Abstract
AbstractIn this paper, we propose a sparse point-plane odometry used in structured environments. Compared to a point-based odometry, we add additional planar constraints into the process of optimization, making the results more reliable. A novel grid-based plane detection algorithm is proposed to cluster sparse points in the same planes. Then, the planes are parameterized by inverse normal and take part in the windowed optimization. By reducing the size of Hessian Matrix, the process of optimization converges faster. Compared to the original point-based odometry, the proposed method performs better on both robustness and efficiency in structured environments.
Publisher
Cambridge University Press (CUP)
Subject
Computer Science Applications,General Mathematics,Software,Control and Systems Engineering,Control and Optimization,Mechanical Engineering,Modeling and Simulation
Reference24 articles.
1. [16] Feng, C. , Taguchi, Y. and Kamat, V. R. , “Fast Plane Extraction in Organized Point Clouds Using Agglomerative Hierarchical Clustering,” In: IEEE International Conference on Robotics and Automation (ICRA) (2014) pp. 6218–6225.
2. Visual SLAM algorithms: a survey from 2010 to 2016
3. Probabilistic RGB-D odometry based on points, lines and planes under depth uncertainty
4. [24] Engel, J. , Usenko, V. and Cremers, D. , “A photometrically calibrated benchmark for monocular visual odometry,” arXiv 2016, arXiv:1607.02555.
5. [23] Yang, S. , Song, Y. , Kaess, M. and Scherer, S. , “Pop-up SLAM: Semantic Monocular Plane SLAM for Low-Texture Environments,” In: IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (2016) pp. 1222–1229.
Cited by
1 articles.
订阅此论文施引文献
订阅此论文施引文献，注册后可以免费订阅5篇论文的施引文献，订阅后可以查看论文全部施引文献