An Acceleration Method Using CUDA based on ORB-SLAM2

In order to solve the problem of poor real-time performance and low position accuracy when mobile robots apply the technology of simultaneous localization and mapping, this paper proposed an improved algorithm using CUDA parallel computing to do the ORB feature extraction and matching portion to reduce the running time of the algorithm based on open source ORB - SLAM2 visual SLAM algorithm. After that.,testing the algorithm with public stereo datasets by doing some comparative experiments using the laptop and camera. the experiment result shows that the proposed method by this paper can greatly reduce the time of extracting feature points, calculating feature descriptors and image matching, and also improve the accuracy of localization. Therefore, the improved algorithm can improve the operation efficiency of the algorithm and the robustness of mobile robot SLAM.

[1]  G LoweDavid,et al.  Distinctive Image Features from Scale-Invariant Keypoints , 2004 .

[2]  Satoshi Fujimoto,et al.  ORB-SLAM map initialization improvement using depth , 2016, 2016 IEEE International Conference on Image Processing (ICIP).

[3]  Andrew J. Davison,et al.  DTAM: Dense tracking and mapping in real-time , 2011, 2011 International Conference on Computer Vision.

[4]  Davide Scaramuzza,et al.  SVO: Fast semi-direct monocular visual odometry , 2014, 2014 IEEE International Conference on Robotics and Automation (ICRA).

[5]  Mohamed S. Shehata,et al.  Image Matching Using SIFT, SURF, BRIEF and ORB: Performance Comparison for Distorted Images , 2017, ArXiv.

[6]  Gary R. Bradski,et al.  ORB: An efficient alternative to SIFT or SURF , 2011, 2011 International Conference on Computer Vision.

[7]  Daniel Cremers,et al.  Direct Sparse Odometry , 2016, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[8]  G. Klein,et al.  Parallel Tracking and Mapping for Small AR Workspaces , 2007, 2007 6th IEEE and ACM International Symposium on Mixed and Augmented Reality.

[9]  Daniel Cremers,et al.  LSD-SLAM: Large-Scale Direct Monocular SLAM , 2014, ECCV.

[10]  Li Qun.  LOCALIZATION APPROACHES FOR INDOOR AUTONOMOUS MOBILE ROBOTS: A REVIEW , 2003 .

[11]  J. M. M. Montiel,et al.  ORB-SLAM: A Versatile and Accurate Monocular SLAM System , 2015, IEEE Transactions on Robotics.

[12]  Gerhard P. Hancke,et al.  Pose Estimation of a Mobile Robot Based on Fusion of IMU Data and Vision Data Using an Extended Kalman Filter , 2017, Sensors.

[13]  Changku Sun,et al.  Vision and dual IMU integrated attitude measurement system , 2018, International Conference on Optical Instruments and Technology.

[14]  Jian Fu,et al.  A Method of SIFT Simplifying and Matching Algorithm Improvement , 2016, 2016 International Conference on Industrial Informatics - Computing Technology, Intelligent Technology, Industrial Information Integration (ICIICII).

[15]  Gustavo Olague,et al.  CUDA-based parallelization of a bio-inspired model for fast object classification , 2017, Neural Computing and Applications.

[16]  Olivier Stasse,et al.  MonoSLAM: Real-Time Single Camera SLAM , 2007, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[17]  Xue Wu,et al.  Improved algorithm based on SURF for image registration , 2017 .

[18]  Juan D. Tardós,et al.  ORB-SLAM2: An Open-Source SLAM System for Monocular, Stereo, and RGB-D Cameras , 2016, IEEE Transactions on Robotics.

[19]  Kevin Skadron,et al.  Scalable parallel programming , 2008, 2008 IEEE Hot Chips 20 Symposium (HCS).

[20]  Alexandre Hostettler,et al.  ORBSLAM-Based Endoscope Tracking and 3D Reconstruction , 2016, CARE@MICCAI.