基于点云曲面拟合的自适应阈值地面分割算法

doi:10.3778/j.issn.1002-8331.2404-0069

摘要/Abstract

摘要： 在无人驾驶矿用卡车自主运行生产的过程中，激光雷达点云处理中的地面分割是矿区目标检测的关键部分。该方法的主要目的是解决传统地面分割算法无法分割矿区崎岖路面的问题，提高数据处理的精度，保证数据处理的速度。在点云的每一帧中，以激光雷达为中心将点云基于密度进行栅格划分，在每个栅格区域中选择最低点作为拟合种子点。根据选出的种子点，使用移动最小二乘法拟合曲面模型，在拟合过程中引入高斯型权函数、正余弦基函数、正交函数集的策略，缩短曲面拟合时间，还原曲面原始形态，并提出自适应阈值的分割方法，精确分割矿区崎岖路面。该算法既可以分割水平路面，又可以分割崎岖路面，大幅提高了矿区地面分割的鲁棒性。在鄂尔多斯某露天矿区实际复杂场景测试下，召回率为94.25%，每帧数据分割的平均耗时为26.2?ms。结果表明，该方法具有较高的精度，效率满足实时性要求。

关键词: 无人驾驶矿用卡车, 点云处理, 地面分割, 移动最小二乘法, 固态式激光雷达

Abstract: In the process of autonomous operation and production of unmanned mining trucks, the ground segmentation in LiDAR point cloud processing is a key part of mine target detection. The main purpose of this method is to solve the problem that the traditional ground segmentation algorithm cannot segment the rugged road surface in mining area, improve the accuracy of data processing and ensure the speed of data processing. In each frame of the point cloud, the point cloud is divided into grids based on density with LiDAR as the center, and the lowest point in each grid region is selected as the fitting seed point. According to the selected seed points, moving least square method is used to fit the surface model, and the strategies of Gaussian weight function, sine and cosine basis function and orthogonal function set are introduced in the fitting process to shorten the fitting time of the surface, which restores the original shape of the surface, and a segmentation method of adaptive threshold is proposed to accurately segment the ground on the rough pavement in the mining area. The proposed algorithm can segment both horizontal and rugged pavement, greatly improving the robustness of mining ground segmentation. In the actual complex scene test of an open-pit mining area in Ordos, the recall rate is 94.25%, and the average time of each frame of data segmentation is 26.2 ms. Experimental results show that the method has high precision and the efficiency meets the real-time requirement.

Key words: unmanned mining truck, point cloud processing, ground segmentation, moving least square method, solid-state LiDAR

李兆强, 吴巧俊, 熊福力, 张岳. 基于点云曲面拟合的自适应阈值地面分割算法[J]. 计算机工程与应用, 2025, 61(16): 337-347.

LI Zhaoqiang, WU Qiaojun, XIONG Fuli, ZHANG Yue. Adaptive Threshold Ground Segmentation Algorithm Based on Point Cloud Surface Fitting[J]. Computer Engineering and Applications, 2025, 61(16): 337-347.

