RTM框架下基于点线特征的视觉SLAM算法
|
摘要
针对图像纹理较为单一及相对模糊时仅仅依靠点特征难以实现精确位姿估计的问题,采用分散模块化技术提出了一种基于点线特征的视觉SLAM(同时定位与地图创建)算法.首先,提取相机采集环境中的点特征及线特征,并根据帧间特征匹配进行跟踪;随后,采用改进的NICP(normal iterative closest point)算法与关键帧匹配策略构建里程计系统.在此基础上,引入基于点线特征词典的闭环检测与GTSAM(Georgia Tech smoothing and mapping library)图优化方法获取具有全局一致性位姿的3维点云地图.以机器人技术中间件构筑系统框架,在提高系统实时性的同时增强功能模块的可扩展性与灵活性.标准数据集与实际实验室场景下的实验结果验证了所提方法的可行性和有效性.
When the image texture is simple and relatively indistinct, it is difficult to implement pose estimation based on point features. For this problem, a visual SLAM(simultaneous localization and mapping) algorithm based on point-line features is proposed by using the distributed modular technology. Firstly, the point and line features in the environment captured by the camera are extracted and tracked according to the inter-frame feature matching. Then, the improved NICP(normal iterative closest point) algorithm and the key frame matching strategy are used to build the odometer system. Based on this, the loop detection based on point-line feature dictionary and the graph optimization method of GTSAM(Georgia Tech smoothing and mapping library) are introduced to obtain 3D point cloud map with globally consistent poses. A system framework is developed with robot technology middleware to enhance the scalability and flexibility of the functional modules while improving the real-time performance of the system. The experimental results on the standard datasets and in laboratory scenes verify the feasibility and effectiveness of the proposed method.
When the image texture is simple and relatively indistinct, it is difficult to implement pose estimation based on point features. For this problem, a visual SLAM(simultaneous localization and mapping) algorithm based on point-line features is proposed by using the distributed modular technology. Firstly, the point and line features in the environment captured by the camera are extracted and tracked according to the inter-frame feature matching. Then, the improved NICP(normal iterative closest point) algorithm and the key frame matching strategy are used to build the odometer system. Based on this, the loop detection based on point-line feature dictionary and the graph optimization method of GTSAM(Georgia Tech smoothing and mapping library) are introduced to obtain 3D point cloud map with globally consistent poses. A system framework is developed with robot technology middleware to enhance the scalability and flexibility of the functional modules while improving the real-time performance of the system. The experimental results on the standard datasets and in laboratory scenes verify the feasibility and effectiveness of the proposed method.
引文
[1] Klein G, Murray D. Parallel tracking and mapping for small AR workspaces[C]//IEEE and ACM International Symposium on Mixed and Augmented Reality. Piscataway, USA:IEEE, 2007.DOI:10.1109/ISMAR.2007.4538852.
[2] Mur-Artal R, Montiel J M M, Tardós J D. ORB-SLAM:A versatile and accurate monocular SLAM system[J]. IEEE Transactions on Robotics, 2015, 31(5):1147-1163.
[3] Engel J, Sch?ps T, Cremers D. LSD-SLAM:Large-scale direct monocular SLAM[C]//13th European Conference on Computer Vision. Berlin, Germany:Springer, 2014:834-849.
[4] Neria J E, Ribeiro M I, Tardós J D. Mobile robot localization and map building using monocular vision[C/OL]//Proceedings of the 5th Symposium for Intelligent Robotics Systems. 1997:275-284. http://webdiis.unizar.es/GRPTR/pubs/Neira SIRS1997.pdf.
[5] Zhang G X, Lee J H, Lim J, et al. Building a 3-D line-based map using stereo SLAM[J]. IEEE Transactions on Robotics, 2015,31(6):1364-1377.
[6] Lu Y, Song D Z. Robust RGB-D odometry using point and line features[C]//IEEE International Conference on Computer Vision. Piscataway, USA:IEEE, 2015:3934-3942.
