结合区域增长与RANSAC的机载LiDAR点云屋顶面分割

doi:10.11947/j.AGCS.2021.20200270

摘要/Abstract

摘要： 建筑物屋顶面大小差异较大、形状复杂、数量不确定等特点，以及机载LiDAR点云密度不均、分布不规则、缺乏语义信息等特性，对屋顶面的准确分割造成了很大干扰，因此现有分割方法的精度和适用性仍有待提高。针对上述问题，本文提出一种结合区域增长与RANSAC的机载LiDAR点云屋顶面分割方法。首先，引入稳健的法向量估计算法计算点云法向量，利用提出的迭代区域增长策略和RANSAC提取多个可靠屋顶面片；然后，基于可靠屋顶面片参数和RANSAC计算内点的思想，迭代合并可靠屋顶面片，并精化屋顶面参数；最后，计算未能通过前面步骤分割的点到各屋顶面的垂直距离，将其标记为距离最小且小于阈值的屋顶面，并通过局部范围内投票的方式精化屋顶面分割结果。利用多个具有代表性的建筑物点云和一组区域建筑物点云进行试验，结果表明，所提出的方法可有效地分割不同复杂程度的建筑物屋顶面，并能较好地分割面积较小的屋顶面，以屋顶面和单点为评价单元的平均分割正确率为95.56%和97.93%，分割的结果可为建筑物三维模型重建、点云精简等应用提供可靠的信息。

关键词: 屋顶面分割, 迭代区域增长, RANSAC, 机载LiDAR点云

Abstract: Roofs of a building have the characteristics of greatly different size, complex shape and uncertain number, and airborne LiDAR point cloud has the characteristics of uneven density, irregular distribution and without any semantic information, which make many existing airborne LiDAR roof segmentation methods ineffective and their applicability and precision still need to be improved. Thus, an airborne LiDAR roof segmentation method combining region growing with random sample consensus is proposed in the paper. Firstly, the robust normal estimation is introduced to calculate point cloud normal, a proposed iterative region growing strategy and random sample consensus are applied to extract many reliable roof patches. Then, an iterative process is performed to merge these roof patches based on their parameters and the idea of inlier selection of random sample consensus(RANSAC), and roof parameters are refined by the process. Finally, the orthogonal distance of points which are not segmented by the previous steps to each roof is calculated, and points are assigned to the corresponding roof with the minimum orthogonal distance and less than the threshold, and the roof segmentation results are refined by voting in the local neighborhood. Multiple representative building point clouds and a group of regional building point clouds are used in the experiment. The results show that the proposed method can effectively segment roofs of buildings with different complexity, and can also effectively segment roofs with small area, the average segmentation correctness is 95.56% and 97.93% by using a roof and a single point as the basic evaluation unit. The results can provide reliable information for applications such as three-dimensional building model reconstruction and point cloud reduction.

Key words: roof segmentation, iterative region growing, RANSAC, airborne LiDAR point cloud

中图分类号:

P237

赵传, 郭海涛, 卢俊, 余东行, 林雨准, 姜怀刚. 结合区域增长与RANSAC的机载LiDAR点云屋顶面分割[J]. 测绘学报, 2021, 50(5): 621-633.

ZHAO Chuan, GUO Haitao, LU Jun, YU Donghang, LIN Yuzhun, JIANG Huaigang. Roof segmentation from airborne LiDAR by combining region growing with random sample consensus[J]. Acta Geodaetica et Cartographica Sinica, 2021, 50(5): 621-633.

