结合上下文特征和图割算法的车载点云聚类方法

doi:10.3969/j.issn.0255-8297.2020.06.009

摘要/Abstract

摘要： 可靠、准确的点云聚类是后续高精度场景目标分析与解译的基础.该文提出了一种基于上下文特征和图割算法的车载点云聚类方法.首先用DBSCAN （density-based spatial clustering of applications with noise）对点云数据进行过分割，得到密度可达的超体素；然后引入空间和属性上下文特征来描述超体素间的关联，并用于定义超体素构建的图模型边的权值；最后基于多标记的图割优化算法得到最佳超体素聚簇.实验结果表明，该方法能够有效改善点云聚类过分割，从而提高聚类的精度.

关键词: DBSCAN, 超体素, 上下文特征, 图割, 点云聚类

Abstract: Reliable and accurate clustering of point cloud is the basis for subsequent high-precision analysis and interpretation of scene target. This paper presents a method for vehicle point cloud segment using supervoxel and graph-cut with contextual feature. Firstly, density-based spatial clustering of applications with noise (DBSCAN) is used to segment the point cloud data, and density-reachable supervoxels can be obtained. Secondly, spatial and attribute contextual features are introduced to describe the correlation between supervoxels and to define the weights of the edges of the graph model constructed by supervoxels. Finally, the optimal supervoxel clustering is obtained based on multi-label graph-cut optimization algorithm. Experimental results show that the proposed method has improved accuracy and performance on over segmentation in clustering.

Key words: density-based spatial clustering of applications with noise (DBSCAN), supervoxel, contextual feature, graph cut, point cloud clustering

中图分类号:

P208

刘亚文, 张颖. 结合上下文特征和图割算法的车载点云聚类方法[J]. 应用科学学报, 2020, 38(6): 924-935.

LIU Yawen, ZHANG Ying. Vehicle Point Cloud Clustering Based on Contextual Feature and Graph Cut[J]. Journal of Applied Sciences, 2020, 38(6): 924-935.

参考文献

[1] Yang J Y, Gan Z Q, Li K, et al. Graph-based segmentation for RGB-D data using 3-D geometry enhanced super pixels[J]. IEEE Transactions on Cybernetics, 2017, 45(5):927-940.
[2] 曾绍琴, 李光强, 廖志强. 空间聚类方法分类[J]. 测绘科学, 2012, 37(5):103-106. Zeng S Q, Li G Q, Liao Z Q. A new category of spatial clustering methods[J]. Science of Surveying and Mapping, 2012, 37(5):103-106. (in Chinese)
[3] Fan Y, Wang M, Geng N, et al. A self-adaptive segmentation method for a point cloud[J]. The Visual Computer, 2018(34):659-673.
[4] Ioannou Y, Taati B, Harrap R, et al. Difference of normals as a multi-scale operator in unorganized point clouds[C]//Proceedings of the 2012 Second International Conference on 3D Imaging, Modeling, Processing, Visualization & Transmission, Zurich, Switzerland, 2012:501-508.
[5] Lin Y, Wang C, Zhai D, et al. Toward better boundary preserved supervoxel segmentation for 3D point clouds[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2018, 143:39-47.
[6] Lin C R, Liu K H, Chen M S. Dual clustering:integrating data clustering over optimization and constraint domains[J]. IEEE Transactions on Knowledge & Data Engineering, 2005, 17(5):628-637.
[7] Zhou J G, Guan J H, Li P X. A dual clustering algorithm for distributed spatial databases[J]. Geo-Spatial Information Science, 2007, 10(2):137-144.
[8] 李光强, 邓敏, 程涛, 等. 一种基于双重距离的空间聚类方法[J]. 测绘学报,2008, 37(4):482-488. Li G Q, Deng M, Cheng T, et al. A dual distance based spatial clustering method[J]. Acta Geodaetica et Cartographica Sinica, 2008, 37(4):482-488. (in Chinese)
[9] Xu Y, Tuttas S, Hoegner L, et al. Voxel-based segmentation of 3D point clouds from construction sites using a probabilistic connectivity model[J]. Pattern Recognition Letters, 2018, 102:67-74.
[10] Verdoja F, Thomas D, Sugimoto A. Fast 3d point cloud segmentation using supervoxels with geometry and color for 3d scene understanding[C]//IEEE International Conference on Multimedia and Expo, Hongkong, China, 2017:1285-1290.
[11] Stein S, Schoeler M, Papon J, et al. Object partitioning using local convexity[C]//IEEE Conference on Computer Vision and Pattern Recognition, Ohio, USA, 2014:304-311.
[12] Xu Y, Hoegner L, Tuttas S, et al. Voxel- and graph-based point cloud segmentation of 3D scenes using perceptual grouping laws[C]//ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Hannover, Germany, 2017:43-50.
[13] Green W R, Grobler H. Normal distribution transform graph-based point cloud segmentation[C]//Pattern Recognition Association of South Africa and Robotics and Mechatronics International Conference, IEEE, Port Elizabeth, South Africa, 2015:54-59.
[14] Hackel T, Wegner J D, Schindler K. Contour detection in unstructured 3D point clouds[C]//IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, USA, 2016:1610-1618.
[15] Landrieu L, Boussaha M. Point cloud oversegmentation with graph-structured deep metric learning[C]//IEEE Conference on Computer Vision and Pattern Recognition, Los Angeles, USA, 2019:1-15.
[16] Tran T N, Drab K, Daszykowski M. Revised DBSCAN algorithm to cluster data with dense adjacent clusters[J]. Chemometrics & Intelligent Laboratory Systems, 2013, 120(2):92-96.
[17] 唐云龙. 基于车载激光点云数据的典型地物分类与提取[D]. 北京:北京工业大学, 2015.
[18] Tan B, Zhong R F, Li Q. Objects classification with vehicle-borne laser scanning data[J]. Journal of Remote Sensing, 2012, 16(1):50-66.
[19] 张颖, 刘亚文, 苗堃. 基于空间上下文关联的车载点云聚类方法[J]. 测绘地理信息, 2019, 44(4):116-121. Zhang Y, Liu Y W, Miao K. A spatial context_based clustering approach for vehicle-borne laser scanning data[J]. Journal of Geomatics, 2019, 44(4):116-121. (in Chinese)
[20] 李英昌. 基于上下文信息的语义图像分类研究[D]. 杭州:浙江大学, 2014.
[21] 徐胜军, 韩九强, 刘光辉. 基于马尔可夫随机场的图像分割方法综述[J]. 计算机应用研究,2013, 30(9):2576-2582. Xu S J, Han J Q, Liu G H. Survey of image segmentation methods based on Markov random fields[J]. Application Research of Computers, 2013, 30(9):2576-2582. (in Chinese)
[22] Boykov Y, Veksler O, Zabih R. Fast approximate energy minimization via graph cuts[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2001, 20(12):1222-1239.
[23] Boykov Y, Kolmogorov V. An experimental comparison of min-cut/max-flow algorithms for energy minimization in vision[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2004, 26(9):1124-1137.