基于下降图像三维地形重建的参数自适应生成

孟偲; 周娜; 张亚妮

doi:10.13700/j.bh.1001-5965.2013.0464

基于下降图像三维地形重建的参数自适应生成

doi: 10.13700/j.bh.1001-5965.2013.0464

北京航空航天大学宇航学院, 北京 100191

基金项目:

中央高校基本科研业务费资助项目（YWF-13-JQCJ-026）

详细信息

作者简介:
孟偲（1977-），男，新疆沙湾人，副教授，Tsai@buaa.edu.cn.

中图分类号: V11
计量
- 文章访问数: 1334
- HTML全文浏览量: 217
- PDF下载量: 520
- 被引次数: 0
出版历程
- 收稿日期: 2013-08-12
- 网络出版日期: 2014-07-20

Adaptive parameter generation in terrain reconstruction with descent images

School of Astronautics, Beijing University of Aeronautics and Astronautics, Beijing 100191, China

摘要

摘要: 针对基于单应性下降图像着陆区三维地形重建算法中重建参数对重建结果的影响，提出一种参数自适应生成算法计算最优重建参数.首先，根据匹配特征点对应的空间点拟合出地形近似平面，作为系列平行虚拟切分平面的参考平面，从而确定切分平面的法向和基本深度.其次，在地形近似平面两侧，分别计算特征点对对应的空间点到地形拟合平面最大距离，用于估计重建区地形高度差异.最后，根据地形高度差异和相机距地形拟合平面距离，寻找最优切分平面个数.根据这些重建参数即可计算地形重建所需单应矩阵.对比不同重建参数下的重建结果表明该参数设定方法可以达到更优重建效果.
- 下降图像 /
- 地形重建 /
- 三维重建 /
- 参数 /
- 自适应
Abstract: In homography based landing area terrain reconstruction with descent images, the reconstruction parameters greatly influence the reconstruction results. Therefore a method for adaptive parameter generation was proposed to find the best parameters automatically. Firstly, an approximate terrain plane was fitted by the spatial points which were determined by the matched feature points in the descent images. The fitted terrain plane worked as the reference plane of the series of parallel slices and determined the normal vector and basic depth. Secondly, at both sides of the fitted terrain plane, the furthest distances between the 3D points and the fitted terrain plane were calculated separately, and thus the terrain height difference could be reckoned. At last, the best account of virtual slice planes was determined with the terrain height difference and the basic depth information of the fitted terrain plane. Through these steps, the homographies which were needed to recover the terrain of landing area could be worked out by these parameters. The comparison of reconstruction results of different parameters shows that the parameters obtained by the proposed method can lead to better reconstruction results without redundant calculation burden.
- descent images /
- terrain reconstruction /
- 3D reconstruction /
- parameter /
- adaptive generation

HTML全文

参考文献(16)

[1]	Matthies L, Olson C, Tharp G, et al.Visual localization methods for Mars rovers using lander, rover, and descent imagery[C]//Proceedings of International Symposium on Artificial Intelligence, Robotics, and Automation in Space.Tokyo:[s.n.], 1997:413-418
[2]	Matthies L, Maimone M, Johnson A, et al.Computer vision on Mars[J].International Journal of Computer Vision, 2007, 75(1): 67-92
[3]	佟帅, 徐晓刚, 易成涛, 等.基于视觉的三维重建技术综述[J].计算机应用研究, 2011, 28(7):2411-2417 Tong Shuai, Xu Xiaogang, Yi Chengtao, et al.Overview on vision-based 3D reconstruction[J].Application Research of Computers, 2011, 28(7):2411-2417(in Chinese)
[4]	Olson C F, Abi-Rached H, Ye M, et al.Wide-baseline stereo vision for Mars rovers[C]//2003 IEEE/RSJ International Conference on Intelligent Robots and Systems.Las Vegas, NV:IEEE, 2003, 2:1302-1307
[5]	王炜强, 徐进, 杜歆, 等.基于宽基线立体视觉的远距离三维重建[J].浙江大学学报:工学版, 2010, 44(6):1073-1078 Wang Weiqiang, Xu Jin, Du Xin, et al.Long-range 3D reconstruction based on wide-baseline stereo vision[J].Journal of Zhejiang University:Engineering Science, 2010, 44(6):1073-1078(in Chinese)
[6]	Oliensis J, Genc Y.Fast algorithms for projective multi-frame structure from motion[C]//IEEE International Conference on Computer Vision.Piscataway, NJ:IEEE, 1999:536-543
[7]	Soatto S, Perona P.Reducing "structure from motion":a general framework for dynamic vision.1.modeling[J].Pattern Analysis and Machine Intelligence, IEEE Transactions on, 1998, 20(9):933-942
[8]	Tomasi C, Kanade T.Shape and motion from image streams under orthography:a factorization method[J].International Journal of Computer Vision, 1992, 9(2):137-154
[9]	Sturm P, Triggs B.A factorization based algorithm for multi-image projective structure and motion[C]//Computer Vision—ECCV '96.Heidelberg, Berlin:Springer, 1996:709-720
[10]	Xue X, Meng C, Jia Y.An improved method for terrain mapping from descent images[C]//Foundations of Intelligent Systems.Heidelberg, Berlin:Springer, 2012:547-555
[11]	Li R, Ma F, Xu F, et al.Mars rover localization using descent and rover imagery-result of the field test at Silver Lake, CA[C/CD]//ASPRS Annual Conference.Washington DC:ASPRS, 2000:22-26
[12]	Ma F, Di K, Li R, et al.Incremental mars rover localization using descent and rover imagery[C/CD]//ASPRS Annual Conference, St.Louis, Mo:ASPRS, 2001:25-27
[13]	Xiong Y, Olson C F, Matthies L H.Computing depth maps from descent images[J].Machine Vision and Applications, 2005, 16(3): 139-147
[14]	裴明涛, 贾云德. 摄像机沿光轴运动下的场景三维重建[J].计算机辅助设计与图形学学报, 2005, 17(3):534-539 Pei Mingtao, Jia Yunde.3d reconstruction under camera motion along optic axis[J].Journal of Computer Aided Design and Computer Graphics, 2005, 17(3):534-539(in Chinese)
[15]	张大志, 王勇涛, 田金文, 等.基于单目视觉系统的远距离场景重建算法研究[J].宇航学报, 2008, 29(1):289-294 Zhang Dazhi, Wang Yongtao, Tian Jinwen, et al.Efficient 3D reconstruction using monocular vision[J].Journal of Astronautics 2008, 29(1):289-294(in Chinese)
[16]	Meng C, Zhou N, Xue X, et al.Homography-based depth recovery with descent images[J].Machine Vision and Applications, 2013, 24(5):1093-1106