视觉点云质量优化支持的道路杆状物变化检测

doi:10.11947/j.AGCS.2026.20250310

摘要/Abstract

摘要：

视觉传感器是目前道路众源变化检测中最常见的感知传感器。针对视觉SLAM在道路杆状物众源变化检测中的精度与稳健性瓶颈，本文提出一套视觉点云质量优化支持的道路杆状物变化检测技术框架。首先，通过融合语义约束、激光点云深度与GNSS全局校正，构建视觉点云优化方法，显著提升了轨迹精度与点云质量；其次，基于优化后的视觉点云，实现道路杆状物的精准提取与定位；然后，引入基于哈希映射的快速匹配策略，完成跨期杆状物的鲁棒变化检测；最后，在同济大学周边采集的两期试验数据上验证了整体流程的有效性。试验表明，本文方法的轨迹平均绝对误差（MAE）与均方根误差（RMSE）分别降低68.39%和65.65%，点云密度平均提升57.97%。在要素定位与变化检测任务中，杆状物要素的平均定位误差在单一方向上为2~3 m，平均平面误差约为3.5 m，杆状物匹配精度达到94.8%，新增与移除杆状物的变化检测准确率与召回率均为100%。结果证明了本文方法的有效性和稳定性，证实了多源数据融合与优化技术在道路杆状物变化检测中的应用价值，为高精地图变化检测与道路设施智能化管理提供了可靠的技术路径。

关键词: 多源数据, 视觉点云, 道路杆状物, 变化检测

Abstract:

Visual sensors are currently the most common perception sensors for crowdsourced road change detection. However, visual SLAM often suffers from limited accuracy and robustness in such applications. To address it, this study proposes a road pole-like object change detection technology framework supported by visual point cloud quality optimization. First, a visual point cloud optimization method is constructed by fusing semantic constraints, LiDAR point cloud depth, and GNSS global correction, significantly improving trajectory accuracy and point cloud quality. Second, based on the optimized visual point cloud, accurate extraction and localization of road pole-like objects are achieved. Then, a fast-matching strategy based on hash mapping is introduced to achieve robust change detection of pole-like objects across different periods. Finally, the effectiveness of the overall process is verified using two phases of experimental data collected around Tongji University. Experiments show that the proposed method reduces the mean absolute error (MAE) and root mean square error (RMSE) of the trajectory by 68.39% and 65.65%, respectively, while increasing the point cloud density by an average of 57.97%. In the tasks of element localization and change detection, the mean localization error of pole-like objects in a single direction is 2~3 m, and the mean plane error is around 3.5 m. The matching accuracy of pole-like objects reaches 94.8%, and the accuracy and recall rates of change detection for added and removed pole-like objects both reach 100%. The results demonstrate the effectiveness and stability of the proposed method, confirm the application value of multi-source data fusion and optimization technology in road pole-like object change detection, and provide a reliable technical path for high-definition map change detection and intelligent management of road facilities.

Key words: multi-source data, visual point cloud, road pole-like objects, change detection

中图分类号:

P237

胡浩鹏, 吴杭彬, 战仕浩, 温在豪, 刘春. 视觉点云质量优化支持的道路杆状物变化检测[J]. 测绘学报, 2026, 55(2): 344-358.

Haopeng HU, Hangbin WU, Shihao ZHAN, Zaihao WEN, Chun LIU. Road pole-like object change detection supported by visual point cloud quality optimization[J]. Acta Geodaetica et Cartographica Sinica, 2026, 55(2): 344-358.