[7] Zuo X X, Xie X J, Liu Y, et al. Robust visual SLAM with point and line features[C]//IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway, USA:IEEE, 2017:1775-1782.
[8] Rublee E, Rabaud V, Konolige K, et al. ORB:An efficient alternative to SIFT or SURF[C]//IEEE International Conference on Computer Vision. Piscataway, USA:IEEE, 2011:2564-2571.
[9] 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, 2010,32(4):722-732.
[10] Rosten E, Porter R, Drummond T. Faster and better:A machine learning approach to corner detection[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2010, 32(1):105-119.
[11] Zhang L, Koch R. An efficient and robust line segment matching approach based on LBD descriptor and pairwise geometric consistency[J]. Journal of Visual Communication and Image Representation, 2013, 24(7):794-805.
[12] Min D B, Lu J B, Nguyen V A, et al. Weighted mode filtering and its applications to depth video enhancement and coding[C]//IEEE International Conference on Acoustics, Speech and Signal Processing. Piscataway, USA:IEEE, 2012:5433-5436.
[13] Serafin J, Grisetti G. NICP:Dense normal based point cloud registration[C]//IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway, USA:IEEE, 2015:742-749.
[14] Carlone L, Kira Z, Beall C, et al. Eliminating conditionally independent sets in factor graphs:A unifying perspective based on smart factors[C]//IEEE International Conference on Robotics and Automation. Piscataway, USA:IEEE, 2014:4290-4297.
[15]贾松敏,张国梁.基于模糊树图与DS证据理论的机器人功能模块粒度划分方法[J].机器人,2016,38(6):696-703.Jia S M, Zhang G L. Granularity partition method for robot functional modules based on fuzzy dendrogram and DS evidence theory[J]. Robot, 2016, 38(6):696-703.
[16] Pomerleau F, Magnenat S, Colas F, et al. Tracking a depth camera:Parameter exploration for fast ICP[C]//IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway, USA:IEEE, 2011:3824-3829.
[17] Steinbrücker F, Sturm J, Cremers D. Real-time visual odometry from dense RGB-D images[C]//IEEE International Conference on Computer Vision. Piscataway, USA:IEEE, 2011:719-722.
[18] Segal A V, Haehnel D, Thrun S. Generalized-ICP[C]//Robotics:Science and Systems V. 2009. DOI:10.15607/RSS.2009.V.021.
[19] Williams B, Cummins M, Neira J, et al. A comparison of loop closing techniques in monocular SLAM[J]. Robotics and Autonomous Systems, 2009, 57(12):1188-1197.
[20] Galvez-López D, Tardós J D. Bags of binary words for fast place recognition in image sequences[J]. IEEE Transactions on Robotics, 2012, 28(5):1188-1197.
[21] Sturm J, Engelhard N, Endres F, et al. A benchmark for the evaluation of RGB-D SLAM systems[C]//IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway, USA:IEEE, 2012:573-580.
[22] Mur-Artal R, Tardós J D. ORB-SLAM2:An open-source SLAM system for monocular, stereo, and RGB-D cameras[J].IEEE Transactions on Robotics, 2017, 33(5):1255-1262.
[23] Kerl C, Sturm J, Cremers D. Dense visual SLAM for RGB-D cameras[C]//IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway, USA:IEEE, 2013:2100-2106.
[24] Endres F, Hess J, Sturm J, et al. 3-D mapping with an RGB-D camera[J]. IEEE Transactions on Robotics, 2014, 30(1):177-187.
[2] Mur-Artal R, Montiel J M M, Tardós J D. ORB-SLAM:A versatile and accurate monocular SLAM system[J]. IEEE Transactions on Robotics, 2015, 31(5):1147-1163.
[3] Engel J, Sch?ps T, Cremers D. LSD-SLAM:Large-scale direct monocular SLAM[C]//13th European Conference on Computer Vision. Berlin, Germany:Springer, 2014:834-849.
[4] Neria J E, Ribeiro M I, Tardós J D. Mobile robot localization and map building using monocular vision[C/OL]//Proceedings of the 5th Symposium for Intelligent Robotics Systems. 1997:275-284. http://webdiis.unizar.es/GRPTR/pubs/Neira SIRS1997.pdf.
