Two improvements of KinectFusion algorithm are proposed. On one hand, the edge feature points in the environment are matched to improve the robustness, on the other hand, a ground plane point cloud in the model is preset to improve the accuracy. A standard RGB-D (RGB-Depth) SLAM (simultaneous localization and mapping) benchmark dataset and a data of the lab environment are modeled, and the comparison results show that, both the robustness and the accuracy are improved obviously after the improvement. The improved algorithm decreases the modeling error from 4.5% to 1.5% in a room of 6m×3m×3m. Although the efficiency is influenced, the running speed of the algorithm is still very high, and the user experience during modeling is good.
[1] Henry P, Krainin M, Herbst E, et al. RGB-D mapping: Using depth cameras for dense 3D modeling of indoor environments[C]//RSS Workshop on RGB-D Cameras. 2010.[2] Fioraio N, Konolige K. Realtime visual and point cloud SLAM[C]//RSS Workshop on RGB-D Cameras. 2011.[3] Izadi S, Kim D, Hilliges O, et al. KinectFusion: Real-time 3D reconstruction and interaction using a moving depth camera[C]//Proceedings of the 24th Annual ACM Symposium on User Interface Software and Technology. New York, USA: ACM, 2011: 559-568.[4] Newcombe R A, Izadi S, Hilliges O, et al. KinectFusion: Real-time dense surface mapping and tracking[C]//10th IEEE International Symposium on Mixed and Augmented Reality. Piscataway, USA: IEEE, 2011: 127-136.[5] Zeng M, Zhao F K, Zheng J X, et al. Octree-based fusion for realtime 3D reconstruction[J]. Graphical Models, 2013, 75(3): 126-136. [6] Zeng M, Zhao F K, Zheng J X, et al. A memory-efficient KinectFusion using octree[C]//First International Conference Computational Visual Media. Berlin, Germany: Springer, 2012: 234-241.[7] Whelan T, McDonald J, Kaess M, et al. Kintinuous: Spatially extended KinectFusion[C]//RGB-D Workshop at Robotics: Science and Systems. 2012.[8] Herrera D C, Kannala J, Heikkila J. Joint depth and color camera calibration with distortion correction[J]. IEEE Transactionson Pattern Analysis and Machine Intelligence, 2012, 34(10): 2058-2064. [9] Curless B, Levoy M. A volumetric method for building complex models from range images[C]//Proceedings of ACM SIGGRAPH. New York, USA: ACM, 1996: 303-312.[10] 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. [11] 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.[12] PCL. PCL Library[OL] (2013-03-06)[2013-03-06]. http://www. Pointcloud s.org/.