图/表 28

图1

图2

图3

表1

图4

表2

表3

图5

表4

图6

图7

表5

表6

表7

图8

图9

图10

图11

表8

表9

图12

表10

图13

图14

表11

图15

表12

图16

参考文献 39

[1]	张敏, 王雁河, 孙周, 等. 交能融合应用场景与技术体系研究[J]. 综合智慧能源, 2025, 47(2): 13-28.
	ZHANG Min, WANG Yanhe, SUN Zhou, et al. Research on application scenarios and technical system for transportation-energy integration[J]. Integrated Intelligent Energy, 2025, 47(2): 13-28.
[2]	杨蒙蒙, 江昆, 温拓朴, 等. 自动驾驶高精度地图众源更新技术现状与挑战[J]. 中国公路学报, 2023, 36(5): 244-259.
	YANG Mengmeng, JIANG Kun, WEN Tuopu, et al. Review on status and challenges of crowdsourced updating of highly automated driving maps[J]. China Journal of Highway and Transport, 2023, 36(5): 244-259.
[3]	XIAO Guiming, LEE Jaeyoung, JIANG Qianshan, et al. Safety improvements by intelligent connected vehicle technologies: a meta-analysis considering market penetration rates[J]. Accident Analysis and Prevention, 2021, 159: 106234.
[4]	LI Qingquan. Dynamic and precise engineering surveying[M]. Berlin: Springer, 2023.
[5]	FENG Jie, CHEN Hongbin, DENG Xianjun, et al. Confident information coverage hole prediction and repairing for healthcare big data collection in large-scale hybrid wireless sensor networks[J]. IEEE Internet of Things Journal, 2021, 8(23): 16801-16813.
[6]	范红超, 孔格菲, 杨岸然. 众源地理信息研究现状与展望[J]. 测绘学报, 2022, 51(7): 1653-1668. DOI: . doi: 10.11947/j.AGCS.2022.20220192
	FAN Hongchao, KONG Gefei, YANG Anran. Current status and prospects of research for volunteered geographic information[J]. Acta Geodaetica et Cartographica Sinica, 2022, 51(7): 1653-1668. DOI: . doi: 10.11947/j.AGCS.2022.20220192
[7]	MORENO C, GONZÁLEZ A, OLAZAGOITIA J L, et al. The acquisition rate and soundness of a low-cost data acquisition system (LC-DAQ) for high frequency applications[J]. Sensors, 2020, 20(2): 524.
[8]	ADADI A. A survey on data-efficient algorithms in big data era[J]. Journal of Big Data, 2021, 8(1): 24.
[9]	冉蓉, 徐兴华, 邱少华, 等. 基于深度卷积神经网络的裂纹检测方法综述[J]. 计算机工程与应用, 2021, 57(9): 23-35.
	RAN Rong, XU Xinghua, QIU Shaohua, et al. Review of crack detection methods based on deep convolutional neural networks[J]. Computer Engineering and Applications, 2021, 57(9): 23-35.
[10]	邹云昊, 李和松, 聂婧, 等. 基于RAW数据的图像处理及应用[J]. 信号处理, 2025, 41(2): 224-240.
	ZOU Yunhao, LI Hesong, NIE Jing, et al. Image processing and applications based on RAW sensor data[J]. Journal of Signal Processing, 2025, 41(2): 224-240.
[11]	CHENG Jun, ZHANG Liyan, CHEN Qihong, et al. A review of visual SLAM methods for autonomous driving vehicles[J]. Engineering Applications of Artificial Intelligence, 2022, 114: 104992.
[12]	LIU Hong, PAN Shuguo, GAO Wang, et al. LiDAR-inertial real-time state estimator with rod-shaped and planar feature[J]. Remote Sensing, 2022, 14(16): 4031.
[13]	YU Xinyi, HE Weiqi, QIAN Xuecheng, et al. Real-time rail recognition based on 3D point clouds[J]. Measurement Science and Technology, 2022, 33(10): 105207.
[14]	蔡怀宇, 杨朝乾, 崔子扬, 等. 图像引导和点云空间约束的公路洒落物检测定位方法[J]. 光电工程, 2024, 51(3): 230317.
	CAI Huaiyu, YANG Zhaoqian, CUI Ziyang, et al. Image-guided and point cloud space-constrained method for detection and localization of abandoned objects on the road[J]. Opto-Electronic Engineering, 2024, 51(3): 230317.
[15]	SAHILI A R, HASSAN S, SAKHRIEH S M, et al. A survey of visual SLAM methods[J]. IEEE Access, 2023, 11: 139643-139677.
[16]	CHEN Xinyu, YU Yantao. An unsupervised low-light image enhancement method for improving V-SLAM localization in uneven low-light construction sites[J]. Automation in Construction, 2024, 162: 105404.
[17]	PENG Hongrui, ZHAO Ziyu, WANG Liguan. A review of dynamic object filtering in SLAM based on 3D LiDAR[J]. Sensors, 2024, 24(2): 645.
[18]	THEODOROU C, VELISAVLJEVIC V, DYO V. Visual SLAM for dynamic environments based on object detection and optical flow for dynamic object removal[J]. Sensors, 2022, 22(19): 7553.
[19]	WANG Ke, ZHAO Guoliang, LU Jianbo. A deep analysis of visual SLAM methods for highly automated and autonomous vehicles in complex urban environment[J]. IEEE Transactions on Intelligent Transportation Systems, 2024, 25(9): 10524-10541.
[20]	LU Yifan, LI Quanhao, LIU Baoan, et al. Robust collaborative 3D object detection in presence of pose errors[C]//Proceedings of 2023 IEEE International Conference on Robotics and Automation (ICRA). London: IEEE, 2023: 4812-4818.
[21]	HOSAIN M T, ZAMAN A, ABIR M R, et al. Synchronizing object detection: applications, advancements and existing challenges[J]. IEEE Access, 2024, 12: 54129-54167.
[22]	YE Longjie, XIAO Wen, WENG Qihao. Supervoxel-based instance segmentation of pole-like facilities from mobile laser scanning data using pyramid cascaded fisher vector modeling[J]. IEEE Transactions on Geoscience and Remote Sensing, 2025, 63: 5701519.
[23]	LI Gang, YU Jian, HUANG Huilan, et al. Dynamic object removal and dense mapping for accurate visual SLAM in outdoor environments[J]. Measurement, 2025, 256: 118172.
[24]	LIU Wei, ANGUELOV D, ERHAN D, et al. SSD: single shot MultiBox detector[C]//Proceedings of 2016 Computer Vision-ECCV. Cham: Springer, 2016: 21-37.
[25]	LIM H, HWANG S, MYUNG H. ERASOR: egocentric ratio of pseudo occupancy-based dynamic object removal for static 3D point cloud map building[J]. IEEE Robotics and Automation Letters, 2021, 6(2): 2272-2279.
[26]	BESL P J, MCKAY N D. Method for registration of 3D shapes[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1992, 14(2): 239-256.
[27]	RODRIGUEZ O. Des lois geometriques qui regissent les desplacements d'un systeme solide dans l'espace et de la variation des coordonnees provenant de deplacements consideres independamment des causes qui peuvent les produire[J]. Vournalde Mathématiques Pureset Appliquées, 1840, 5: 380-440.
[28]	DEMMEL J, KAHAN W. Accurate singular values of bidiagonal matrices[J]. SIAM Journal on Scientific and Statistical Computing, 1990, 11(5): 873-912.
[29]	PANDE B, PADAMWAR K, BHATTACHARYA S, et al. A review of image annotation tools for object detection[C]//Proceedings of 2022 International Conference on Applied Artificial Intelligence and Computing (ICAAIC). Salem: IEEE, 2022: 976-982.
[30]	REDMON J, DIVVALA S, GIRSHICK R, et al. You only look once: unified, real-time object detection[C]//Proceedings of 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Las Vegas: IEEE, 2016: 779-788.
[31]	YAO A C. Uniform hashing is optimal[J]. Journal of the ACM, 1985, 32(3): 687-693.
[32]	WANG Yanjun, LI Shaochun, WANG Mengjie, et al. A simple deep learning network for classification of 3D mobile LiDAR point clouds[J]. Journal of Geodesy and Geoinformation Science, 2021, 4(3): 49-59.
[33]	YANG Juntao, KANG Zhizhong, YANG Zhou. A robust model fitting-based method for transmission line extraction from airborne LiDAR point cloud data[J]. Journal of Geodesy and Geoinformation Science, 2021, 4(3): 60-71.
[34]	GENG Yuxin, ZHONG Ruofei, HUANG Yuqin, et al. Automatic extraction method of 3D feature guidelines for complex cultural relic surfaces based on point cloud[J]. Journal of Geodesy and Geoinformation Science, 2024, 7(1): 16-41.
[35]	YI Hao, LIU Bo, ZHAO Bin, et al. Small object detection algorithm based on improved YOLOv8 for remote sensing[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2024, 17: 1734-1747.
[36]	WILLMOTT C J, MATSUURA K. Advantages of the mean absolute error (MAE) over the root mean square error (RMSE) in assessing average model performance[J]. Climate Research, 2005, 30(1): 79-82.
[37]	BESCOS B, FÁCIL J M, CIVERA J, et al. DynaSLAM: tracking, mapping, and inpainting in dynamic scenes[J]. IEEE Robotics and Automation Letters, 2018, 3(4): 4076-4083.
[38]	CASELITZ T, STEDER B, RUHNKE M, et al. Monocular camera localization in 3D LiDAR maps[C]//Proceedings of 2016 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). Daejeon: IEEE, 2016: 1926-1931.
[39]	CHU Tianxing, GUO Ningyan, BACKÉN S, et al. Monocular camera/IMU/GNSS integration for ground vehicle navigation in challenging GNSS environments[J]. Sensors, 2012, 12(3): 3162-3185.

