE, 2019.
[14]	CHEN X, MILIOTO A, PALAZZOLO E, et al. SuMa++: Efficient LiDAR-based semantic SLAM[C]. International Conference on Intelligent Robots and Systems(IROS), Macau, China: IEEE, 2019.
[15]	GOLOVINSKIY A, FUNKHOUSER T. Min-cut based segmentation of point clouds[C]. IEEE International Conference on Computer Vision Workshops, Kyoto, Japan: IEEE, 2009.
[16]	刘亚坤, 李永强, 刘会云, 等. 基于改进RANSAC算法的复杂建筑物屋顶点云分割[J]. 地球信息科学学报, 2021, 23(8): 1497-1507. https://www.cnki.com.cn/Article/CJFDTOTAL-DQXX202108014.htm LIU Yakun, LI Yongqiang, LIU Huiyun, et al. An improved RANSAC algorithm for point cloud segmentation of complex building roofs[J]. Journal of Geo-information Science, 2021, 23(8): 1497-1507. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-DQXX202108014.htm
[17]	AI C, TSAI Y J. Critical assessment of an enhanced traffic sign detection method using mobile LiDAR and INS technologies[J]. Journal of Transportation Engineering, 2015, 141(5): 1-12.
[18]	PU Shi, RUTZINGER M, VOSSELMAN G, et al. Recognizing basic structures from mobile laser scanning data for road inventory studies[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2011, 66(6): S28-S39. doi: 10.1016/j.isprsjprs.2011.08.006
[19]	TAO Qianwen, HU Zhaozheng, HUANG Gang, et al. LiDAR-only vehicle localization based on map generation[C]. Transportation Research Board(TRB), 2019 Annual Meeting, Washington, D. C., America: National Academies of USA, 2019.