A deep neural network model for road extraction of MLS LiDAR point cloud

doi:10.13474/j.cnki.11-2246.2023.0351

Abstract

Abstract: PointNet++ has shown better performance than traditional methods in MLS LiDAR point cloud road extraction, but there are still the phenomena of over segmentation or under segmentation for road edge extraction.To address this issue, an improved neighborhood enhancement coding network E-PointNet++ is proposed. By introducing a neighborhood enhancement coding module before feature extraction, the connection between local neighborhood points is established to improve the network's road edge segmentation ability.Comparative experiments are conducted on two datasets, and E-PointNet++ shows significantly better performance than other methods, with accuracy, integrity and detection quality all exceeding 97%. This method performs robustly on different datasets and scenarios.

Key words: MLS LiDAR point cloud, deep learning, road extraction, edge segmentation, neighborhood enhanced coding

CLC Number:

P237

LIU Jin, YANG Ronghao, WEN Wen, TAN Junxiang, LAN Qinglong, GAO Xiang, TANG Hong. A deep neural network model for road extraction of MLS LiDAR point cloud[J]. Bulletin of Surveying and Mapping, 2023, 0(12): 8-12,18.

References

[1] MA Lingfei,LI Ying,LI J,et al. Mobile laser scanned point-clouds for road object detection and extraction: a review[J]. Remote Sensing,2018,10(10): 1531.
[2] 王羽尘,于斌,陈晓阳,等. 基于激光雷达点云的道路几何信息提取与数字化建模研究[J]. 中国公路学报,2023,36(3):45-60.
[3] 惠振阳,胡友健,康妍斐. 基于反射强度偏度平衡的道路点云提取算法[J]. 激光与光电子学进展,2018,55(2):439-445.
[4] ZHANG Wuming,QI Jianbo,WAN Peng,et al. An easy-to-use airborne LiDAR data filtering method based on cloth simulation[J]. Remote Sensing,2016,8(6):501.
[5] 方莉娜,杨必胜. 车载激光扫描数据的结构化道路自动提取方法[J]. 测绘学报,2013,42(2):260-267.
[6] 胡啸,黄明,周海霞. 车载激光扫描数据的高速道路自动提取方法[J]. 测绘科学,2019,44(3):101-106.
[7] 卢健,贾旭瑞,周健,等. 基于深度学习的三维点云分割综述[J]. 控制与决策,2023,38(3):595-611.
[8] SU Hang,MAJI S,KALOGERAKIS E,et al. Multi-view convolutional neural networks for 3D shape recognition[C]//Proceedings of 2015 IEEE International Conference on Computer Vision (ICCV). Santiago:IEEE,2015:945-953.
[9] MENG H Y,GAO Lin,LAI Yukun,et al. VV-net:voxel VAE net with group convolutions for point cloud segmentation[C]//Proceedings of 2019 IEEE/CVF International Conference on Computer Vision (ICCV). Seou:IEEE,2019:8500-8508.
[10] MATURANA D,SCHERER S. VoxNet:a 3D Convolutional Neural Network for real-time object recognition[C]//Proceedings of 2015 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).Hamburg:IEEE,922-928.
[11] BOULCH A,GUERRY J,LE SAUX B,et al. SnapNet:3D point cloud semantic labeling with 2D deep segmentation networks[J]. Computers & Graphics,2018,71:189-198.
[12] CHARLES R Q,HAO Su,MO Kaichun,et al. PointNet:deep learning on point sets for 3D classification and segmentation[C]// Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Honolulu:IEEE,2017:652-660.
[13] 杨晓文,王爱兵,韩燮,等. 基于KNN-PointNet的点云语义分割[J]. 激光与光电子学进展,2021,58(24):272-279.
[14] QI C R,YI Li,SU Hao,et al. PointNet++:deep hierarchical feature learning on point sets in a metric space[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems. New York:ACM,2017:5105-5114.
[15] XU Mutian,DING Runyu,ZHAO Hengshuang,et al. PAConv:position adaptive convolution with dynamic kernel assembling on point clouds[C]// Proceedings of 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Nashville:IEEE,2021:3173-3182.
[16] 焦晨,王宝锋,易耀华. 点云数据滤波算法研究[J]. 国外电子测量技术,2019,38(11):18-22.
[17] DESCHAUD J E,DUQUE D,RICHA J P,et al. Paris-CARLA-3D:a real and synthetic outdoor point cloud dataset for challenging tasks in 3D mapping[J]. Remote Sensing,2021,13(22):4713.
[18] 冉妮妮. 城市结构化道路路面点云提取方法研究[J]. 测绘技术装备,2022,24(1):57-61.
[19] 管郡智,潘卫清. 基于多区域RANSAC的地面点云提取算法[J]. 电子技术与软件工程,2020(14):176-177.