变化情况	前期有杆状物	后期有杆状物	前后有语义变化	检测方法
新增	否	是	否	在杆状物匹配阶段，若当前期目标的候选匹配集合为空，则标记为新增候选。通过核密度估计过滤因噪声或临时障碍物导致的误检
移除	是	否	否	移除检测需验证点云缺失区域的持续性，并确认底部结构是否匹配历史数据，避免误判
语义变化	是	是	否	对两期同一杆状物的语义标签进行一致性校验，确定变更候选集，随后通过试验设备采集的图像数据多重验证，排除遮挡或误检

期	采集时间	GNSS	Mono	采集周期/s	采集长度/m	图像数
D1	2024-10	√	√	240	3 243.21	2401
D2	2025-04	√	√	247	3 263.67	2409

期	ORB-SLAM2			本文方法			优化提升
期	MAE/m	RMSE/m	时间/s	MAE/m	RMSE/m	时间/s	MAE/（%）	RMSE/（%）
D1	5.798	5.440	694	1.625	1.762	952	71.97	67.60
D2	5.941	6.047	682	2.091	2.196	972	64.80	63.69

期	ORB-SLAM2点云数量	本文方法点云数量	优化提升/（%）
D1	156 486	250 213	59.89
D2	140 852	219 792	56.04

期	DynaSLAM			本文方法（仅语义优化）
期	MAE/m	RMSE/m	时间/s	MAE/m	RMSE/m	时间/s
D1	3.763	3.424	1029	2.082	2.201	798
D2	3.913	3.664	1033	2.461	2.515	811