Enhanced Geometric Map: a 2D&3D Hybrid City Model of Large Scale Urban Environment for Robot Navigation
LI Haifeng1,2, HU Zunhe1, LIU Jingtai3,4
1. College of Computer Science and Technology, Civil Aviation University of China, Tianjin 300300, China;
2. Fujian Provincial Key Laboratory of Information Processing and Intelligent Control, Fuzhou 350121, China;
3. Institute of Robotics and Information Automatic System, Nankai University, Tianjin 300071, China;
4. Tianjin Key Laboratory of Intelligent Robotics, Tianjin 300071, China
Enhanced Geometric Map: a 2D&3D Hybrid City Model of Large Scale Urban Environment for Robot Navigation
LI Haifeng1,2, HU Zunhe1, LIU Jingtai3,4
1. College of Computer Science and Technology, Civil Aviation University of China, Tianjin 300300, China;
2. Fujian Provincial Key Laboratory of Information Processing and Intelligent Control, Fuzhou 350121, China;
3. Institute of Robotics and Information Automatic System, Nankai University, Tianjin 300071, China;
4. Tianjin Key Laboratory of Intelligent Robotics, Tianjin 300071, China
LI Haifeng, HU Zunhe, LIU Jingtai. Enhanced Geometric Map: a 2D&3D Hybrid City Model of Large Scale Urban Environment for Robot Navigation[J]. 机器人, 2016, 38(3): 311-321.DOI: 10.13973/j.cnki.robot.2016.0321.
LI Haifeng, HU Zunhe, LIU Jingtai. Enhanced Geometric Map: a 2D&3D Hybrid City Model of Large Scale Urban Environment for Robot Navigation. ROBOT, 2016, 38(3): 311-321. DOI: 10.13973/j.cnki.robot.2016.0321.
摘要To facilitate scene understanding and robot navigation in large scale urban environment, a two-layer enhanced geometric map (EGMap) is designed using videos from a monocular onboard camera. The 2D layer of EGMap consists of a 2D building boundary map from top-down view and a 2D road map, which can support localization and advanced map-matching when compared with standard polyline-based maps. The 3D layer includes features such as 3D road model, and building facades with coplanar 3D vertical and horizontal line segments, which can provide the 3D metric features to localize the vehicles and flying-robots in 3D space. Starting from the 2D building boundary and road map, EGMap is initially constructed using feature fusion with geometric constraints under a line feature-based simultaneous localization and mapping (SLAM) framework iteratively and progressively. Then, a local bundle adjustment algorithm is proposed to jointly refine the camera localizations and EGMap features. Furthermore, the issues of uncertainty, memory use, time efficiency and obstacle effect in EGMap construction are discussed and analyzed. Physical experiments show that EGMap can be successfully constructed in large scale urban environment and the construction method is demonstrated to be very accurate and robust.
Abstract：To facilitate scene understanding and robot navigation in large scale urban environment, a two-layer enhanced geometric map (EGMap) is designed using videos from a monocular onboard camera. The 2D layer of EGMap consists of a 2D building boundary map from top-down view and a 2D road map, which can support localization and advanced map-matching when compared with standard polyline-based maps. The 3D layer includes features such as 3D road model, and building facades with coplanar 3D vertical and horizontal line segments, which can provide the 3D metric features to localize the vehicles and flying-robots in 3D space. Starting from the 2D building boundary and road map, EGMap is initially constructed using feature fusion with geometric constraints under a line feature-based simultaneous localization and mapping (SLAM) framework iteratively and progressively. Then, a local bundle adjustment algorithm is proposed to jointly refine the camera localizations and EGMap features. Furthermore, the issues of uncertainty, memory use, time efficiency and obstacle effect in EGMap construction are discussed and analyzed. Physical experiments show that EGMap can be successfully constructed in large scale urban environment and the construction method is demonstrated to be very accurate and robust.
基金资助:National Natural Science Foundation of China (61305107, U1333109); the Fundamental Research Funds for the Central Universities (3122016B006); the Scientific Research Funds for Civil Aviation University of China (2012QD23X)
通讯作者:
LI Haifeng, hfli@cauc.edu.cn
E-mail: hfli@cauc.edu.cn
作者简介: LI Haifeng (1984-)，male, ph.D, lecturer. His research interestincludes computer vision, robot navigation. HU Zunhe (1990-)，male, master candidate. His researchinterest includes computer vision, visual SLAM. LIU Jingtai (1964-)，male, ph.D, professor. His research interestincludes robotic technology, computer applications,automatic information system, intelligence scienceand technology.
