Planar Homography based Monocular SLAM Initialization Method

Simultaneous Localization and Mapping (SLAM) is a popular topic in autonomous robots navigation. It has been studied for decades in both computer vision and robotics communities. Monocular systems is more cost effective compared to RGBD or Stereo systems; however, it is relatively complicated to initialize due to scale uncertainty. Under certain conditions, it is assumed that the camera only moves in a planar scene, which provides us with homography constraints. In this paper, the efficiency of monocular initialization was improved based on the open source platform ORB-SLAM2, employing the algorithm based on planar homography constraints. We compared the improved algorithm with the source code of ORB-SLAM2 on the public datasets. It showed that our algorithm has better stability and robustness in the planar scene dataset and more initializing map points.

[1]  Shaojie Shen,et al.  Monocular Visual–Inertial State Estimation With Online Initialization and Camera–IMU Extrinsic Calibration , 2017, IEEE Transactions on Automation Science and Engineering.

[2]  S. Jia,et al.  Mobile Robot Simultaneous Localization and Mapping Based on a Monocular Camera , 2016, J. Robotics.

[3]  Vincent Lepetit,et al.  Instant Outdoor Localization and SLAM Initialization from 2.5D Maps , 2015, IEEE Transactions on Visualization and Computer Graphics.

[4]  Rob Fergus,et al.  Depth Map Prediction from a Single Image using a Multi-Scale Deep Network , 2014, NIPS.

[5]  Bernhard P. Wrobel,et al.  Multiple View Geometry in Computer Vision , 2001 .

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

[7]  Sheng Liu,et al.  Fast initialization for feature-based monocular slam , 2017, 2017 IEEE International Conference on Image Processing (ICIP).

[8]  Michael Gassner,et al.  SVO: Semidirect Visual Odometry for Monocular and Multicamera Systems , 2017, IEEE Transactions on Robotics.

[9]  Anders Robertsson,et al.  Visual Odometry from Two Point Correspondences and Initial Automatic Camera Tilt Calibration , 2017, VISIGRAPP.

[10]  Anders Heyden,et al.  Planar Motion and Hand-eye Calibration using Inter-image Homographies from a Planar Scene , 2013, VISAPP.

[11]  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.

[12]  Chunhua Shen,et al.  Depth and surface normal estimation from monocular images using regression on deep features and hierarchical CRFs , 2015, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[13]  Friedrich Fraundorfer,et al.  Minimal solutions for the rotational alignment of IMU-camera systems using homography constraints , 2018, Comput. Vis. Image Underst..

[14]  Rob Fergus,et al.  Predicting Depth, Surface Normals and Semantic Labels with a Common Multi-scale Convolutional Architecture , 2014, 2015 IEEE International Conference on Computer Vision (ICCV).

[15]  Qifeng Yu,et al.  Planar self-calibration for stereo cameras with radial distortion. , 2017, Applied optics.

[16]  Pascal Vasseur,et al.  Visual Odometry Using a Homography Formulation with Decoupled Rotation and Translation Estimation Using Minimal Solutions , 2018, 2018 IEEE International Conference on Robotics and Automation (ICRA).

[17]  David Nistér,et al.  An efficient solution to the five-point relative pose problem , 2003, 2003 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2003. Proceedings..