参考文献

[1] 汪利斌, 胡翰, 朱庆, 等. 局部表面参数化的实景三角网模型语义增强方法[J]. 测绘学报, 2020, 49(2):225-234. DOI:10.11947/j.AGCS.2020.20180588. WANG Libin, HU Han, ZHU Qing, et al. A semantic enhancement method for photorealistic mesh model based on local parameterization[J]. Acta Geodaetica et Cartographica Sinica, 2020, 49(2):225-234. DOI:10.11947/j.AGCS.2020.20180588.
[2] 赵瑞斌, 庞明勇, 张燕玲, 等. 机载LiDAR点云中精细建筑物顶面的渐进提取[J]. 计算机辅助设计与图形学学报, 2017, 29(4):624-631. ZHAO Ruibin, PANG Mingyong, ZHANG Yanling, et al. Progressively extracting accurate building roofs from airborne LiDAR data[J]. Journal of Computer-Aided Design & Computer Graphics, 2017, 29(4):624-631.
[3] 赵传, 张保明, 陈小卫, 等. 一种利用点云邻域信息的建筑物屋顶面高精度自动提取方法[J]. 测绘学报, 2017, 46(9):1123-1134. DOI:10.11947/j.AGCS.2017.20160518. ZHAO Chuan, ZHANG Baoming, CHEN Xiaowei, et al. Accurate and automatic building roof extraction using neighborhood information of point clouds[J]. Acta Geodaetica et Cartographica Sinica, 2017, 46(9):1123-1134. DOI:10.11947/j.AGCS.2017.20160518.
[4] LI Li, YAO Jian, TU Jingmin, et al. Roof plane segmentation from airborne LiDAR data using hierarchical clustering and boundary relabeling[J]. Remote Sensing, 2020, 12(9):1363-1384.
[5] LI Lin, YANG Fan, ZHU Haihong, et al. An improved RANSAC for 3D point cloud plane segmentation based on normal distribution transformation cells[J]. Remote Sensing, 2017, 9(5):432-448.
[6] DONG Zhen, YANG Bisheng, HU Pingbo, et al. An efficient global energy optimization approach for robust 3D plane segmentation of point clouds[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2018, 137:112-133.
[7] YAN Jixing, SHAN Jie, JIANG Wanshou. A global optimization approach to roof segmentation from airborne LiDAR point clouds[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2014, 94:183-193.
[8] 李明磊, 李广云, 王力, 等. 采用八叉树体素生长的点云平面提取[J]. 光学精密工程, 2018, 26(1):172-183. LI Minglei, LI Guangyun, WANG Li, et al. Planar feature extraction from unorganized point clouds using octree voxel-based region growing[J]. Optics and Precision Engineering, 2018, 26(1):172-183.
[9] WANG Miao, TSENG Y H. Automatic segmentation of LiDAR data into coplanar point clusters using an octree-based split-and-merge algorithm[J]. Photogrammetric Engineering and Remote Sensing, 2010, 76(4):407-420.
[10] VO A V, TRUONG-HONG L, LAEFER D F, et al. Octree-based region growing for point cloud segmentation[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2015, 104:88-100.
[11] WANG Yinghui, HAO Wen, NING Xiaojuan, et al. Automatic segmentation of urban point clouds based on the Gaussian map[J]. The Photogrammetric Record, 2013, 28(144):342-361.
[12] FERRAZ A, BRETAR F, JACQUEMOUD S, et al. 3D segmentation of forest structure using a mean-shift based algorithm[C]//Proceedings of 2010 IEEE International Conference on Image Processing. Hong Kong, China:IEEE, 2010:1413-1416.
[13] SAMPATH A, SHAN Jie. Segmentation and reconstruction of polyhedral building roofs from aerial LiDAR point clouds[J]. IEEE Transactions on Geoscience and Remote Sensing, 2010, 48(3):1554-1567.
[14] BIOSCA J M, LERMA J L. Unsupervised robust planar segmentation of terrestrial laser scanner point clouds based on fuzzy clustering methods[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2008, 63(1):84-98.
[15] ZHANG Chunsun, HE Yuxiang, FRASER C S. Spectral clustering of straight-line segments for roof plane extraction from airborne LiDAR point clouds[J]. IEEE Geoscience and Remote Sensing Letters, 2018, 15(2):267-271.
[16] XU Xiaobin, LUO Minzhou, TAN Zhiying, et al. Plane segmentation and fitting method of point clouds based on improved density clustering algorithm for laser radar[J]. Infrared Physics & Technology, 2019, 96:133-140.
[17] GU Yanfeng, CAO Zhimin, DONG Limin. A hierarchical energy minimization method for building roof segmenta-tion from airborne LiDAR data[J]. Multimedia Tools and Applications, 2017, 76(3):4197-4210.