[5] Zhang G X, Lee J H, Lim J, et al. Building a 3-D line-based map using stereo SLAM[J]. IEEE Transactions on Robotics, 2015,31(6):1364-1377.
[6] Lu Y, Song D Z. Robust RGB-D odometry using point and line features[C]//IEEE International Conference on Computer Vision. Piscataway, USA:IEEE, 2015:3934-3942.
[7] Zuo X X, Xie X J, Liu Y, et al. Robust visual SLAM with point and line features[C]//IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway, USA:IEEE, 2017:1775-1782.
[8] Rublee E, Rabaud V, Konolige K, et al. ORB:An efficient alternative to SIFT or SURF[C]//IEEE International Conference on Computer Vision. Piscataway, USA:IEEE, 2011:2564-2571.
[9] 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, 2010,32(4):722-732.
[10] Rosten E, Porter R, Drummond T. Faster and better:A machine learning approach to corner detection[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2010, 32(1):105-119.
[11] Zhang L, Koch R. An efficient and robust line segment matching approach based on LBD descriptor and pairwise geometric consistency[J]. Journal of Visual Communication and Image Representation, 2013, 24(7):794-805.
[12] Min D B, Lu J B, Nguyen V A, et al. Weighted mode filtering and its applications to depth video enhancement and coding[C]//IEEE International Conference on Acoustics, Speech and Signal Processing. Piscataway, USA:IEEE, 2012:5433-5436.
[13] Serafin J, Grisetti G. NICP:Dense normal based point cloud registration[C]//IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway, USA:IEEE, 2015:742-749.
[14] Carlone L, Kira Z, Beall C, et al. Eliminating conditionally independent sets in factor graphs:A unifying perspective based on smart factors[C]//IEEE International Conference on Robotics and Automation. Piscataway, USA:IEEE, 2014:4290-4297.
[15]贾松敏,张国梁.基于模糊树图与DS证据理论的机器人功能模块粒度划分方法[J].机器人,2016,38(6):696-703.Jia S M, Zhang G L. Granularity partition method for robot functional modules based on fuzzy dendrogram and DS evidence theory[J]. Robot, 2016, 38(6):696-703.
[16] Pomerleau F, Magnenat S, Colas F, et al. Tracking a depth camera:Parameter exploration for fast ICP[C]//IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway, USA:IEEE, 2011:3824-3829.
[17] Steinbrücker F, Sturm J, Cremers D. Real-time visual odometry from dense RGB-D images[C]//IEEE International Conference on Computer Vision. Piscataway, USA:IEEE, 2011:719-722.
[18] Segal A V, Haehnel D, Thrun S. Generalized-ICP[C]//Robotics:Science and Systems V. 2009. DOI:10.15607/RSS.2009.V.021.
[19] Williams B, Cummins M, Neira J, et al. A comparison of loop closing techniques in monocular SLAM[J]. Robotics and Autonomous Systems, 2009, 57(12):1188-1197.
[20] Galvez-López D, Tardós J D. Bags of binary words for fast place recognition in image sequences[J]. IEEE Transactions on Robotics, 2012, 28(5):1188-1197.
[21] Sturm J, Engelhard N, Endres F, et al. A benchmark for the evaluation of RGB-D SLAM systems[C]//IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway, USA:IEEE, 2012:573-580.
[22] Mur-Artal R, Tardós J D. ORB-SLAM2:An open-source SLAM system for monocular, stereo, and RGB-D cameras[J].IEEE Transactions on Robotics, 2017, 33(5):1255-1262.
[23] Kerl C, Sturm J, Cremers D. Dense visual SLAM for RGB-D cameras[C]//IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway, USA:IEEE, 2013:2100-2106.
[24] Endres F, Hess J, Sturm J, et al. 3-D mapping with an RGB-D camera[J]. IEEE Transactions on Robotics, 2014, 30(1):177-187.