Abstract The 3D reconstruction system is mostly based on the simultaneous localization and mapping(SLAM) system of the feature point method and the direct method. The SLAM of feature point method cannot obtain good reconstruction results in the absence of feature points, while the SLAM of direct method has difficulty in estimating the pose with a fast-moving camera, and consequently, reconstruction results are unsatisfactory. To solve these problems, a dense 3D scene reconstruction system with a depth camera (RGB-D camera) based on semi-direct SLAM is proposed in this paper. The feature point method is exploited to estimate the camera pose in feature-rich areas. In the area of missing feature points, the direct method is utilized to estimate the pose of the camera. Then, the three-dimensional map is constructed by the optimized camera pose. The furfel model and the deformation map are utilized to estimate the pose of the point cloud and fuse point cloud. Finally, the ideal 3D reconstruction model is obtained. Experiments show that the system can be applied to all three-dimensional reconstruction of various occasions and acquire the ideal three-dimensional reconstruction model.
About author:: XU Haonan, master student. His research interests include visual SLAM and 3D reconstruction. YU Lei, Ph.D., associate professor. His research interests include artificial intelligence and 3D reconstruction. FEI Shumin, Ph.D., professor. His research interests include artificial intelligence and nonlinear system.
[1] TONG J, ZHOU J, LIU L G, et al. Scanning 3D Full Human Bodies Using Kinects. IEEE Transactions on Visualization and Computer Graphics, 2012, 18(4): 643-650. [2] DAVISON A J, REID I D, MOLTON N D, et al. MonoSLAM: Real-Time Single Camera SLAM. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2007, 29(6): 1052-1067. [3] KLEIN G, MURRAY D. Parallel Tracking and Mapping for Small AR Workspaces//Proc of the 6th IEEE and ACM International Symposium on Mixed and Augmented Reality. Washington, USA: IEEE, 2007: 225-234. [4] CIVERA J, DAVISON A J, MONTIELJ M. Inverse Depth Parametrization for Monocular SLAM. IEEE Transactions on Robotics, 2008, 24(5): 932-945. [5] HUANG G P, MOURIKIS A I, ROUMELIOTIS S I. A Quadratic-Complexity Observability-Constrained Unscented Kalman Filter for SLAM. IEEE Transactions on Robotics, 2013, 29(5): 1226-1243. [6] MUR-ARTAL R, MONTIEL J, TARDS J D. ORB-SLAM: A Versatile and Accurate Monocular SLAM System. IEEE Transactions on Robotics, 2017, 31(5): 1147-1163. [7] 王 楠,马书根,李 斌,等.震后建筑内部层次化SLAM的地图模型转换方法.自动化学报, 2015, 41(10): 1723-1733. (WANG N, MA S G, LI B, et al. A Model Transformation of Map Representation for Hierarchical SLAM That Can Be Used for After-earthquake Buildings. Acta Automatica Sinica, 2015, 41(10): 1723-1733.) [8] ENGEL J, SCH PS T, CREMERS D. LSD-SLAM: Large-Scale Direct Monocular SLAM//Proc of the European Conference on Computer Vision. Berlin, Germany: Springer, 2014: 834-849. [9] ENDRES F, HESS J, STURM J, et al. 3D Mapping with an RGB-D Camera. IEEE Transactions on Robotics, 2014, 30(1): 177-187. [10] BACHRACH A, PRENTICE S, HE R J, et al. Estimation, Planning, and Mapping for Autonomous Flight Using an RGB-D Camera in GPS-Denied Environments. International Journal of Robotics Research, 2012, 31(11): 1320-1343. [11] HENRY P, KRAININ M, HERBST E, et al. RGB-D Mapping: Using Depth Cameras for Dense 3D Modeling of Indoor Environments//Proc of the 12th International Symposium on Experimental Robotics. Berlin, Germany: Springer, 2010: 477-492. [12] IZADI S, KIM D, HILLIGES O, et al. KinectFusion: Real-Time 3D Reconstruction and Interaction Using a Moving Depth Camera//Proc of the 24th Annual ACM Symposium on User Interface Software and Technology. New York, USA: ACM, 2011: 559-568. [13] WHELAN T, LEUTENEGGER S, SALAS-MORENO R S. ElasticFusion: Dense SLAM without a Pose Graph[C/OL]. [2017-12-20]. http://www.thomaswhelan.ie/Whelan15rss.pdf. [14] WHELAN T, SALAS-MORENO R F, GLOCKER B. ElasticFu- sion: Real-Time Dense SLAM and Light Source Estimation. International Journal of Robotics Research, 2016, 35(14): 1697-1716. [15] DOU M S, KHAMIS S, DEGTYAREV Y, et al. Fusion4D: Real-Time Performance Capture of Challenging Scenes. ACM Transactions on Graphics, 2016, 35(4). DOI: 10.1145/2897824.2925969. [16] WHELAN T, KAESS M, JOHANNSSON H, et al. Real-Time Large-Scale Dense RGB-D SLAM with Volumetric Fusion. International Journal of Robotics Research, 2015, 34(4/5): 598-626. [17] MUR-ARTAL R, TARDOS J D. ORB-SLAM2: An Open-Source SLAM System for Monocular, Stereo, and RGB-D Cameras. IEEE Transactions on Robotics, 2017, 33(5): 1255-1262.
2018, Vol. 31 
