基于RGB-D的云机器人3D SLAM实验系统
|
摘要
针对室内环境的三维视觉同步定位与地图构建(3D VSLAM)计算量大、耗时长、硬件要求高的问题,提出了一种基于RGB-D的云机器人VSLAM实验平台。采用Kinect传感器,获取环境的RGB图像和深度信息,采用金字塔Lucas-Kanade算法实现帧间FAST特征点的快速追踪与匹配,运用RANSAC算法进行初始配准,提取关键帧。借助于云计算动态供给、弹性计算的优势,将VSLAM中计算消耗大的精确配准、闭环检测和全局优化处理过程卸载至云端进行,以减轻本地处理器的运算负担。实验结果表明,该方法能够有效地减轻VSLAM对硬件的依赖度,缩短SLAM的执行时间并提高构图精度,为云机器人以较低的成本实现先进SLAM算法提供了有效的解决途径。
In view of the problems of large computation, time-consuming and high hardware requirements of 3 D-VSLAM in indoor environment, a cloud robot VSLAM experimental platform based on RGB-D is proposed. The Kinect sensor is used to acquire RGB image and depth information of the environment, the pyramid Lucas-Kanade algorithm is adopted to realize the fast tracking and matching of FAST feature points between frames, and the RANSAC algorithm is used for initial registration and key frame extraction. By virtue of the advantages of dynamic supply and flexible computation in cloud computing, the process of the accurate registration, closed-loop detection and global optimization in VSLAM is unloaded to the cloud to reduce the computing burden of local processors. The experimental results show that this method can effectively reduce the hardware dependence of VSLAM, shorten the execution time of SLAM and improve the composition accuracy. It provides an effective solution for cloud robots to implement the advanced SLAM algorithm at a lower cost.
In view of the problems of large computation, time-consuming and high hardware requirements of 3 D-VSLAM in indoor environment, a cloud robot VSLAM experimental platform based on RGB-D is proposed. The Kinect sensor is used to acquire RGB image and depth information of the environment, the pyramid Lucas-Kanade algorithm is adopted to realize the fast tracking and matching of FAST feature points between frames, and the RANSAC algorithm is used for initial registration and key frame extraction. By virtue of the advantages of dynamic supply and flexible computation in cloud computing, the process of the accurate registration, closed-loop detection and global optimization in VSLAM is unloaded to the cloud to reduce the computing burden of local processors. The experimental results show that this method can effectively reduce the hardware dependence of VSLAM, shorten the execution time of SLAM and improve the composition accuracy. It provides an effective solution for cloud robots to implement the advanced SLAM algorithm at a lower cost.
引文
[1] KEHOE B, PATIL S, ABBEEL P, et al. A survey of research on cloud robotics and automation[J]. IEEE Transactions on Automation Science and Engineering, 2015, 12(2):398-409.
[2] MANZI A, FIORINI L, ESPOSITO R, et al. Design of a cloud robotic system to support senior citizens: the KuBo experience[J]. Autonomous Robots, 2017,43(3):699-709.
[3] HUNZIKER D, GAJAMOHAN M, WAIBEL M, et al. Rapyuta: The roboearth cloud engine[C]// IEEE International Conference on Robotics and Automation (ICRA), 2013: 438-444.
[4] THRUN S, LIU Y, KOLLER D, et al. Simultaneous localization and mapping with sparse extended information filters[J]. The International Journal of Robotics Research, 2004, 23(7/8):693-716.
[5] DURRANT-WHYTE H, Bailey T. Simultaneous localization and mapping[J]. IEEE Robotics and Automation Magazine, 2006, 13(2):99-110.
[6] HENRY P, KRAININ M, HERBST E, et al. RGB-D mapping: Using Kinect-style depth cameras for dense 3D modeling of indoor environments[J]. International Journal of Robotics Research, 2012, 31(5):647-663.
[7] ENDRES F, HESS J, ENGELHARD N, et al. An evaluation of the RGB-D SLAM system[C]// 2012 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2012: 1691-1696.
[8] CLIPP B, LIM J, FRAHM J M, et al. Parallel, real-time visual SLAM[C]// 2010 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).IEEE, 2010: 3961-3968.
[9] 付梦印, 吕宪伟, 刘彤,等. 基于RGB-D数据的实时SLAM算法[J]. 机器人, 2015, 37(6):683-692.
[10] 杨扬, 曹其新, 朱笑笑, 等. 面向机器人手眼协调抓取的3维建模方法[J]. 机器人, 2013, 35(2):151-155.
[11] AYUSH K, AGARWAL N K. Real time visual SLAM using cloud computing[C]//2013 Fourth International Conference on Computing, Communications and Networking Technologies (ICCCNT). IEEE, 2013: 1-7.
[12] BENAVIDEZ P, MUPPIDI M, RAD P, et al. Cloud-based realtime robotic Visual SLAM[C]// 2015 9th Annual IEEE International Systems Conference (SysCon). IEEE, 2015: 773-777.
