全部 标题 作者
关键词 摘要

OALib Journal期刊
ISSN: 2333-9721
费用:99美元

查看量下载量

相关文章

更多...
电子学报  2014 

针对摄像机矩阵估计的增强连续禁忌搜索方法

DOI: 10.3969/j.issn.0372-2112.2014.12.013, PP. 2422-2428

Keywords: 多视几何,摄像机矩阵,全局最优,禁忌搜索

Full-Text   Cite this paper   Add to My Lib

Abstract:

摄像机矩阵估计是机器视觉的一个重要问题.在2范数误差代价函数模型下,最小二乘法简单而有效,但因误差代价函数非凸,容易陷入局部最优.在无穷范数误差代价函数模型下,凸优化方法理论上可以获得全局最优,但计算效率较低,其计算耗时随着问题规模的增大而急剧增加.现代优化论中的增强连续禁忌搜索(Enhancedcontinuoustaboosearch,ECTS)方法具有逃离局部最优的优良性质,因此本文在2范数误差代价函数模型下提出一种针对摄像机矩阵估计的ECTS算法.在ECTS置信区间序列构造及最大置信区间选择环节,本文提出了一种非迭代的方法获取包含全局最优解的凸包.在增强禁忌搜索环节,本文提出了一种基于伪凸函数的候选解邻域构造方法.同时,给出了本文算法以概率1收敛于全局最优的理论证明.对虚拟场景和真实场景的实验结果表明本文算法可以快速获取摄像机矩阵估计的全局最优解.

References

[1]  Hartley R,Zisserman A.Multiple View Geometry in Computer Vision[M].Cambridge:Cambridge University Press,2004.86-98.
[2]  Triggs B,McLauchlan P,et al.Bundle Adjustment a Modern Synthesis[M].Berlin Heidelberg:Springer,2000.298-372.
[3]  Chelouah R,Siarry P.Tabu search applied to global optimization[J].European Journal of Operational Research,2000,123(2):256-270.
[4]  陈晓峰,姜慧研.量子禁忌搜索算法的研究[J].电子学报,2013,41(11):2161-2166. Chen X F,Jiang H Y.Research of quantum tabu search algorithm[J].Acta Electronica Sinica,2013,41(11):2161-2166.(in Chinese)
[5]  许鹏飞,苗启广,李伟生,张军英.基于函数复杂度的自适应模拟退火和禁忌搜索新算法[J].电子学报,2012,40(6):1218-1222. Xu P F,Miao Q G,Li W S,Zhang J Y.Adaptive simulated annealing algorithm and tabu search algorithm based on the function complexity[J].Acta Electronica Sinica,2012,40(06):1218-1222.(in Chinese)
[6]  Boyd S,Vandenberghe L.Convex Optimization[M].Cambridge Univ Press,2004.95-104.
[7]  Olsson C,Kahl F.Generalized convexity in multiple view geometry[J].Journal of Mathematical Imaging and Vision,2010,38(1):35-51.
[8]  Ji M,Tang H.Global optimizations and tabu search based on memory[J].Applied Mathematics and Computation,2004,159(2):449-457.
[9]  Kahl F.Multiple View Geometry and the L-infinity Norm[DB/OL].http://www.maths.lth.se/matematiklth/personal/fredrik/download.html,2013.
[10]  Werner T,Zisserman A.New techniques for automated architecture reconstruction from photographs proc[A].7th European Conference on Computer Vision[C].Copenhagen,Denmark:ECCV,2002.541-555.
[11]  Ke Q,Kanade T.Quasiconvex optimization for robust geometric reconstruction[A].IEEE International Conference on Computer Vision[C].Beijing:IEEE,2005.986-993.
[12]  Kahl F.Multiple view geometry and the L∞-norm[A].IEEE International Conference on Computer Vision[C].Beijing:IEEE,2005.1002-1009.
[13]  Olsson C,Eriksson A,et al.Efficient optimization of L∞-problems using pseudoconvexity[A].IEEE International Conference on Computer Vision[C].Riode Janeiro:IEEE,2007.1-8.
[14]  Agarwal S,Snavely N,et al.Fast algorithms for L∞-Norm problems in multiview geometry[A] IEEE Conference on Computer Vision and Pattern Recognition[C].Alaska:IEEE,2008.1-8.
[15]  Dai Z,Wu Y,Zhang F,et al.A novel fast method for L∞ problems in multiview geometry[A].Computer Vision-ECCV 2012[C].Florence:Springer Berlin Heidelberg,2012.116-129.
[16]  Hartley R,Kahl F,et al.Verifying global minima for L2 minimization problems in multiple view geometry[J].International Journal of Computer Vision,2013,101(2):288-304.
[17]  张闯,迟健男,张朝晖,王志良.一种基于立体视觉的视线估计方法[J].电子学报,2010,38(05):1008-1014. Zhang C,Chi J N,Zhang Z H,Wang Z L.A novel gaze estimation method based on stereo vision[J].Acta Electronica Sinica,2010,38(05):1008-1014.(in Chinese)
[18]  Glover F.Tabu search methods in artificial intelligence and operations research[J].ORSA Artificial Intelligence,1987,1(2):6.

Full-Text

Contact Us

service@oalib.com

QQ:3279437679

WhatsApp +8615387084133