参考文献

[1] 张晞, 梁斌, 于淼, 等. 露天矿山无人驾驶运输技术现状及发展趋势研究[J]. 煤炭工程, 2022, 54(6): 132-138.
ZHANG X, LIANG B, YU M, et al. Current situation and development direction of unmanned transportation technology in open pit mines[J]. Coal Engineering, 2022, 54(6): 132-138.
[2] MOOSMANN F, PINK O, STILLER C. Segmentation of 3D LiDAR data in non-flat urban environments using a local convexity criterion[C]//Proceedings of the 2009 IEEE Intelligent Vehicles Symposium. Piscataway: IEEE, 2009: 215-220.
[3] DOUILLARD B, UNDERWOOD J, KUNTZ N, et al. On the segmentation of 3D LiDAR point clouds[C]//Proceedings of the 2011 IEEE International Conference on Robotics and Automation. Piscataway: IEEE, 2011: 2798-2805.
[4] ANGUELOV D, TASKARF B, CHATALBASHEV V, et al. Discriminative learning of Markov random fields for segmentation of 3D scan data[C]//Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2005: 169-176.
[5] FISCHLER M A, BOLLES R C. Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography[J]. Communications of the ACM, 1981, 24(6): 381-395.
[6] ZERMAS D, IZZAT I, PAPANIKOLOPOULOS N. Fast segmentation of 3D point clouds: a paradigm on LiDAR data for autonomous vehicle applications[C]//Proceedings of the 2017 IEEE International Conference on Robotics and Automation. Piscataway: IEEE, 2017: 5067-5073.
[7] 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.
[8] LIM H, OH M, MYUNG H. Patchwork: concentric zone-based region-wise ground segmentation with ground likelihood estimation using a 3D LiDAR sensor[J]. IEEE Robotics and Automation Letters, 2021, 6(4): 6458-6465.
[9] KAMMEL S, PITZER B. Lidar-based lane marker detection and mapping[C]//Proceedings of the 2008 IEEE Intelligent Vehicles Symposium. Piscataway: IEEE, 2008: 1137-1142.
[10] HIMMELSBACH M, HUNDELSHAUSEN F V, WUENSCHE H J. Fast segmentation of 3D point clouds for ground vehicles[C]//Proceedings of the 2010 IEEE Intelligent Vehicles Symposium. Piscataway: IEEE, 2010: 560-565.
[11] 张佩翔, 王奇, 高仁璟, 等. 局部阈值自适应的地面点云分割[J]. 光学精密工程, 2023, 31(17): 2564-2572.
ZHANG P X, WANG Q, GAO R J, et al. Ground point cloud segmentation based on local threshold adaptive method[J]. Optics and Precision Engineering, 2023, 31(17): 2564-2572.
[12] LEE S, LIM H, MYUNG H. Patchwork++: fast and robust ground segmentation solving partial under-segmentation using 3D point cloud[C]//Proceedings of the 2022 IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway: IEEE, 2022: 13276-13283.
[13] MIJIT T, FIRKAT E, YUAN X M, et al. LR-Seg: a ground segmentation method for low-resolution LiDAR point clouds[J]. IEEE Transactions on Intelligent Vehicles, 2024, 9(1): 347-356.
[14] 邹兵, 陈鹏, 刘登洪. 一种基于栅格投影的快速地面点云分割算法[J]. 城市勘测, 2021(3): 112-116.
ZOU B, CHEN P, LIU D H. A fast ground point cloud segmentation algorithm based on grid projection[J]. Urban Geotechnical Investigation & Surveying, 2021(3): 112-116.
[15] 任彬, 崔健源, 李刚, 等. 基于自适应阈值的三维点云分段式去噪方法[J]. 光子学报, 2022, 51(2): 311-324.
REN B, CUI J Y, LI G, et al. A three-dimensional point cloud denoising method based on adaptive threshold[J]. Acta Photonica Sinica, 2022, 51(2): 311-324.
[16] 潘文标, 元文浩. 基于密度划分的云数据分块存储方法仿真[J]. 计算机仿真, 2022, 39(8): 456-459.
PAN W B, YUAN W H. Simulation of cloud data block storage method based on density division[J]. Computer Simulation, 2022, 39(8): 456-459.
[17] 曾清红, 卢德唐. 基于移动最小二乘法的曲线曲面拟合[J]. 工程图学学报, 2004, 25(1): 84-89.
ZENG Q H, LU D T. Curve and surface fitting based on moving least-squares methods[J]. Journal of Engineering Graphics, 2004, 25(1): 84-89.
[18] 左勇, 任阳, 杜志华, 等. 基于LP-RANSAC算法的路面点云快速移除算法[J]. 激光与光电子学进展, 2023, 60(14): 376-381.
ZUO Y, REN Y, DU Z H, et al. Rapid removal algorithm of road surface point cloud based on LP-RANSAC algorithm[J]. Laser & Optoelectronics Progress, 2023, 60(14): 376-381.
[19] 王春阳, 丘文乾, 刘雪莲, 等. 基于平面拟合的地面点云精确分割方法[J]. 吉林大学学报 (工学版), 2023, 53(3): 933-940.
WANG C Y, QIU W Q, LIU X L, et al. Accurate segmentation method of ground point cloud based on plane fitting[J]. Journal of Jilin University (Engineering and Technology Edition), 2023, 53(3): 933-940.
[20] 梅圣明, 黄妙华, 柳子晗, 等. 基于三维激光雷达的复杂场景中地面分割方法[J]. 激光与光电子学进展, 2022, 59(10): 422-429.
MEI S M, HUANG M H, LIU Z H, et al. Ground segmentation method in complex scenes based on three-dimensional LiDAR[J]. Laser & Optoelectronics Progress, 2022, 59(10): 422-429.
[21] 张凯, 于春磊, 赵亚丽, 等. 基于自适应阈值的三维激光点云地面分割算法研究[J]. 汽车工程, 2021, 43(7): 1005-1012.
ZHANG K, YU C L, ZHAO Y L, et al. Research on ground segmentation algorithm based on adaptive thresholds for 3D laser point clouds[J]. Automotive Engineering, 2021, 43(7): 1005-1012.
[22] RUMMELHARD L, PAIGWAR A, NèGRE A, et al. Ground estimation and point cloud segmentation using spatiotemporal conditional random field[C]//Proceedings of the 2017 IEEE Intelligent Vehicles Symposium. Piscataway: IEEE, 2017: 1105-1110.
[23] 李剑宇, 王可智, 王纪凯, 等. 郊区场景下基于LiDAR点云的地面分割[C]//第23届中国系统仿真技术及其应用学术年会会议论文集, 2022: 267-270.
LI J Y, WANG K Z, WANG J K, et al. Ground segmentation based on LiDAR point cloud in suburban scenarios[C]//Proceedings of the 23rd China Annual Conference on System Simulation Technology and Its Application, 2022: 267-270.