[18] ADAM A, CHATZILARI E, NIKOLOPOULOS S, et al. H-RANSAC:a hybrid point cloud segmentation com-bining 2D and 3D data[C]//Proceedings of ISPRS Annual Photogrammetry Remote Sensing Spatial Information Science. Riva del Garda, Italy:ISPRS, 2018:1-8.
[19] TARSHA-KURDI F, LANDES T, GRUSSENMEYER P, et al. Model-driven and data-driven approaches using LIDAR data:analysis and comparison[C]//Proceedings of the Photogrammetric Image Analysis. Munich, Germany:Springer, 2007:87-92.
[20] 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, 1980, 24(6):381-395.
[21] SCHNABEL R, WAHL R, KLEIN R. Efficient RANSAC for point-cloud shape detection[J]. Computer Graphics Forum, 2007, 26(2):214-226.
[22] SUN B, MORDOHAI P. Oriented point sampling for plane detection in unorganized point clouds[C]//Proceedings of 2019 International Conference on Robotics and Automation.Montreal,Canada:IEEE, 2019:2917-2923.
[23] XU Bo, JIANG Wanshou, SHAN Jie, et al. Investigation on the weighted RANSAC approaches for building roof plane segmentation from LiDAR point clouds[J]. Remote Sensing, 2016, 8(1):1-23.
[24] WU Teng, HU Xiangyun, YE Lizhe. Fast and accurate plane segmentation of airborne LiDAR point cloud using cross-line elements[J]. Remote Sensing, 2016, 8(5):383-409.
[25] KIM C, HABIB A, PYEON M, et al. Segmentation of planar surfaces from laser scanning data using the magnitude of normal position vector for adaptive neighborhoods[J]. Sensors, 2016, 16(2):140-170.
[26] TARSHA-KURDI F, LANDES T, GRUSSENMEYER P. Hough-transform and extended RANSAC algorithms for automatic detection of 3D building roof planes from LiDAR data[J]. International Archives of Photogrammetry, Remote Sensing and Spatial Information Sciences, 2007, 36:407-412.
[27] WANG Liang, SHEN Chao, DUAN Fuqing, et al. Energy-based multi-plane detection from 3D point clouds[M]//HIROSE A, OZAWA S, DOYA K, et al. International Conference on Neural Information Processing. Kyoto, Japan:Springer, 2016:715-722.
[28] CAO Rujun, ZHANG Yongjun, LIU Xinyi, et al. Roof plane extraction from airborne LiDAR point clouds[J]. International Journal of Remote Sensing, 2017, 38(12):3684-3703.
[29] ZHANG Jie, CAO Junjie, LIU Xiuping, et al. Point cloud normal estimation via low-rank subspace clustering[J]. Computers & Graphics, 2013, 37(6):697-706.
[30] BOULCH A, MARLET R. Fast and robust normal estimation for point clouds with sharp features[J]. Computer Graphics Forum, 2012, 31(5):1765-1774.
[31] ZHANG Jie, CAO Junjie, LIU Xiuping, et al. Multi-normal estimation via pair consistency voting[J]. IEEE Transactions on Visualization and Computer Graphics, 2019, 25(4):1693-1706.
[32] BEN-SHABAT Y, LINDENBAUM M, FISCHER A, et al. Nesti-Net:normal estimation for unstructured 3D point clouds using convolutional neural networks[C]//Proceedings of 2019 IEEE/CVF Computer Vision and Pattern Recognition. Long Beach, California:IEEE, 2019:10112-10120.
[33] 高广, 马洪超, 张良. 利用合成算法从LiDAR数据提取屋顶面[J]. 武汉大学学报(信息科学版), 2014, 39(10):1225-1230. GAO Guang, MA Hongchao, ZHANG Liang. Automatic extraction of building roofs from LiDAR data using a hybridized method[J]. Geomatics and Information Science of Wuhan University, 2014, 39(10):1225-1230.
[34] ISPRS. ISPRS test project on urban classification and 3D building reconstruction[DB/OL]. (2016-8-13)[2020-06-10]. http://www2.isprs.org/commissions/comm3/wg4/detection-and-reconstruction.html.
[35] OpenTopography. NCALM Project. PI:Shane Grigsby, University of Colorado, Boulder[DB/OL]. (2016-9-10)[2020-06-10]. https://doi.org/10.5069/G9ZC80SR.
[36] AWRANGJEB M, FRASER C S. An automatic and threshold-free performance evaluation system for building extraction techniques from airborne LIDAR data[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2014, 7(10):4184-4198.
[37] RABBANI T, HEUVEL F, VOSSELMAN G. Segmentation of point clouds using smoothness constraint[C]//Proceedings of 2006 International Archives of Photogrammetry, Remote Sensing and Spatial Information Sciences.[S.l.]:ISPRS, 2006:248-253.