[13] MOHANARAJAH G, USENKO V, SINGH M, et al. Cloud-based collaborative 3D mapping in real-time with low-cost robots[J]. IEEE Transactions on Automation Science and Engineering, 2015, 12(2):423-431.
[14] BOUGUET J Y. Pyramidal Implementation of the Lucas Kanade Feature Tracker Description of the algorithm[J]. Intel Corporation Microprocessor Research Labs Tech Rep, 2000, 22(2):363-381.
[15] BESL P J, MCKAY H D. A method for registration of 3-D shapes[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 1992, 14(2):239-256.
[16] ZHANG H, LI B, YANG D. Keyframe detection for appearance-based visual SLAM[C]// IEEE/RSJ International Conference on Intelligent Robots and Systems. 2010: 2071-2076.
[17] 梁明杰, 闵华清, 罗荣华. 基于图优化的同时定位与地图创建综述[J]. 机器人, 2013, 35(4):500-512.
[18] KERL C, STURM J, CREMERS D. Dense visual SLAM for RGB-D cameras[C]// IEEE/RSJ International Conference on Intelligent Robots and Systems. 2013:2100-2106.
[19] SILBERMAN N, FERGUS R. Indoor scene segmentation using a structured light sensor[C]// IEEE International Conference on Computer Vision Workshops. IEEE, 2011:601-608.
[2] MANZI A, FIORINI L, ESPOSITO R, et al. Design of a cloud robotic system to support senior citizens: the KuBo experience[J]. Autonomous Robots, 2017,43(3):699-709.
[3] HUNZIKER D, GAJAMOHAN M, WAIBEL M, et al. Rapyuta: The roboearth cloud engine[C]// IEEE International Conference on Robotics and Automation (ICRA), 2013: 438-444.
[4] THRUN S, LIU Y, KOLLER D, et al. Simultaneous localization and mapping with sparse extended information filters[J]. The International Journal of Robotics Research, 2004, 23(7/8):693-716.
[5] DURRANT-WHYTE H, Bailey T. Simultaneous localization and mapping[J]. IEEE Robotics and Automation Magazine, 2006, 13(2):99-110.
[6] HENRY P, KRAININ M, HERBST E, et al. RGB-D mapping: Using Kinect-style depth cameras for dense 3D modeling of indoor environments[J]. International Journal of Robotics Research, 2012, 31(5):647-663.
[7] ENDRES F, HESS J, ENGELHARD N, et al. An evaluation of the RGB-D SLAM system[C]// 2012 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2012: 1691-1696.
[8] CLIPP B, LIM J, FRAHM J M, et al. Parallel, real-time visual SLAM[C]// 2010 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).IEEE, 2010: 3961-3968.
[9] 付梦印, 吕宪伟, 刘彤,等. 基于RGB-D数据的实时SLAM算法[J]. 机器人, 2015, 37(6):683-692.
[10] 杨扬, 曹其新, 朱笑笑, 等. 面向机器人手眼协调抓取的3维建模方法[J]. 机器人, 2013, 35(2):151-155.
[11] AYUSH K, AGARWAL N K. Real time visual SLAM using cloud computing[C]//2013 Fourth International Conference on Computing, Communications and Networking Technologies (ICCCNT). IEEE, 2013: 1-7.
[12] BENAVIDEZ P, MUPPIDI M, RAD P, et al. Cloud-based realtime robotic Visual SLAM[C]// 2015 9th Annual IEEE International Systems Conference (SysCon). IEEE, 2015: 773-777.
[13] MOHANARAJAH G, USENKO V, SINGH M, et al. Cloud-based collaborative 3D mapping in real-time with low-cost robots[J]. IEEE Transactions on Automation Science and Engineering, 2015, 12(2):423-431.
[14] BOUGUET J Y. Pyramidal Implementation of the Lucas Kanade Feature Tracker Description of the algorithm[J]. Intel Corporation Microprocessor Research Labs Tech Rep, 2000, 22(2):363-381.
[15] BESL P J, MCKAY H D. A method for registration of 3-D shapes[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 1992, 14(2):239-256.
[16] ZHANG H, LI B, YANG D. Keyframe detection for appearance-based visual SLAM[C]// IEEE/RSJ International Conference on Intelligent Robots and Systems. 2010: 2071-2076.
[17] 梁明杰, 闵华清, 罗荣华. 基于图优化的同时定位与地图创建综述[J]. 机器人, 2013, 35(4):500-512.
[18] KERL C, STURM J, CREMERS D. Dense visual SLAM for RGB-D cameras[C]// IEEE/RSJ International Conference on Intelligent Robots and Systems. 2013:2100-2106.
[19] SILBERMAN N, FERGUS R. Indoor scene segmentation using a structured light sensor[C]// IEEE International Conference on Computer Vision Workshops. IEEE, 2011:601-608.