DENG Lei1,2, CHEN Bao-hua1,2, HUANG Si-yuan1,2, DUAN Yue-qi1,2, ZHOU Jie1,2,3
1. Department of Automation, Tsinghua University, Beijing 100084, China;
2. Tsinghua National Laboratory for Information Science and Technology, Beijing 100084, China;
3. State Key Laboratory of Intelligent Technology and Systems, Beijing 100084, China
Image-based localization has wide applications including auto navigation,path planning,virtual reality,etc.For the end user,a single image is required.In this paper,we present a ray-based image localization system including ray-based 3D reconstruction and pose graph optimization-based localization.By using the inherit geometry properties of ray-based camera model,the proposed reconstruction algorithm can deal with various wide FOV camera models (panorama and fisheye) to reduce the reconstruction cost and achieve a better reconstruction result.The proposed pose graph optimization localization framework can combine 2D-3D matches information and relative pose information for a better localization accuracy.Experiment result shows the effectiveness of the proposed system.
[1] 左敏,曾广平,涂序彦.无人变电站智能机器人的视觉导航研究[J].电子学报,2012,39(10):2464-2468. Zuo Min,Zeng Guang-ping,Tu Xu-yuan.Research on visual navigation of untended substation patrol robot[J].Acta Electronica Sinica,2012,39(10):2464-2468.(in Chinese)
[2] 郝宗波,洪炳熔.未知环境下基于传感器的移动机器人路径规划[J].电子学报,2006,34(5):953-956. Hao Zong-bo,Hong Bing-rong.Sensor-based path planning for mobile robot in unknown environment[J].Acta Electronica Sinica,2006,34(5):953-956.(in Chinese)
[3] Schindler G,Brown M,Szeliski R.City-scale location recognition[A].Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition[C].Minneapolis,USA:IEEE Press,2007.1-7.
[4] Zamir A,Shah M.Accurate image localization based on google maps street view[A].Lecture Notes in Computer Science[C].Berlin Heidelberg:GER,2010.255-268.
[5] Sattler T,Leibe B,Kobbelt L.Fast image-based localization using direct 2D-to-3D matching[A].Proceedings of IEEE International Conference on Computer Vision[C].Barcelona,ESP:IEEE Press,2011.667-674.
[6] Li Yunpeng,Snavely N,Huttenlocher D,et al.Worldwide pose estimation using 3D point clouds[A].Lecture Notes in Computer Science[C].Berlin Heidelberg:GER,2012.15-29.
[7] Snavely N,Seitz S M,Szeliski R.Photo tourism:exploring photo collections in 3D[J].ACM Transactions on Graphics,2006,25(3):835-846.
[8] Wu ChangChang.Towards linear-time incremental structure from motion[A].International Conference on 3D Vision[C].Seattle,Washington,USA,2013.127-134.
[9] Hartley R,Zisserman A.Multiple View Geometry in Computer Vision[M].England:Cambridge University Press,2004.1187-1865.
[10] Bill Triggs,Philip Mclauchlan,Richard Hartley,Andrew Fitzgibbon.Bundle adjustment-a modern synthesis[A].Vision Algorithms:Theory and Practice,LNCS[C].Berlin Heidelberg:Springer Verlag 2000.153-177.
[11] Lowe D.Distinctive image features from scale-invariant keypoints[J].International Journal of Computer Vision,2004,60(2):91-110.
[12] Fisher 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,1981,24(6):726-740.
[13] Bujnak M,Kukelova Z,Pajdla T.A general solution to the P4P problem for camera with unknown focal length[A].Proceedings of IEEE Conference on Computer Vision and Pattern Recognition[C].Anchorage,USA:IEEE Press,2008.1-8.
[14] Kukelova Z,Bujnak M,Pajdla T.Real-time solution to the absolute pose problem with unknown radial distortion and focal length[A].Proceedings of IEEE International Conference on Computer Vision[C].Sydney,USA:IEEE Press,2013.2816-2823.
2017, Vol. 45 
