﻿ 顾及点云曲率的快速点云表面模型重建算法

# 顾及点云曲率的快速点云表面模型重建算法A Fast Surface Reconstruction Algorithm Considering the Curvature of Point Cloud

Abstract: We describe a fast surface reconstruction algorithm considering the curvature of point cloud from a set of merged range scans. Our key contribution is improving the efficiency of the algorithm by deleting part of visual information. First, Delaunay edges are added to the point cloud to construct Delaunay structure. Then, part of visual information is deleted base of curvature of point cloud, and a graph-cuts problem is established based on the remaining visual information. Finally, a surface model is obtained by solving the graph-cuts problem. We tested our method on several publicly available sets of range scans. The experimental results show that the method can efficiently reconstruct high-quality surface model with rich details and high integrity.

1. 引言

2. 快速点云表面模型重建算法

2.1. Delaunay三角化

2.2. 基于点云曲率删减可视信息

2.3. 利用可视信息加权

$w=\alpha ×\left(1.0-{e}^{\frac{-{d}^{2}}{2×\sigma }}\right)$ (1)

Figure 1. Weighting process of single ray

2.4. 图割求解

$C\left(S,T\right)=\underset{Vp\in S\\left\{s\right\}\wedge Vq\in T\\left\{t\right\}}{\sum }{w}_{pg}+\underset{Vp\in S\\left\{s\right\}}{\sum }{w}_{pt}+\underset{Vp\in T\\left\{t\right\}}{\sum }{w}_{ps}$ (2)

3. 基于点云曲率的光线删减

3.1. 单点光线删减

Figure 2. Full space and free space

3.2. 基于点云曲率的光线删减

Figure 3. Mean curvature of point cloud

4. 实验结果分析

4.1. 数据准备

4.2. 实验结果

Figure 4. Surface model of the point cloud

Figure 5. Comparison of vegetation reconstruction results

Figure 6. Comparison of building reconstruction results

Table 1. System resulting data of standard experiment

5. 结论与展望

NOTES

*通讯作者。

[1] 杨建思, 杜志强, 彭正洪, 等. 数字城市三维景观模型的建模技术[J]. 武汉大学学报(工学版), 2003, 36(3): 37-40.

[2] 詹总谦, 林元培, 艾海滨. 基于3ds Max二次开发的建筑物快速三维重建[J]. 测绘通报, 2016(11): 22-25.

[3] 杨必胜, 梁福逊, 黄荣刚. 三维激光扫描点云数据处理研究进展、挑战与趋势[J]. 测绘学报, 2017, 46(10): 1509-1516.

[4] 张卫龙. 局部信息约束的三维重建方法研究[D]: [博士学位论文]. 武汉: 武汉大学, 2019.

[5] 马东岭, 王晓坤, 李广云. 利用图割算法进行城市密集点云表面模型重建[J]. 测绘通报, 2019(2): 45-48.

[6] Kazhdan, M. (2006) Poisson Surface Reconstruction. Symposium on Geometry Processing, Goslar, 61-70.

[7] Rocchini, C., Cignoni, P., Ganovelli, F., et al. (2001) Marching Intersections: An Efficient Resampling Algorithm for Surface Management. International Conference on Shape Modeling & Applications, Los Alamitos, 296.

[8] Calakli, F. and Taubin, G. (2012) SSD-C: Smooth Signed Distance Colored Surface Reconstruction. In: Ex-panding the Frontiers of Visual Analytics and Visualization, Springer, London, 323-338.
https://doi.org/10.1007/978-1-4471-2804-5_18

[9] Labatut, P., Pons, J.P. and Keriven, R. (2009) Robust and Efficient Surface Reconstruction from Range Data. Computer Graphics Forum, 28, 2275-2290.
https://doi.org/10.1111/j.1467-8659.2009.01530.x

[10] Jancosek, M. and Pajdla, T. (2011) Multi-View Recon-struction Preserving Weakly-Supported Surfaces. Computer Vision and Pattern Recognition (CVPR), Colorado Springs, 3121-3128.
https://doi.org/10.1109/CVPR.2011.5995693

[11] Cao, T., et al. (2014) A GPU Accelerated Algo-rithm for 3D Delaunay Triangulation. In: Proceedings of the Symposium on Interactive 3D Graphics, ACM, New York, 47-54.
https://doi.org/10.1145/2556700.2556710

[12] 詹洋, 尹颜朋. 基于Minimum s-t Cut三维表面重建算法[J]. 现代计算机(专业版), 2018(2): 34-37.

[13] Boykov, Y. and Kolmogorov, V. (2001) An Experimental Comparison of Min-Cut/Max-Flow Algorithms for Energy Minimization in Vision. In: International Workshop on Energy Minimization Methods in Computer Vision and Pattern Recognition, Springer, Berlin, 359-374.
https://doi.org/10.1007/3-540-44745-8_24

[14] Jancosek, M. and Pajdla, T. (2014) Exploiting Visibility Infor-mation in Surface Reconstruction to Preserve Weakly Supported Surfaces. International Scholarly Research Notices, 2014, Article ID: 798595.
https://doi.org/10.1155/2014/798595

[15] Boissonnat, J.D., Devillers, O., Pion, S., et al. (2002) Triangulations in CGAL. Computational Geometry, 22, 5-19.
https://doi.org/10.1016/S0925-7721(01)00054-2

[16] Rusu, R.B. and Cousins, S. (2011) 3D Is Here: Point Cloud Library (PCL). IEEE International Conference on Robotics and Automation, Shanghai, 9-13 May 2011, 1-4.
https://doi.org/10.1109/ICRA.2011.5980567

Top