[1] Quddus W, Ochieng W Y, Noland R B. Current map-matching algorithms for transport applications: State-of-the art and future research directions[J]. Transportation Research Part C: Emerging Technologies, 2007, 15(5): 312-328.
[2] Pinana-Diaz C, Toledo-Moreo R, Javier T, et al. A two-layers based approach of an enhanced-map for urban positioning support[ J]. Sensors, 2012, 12(11): 14508-14524.
[3] Meguro J-I, Murata T, Takiguchi, J-I, et al. GPS multipath mitigation for urban area using omnidirectional infrared camera[J]. IEEE Transactions on Intelligent Transportation Systems, 2009, 10(1): 22-30.
[4] Frueh C, Jain S, Zakhor A. Data processing algorithms for generating textured 3D building facade meshes from laser scans and camera images[J]. International Journal of Computer Vision, 2005, 61(2): 159-184.
[5] Zebedin L, Bauer J, Karner K, et al. Fusion of feature- and areabased information for urban buildings modeling from aerial imagery[ C]//10th European Conference on Computer Vision: Part IV. Berlin, Germany: Springer-Verlag, 2008: 873-886.
[6] Gee A P, Chekhlov D, Calway A, et al. Discovering higher level structure in visual SLAM[J]. IEEE Transactions on Robotics, 2008, 24(5): 980-990.
[7] Martinez-Carranza J, Calway A. Unifying planar and point mapping in monocular SLAM[C]//British Machine Vision Conference. UK: British Machine Vision Association, 2010: 43.1- 43.11.
[8] Cornelis N, Leibe B, Cornelis K, et al. 3D urban scene modeling integrating recognition and reconstruction[J]. International Journal of Computer Vision, 2008, 78(2-3): 121-141.
[9] Seitz S M, Curless B, Diebel J, et al. A comparison and evaluation of multi-view stereo reconstruction algorithms[C]//IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Piscataway, USA: IEEE, 2006: 519-526.
[10] Mouragnon E, Lhuillier M, Dhome M, et al. Generic and realtime structure from motion using local bundle adjustment[J]. Image and Vision Computing, 2009, 27(8): 1178-1193.
[11] Lowe D. Distinctive image features from scale-invariant keypoints[ J]. International Journal of Computer Vision, 2004, 60(2): 91-110.
[12] Civera J, Grasa O G, Davison A J, et al. 1-point RANSAC for EKF-based structure from motion[C]//IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway, USA: IEEE, 2009: 3498-3504.
[13] Smith P, Reid I, Davison A. Real-time monocular SLAM with straight lines[C]//British Machine Vision Conference. UK: British Machine Vision Association, 2006: 17-26.
[14] Lemaire T, Lacroix S. Monocular-vision based SLAM using line segments[C]//IEEE International Conference on Robotics and Automation. Piscataway, USA: IEEE, 2007: 2791-2796.
[15] Li H, Song D Z, Liu Y, et al. A two-view based multilayer feature graph for robot navigation[C]//IEEE International Conference on Robotics and Automation. Piscataway, USA: IEEE, 2012: 3580-3587.
[16] Lu Y, Song D Z, Xu Y L, et al. Automatic building exterior mapping using multilayer feature graphs[C]//IEEE International Conference on Automation Science and Engineering. Piscataway, USA: IEEE, 2013: 162-167.
[17] Lu Y, Song D Z. Visual navigation using heterogeneous landmarks and unsupervised geometric constraints[J]. IEEE Transactions on Robotics, 2015, 31(3): 736-749.
[18] Li H F, Xiang J L, Liu J T. An automatic building extraction method from high resolution satellite image[C]//Chinese Control Conference. Piscataway, USA: IEEE, 2012: 4884-4889.
[19] Fischler 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): 381-395.
[20] von Gioi R G, Jakubowicz J, Morel J M, et al. LSD: A fast line segment detector with a false detection control[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 20010, 32(4): 722-732.
[21] Fan B, Wu F C, Hu Z Y. Line matching leveraged by point correspondences[ C]//IEEE International Conference on Computer Vision and Pattern Recognition. Piscataway, USA: IEEE, 2010: 390-397.
[22] Li H F, Liu J T. An optimal vanishing point detection method with error analysis[J]. Acta Automatica Sinica, 2012, 38(2): 213-219.
[23] Li H F, Wang H P, Liu J T. Error aware multiple vertical planes based visual localization for mobile robots in urban environments[ J]. Science China: Information Sciences, 2015, 58(3): 1-14.
