Awesome

Line and Plane based Visual Odometry (LPVO)

This package provides a MATLAB implementation of ICRA 2018 paper: "Low-Drift Visual Odometry in Structured Environments by Decoupling Rotational and Translational Motion" for the purpose of research and study only. Note that this repository only includes simplified proposed visual odometry example codes to understand how the LPVO works in structured environments.

1. Goal

Our goal is to estimate 6-DoF camera motion with respect to the indoor structured environments. The LPVO exploits line and plane primitives jointly to recognize the spatial regularities of orthogonal structured environments. Lines from RGB images and surface normals from depth images are simultaneously used to perceive environmental regularities accurately and stably. LPVO can track drift-free rotational motion while at least a single plane and a pair of lines parallel to the Manhattan world (MW) axes are visible. Given the absolute camera orientation, we recover the optimal translational motion, which minimizes de-rotated reprojection error.

2. Prerequisites

This package is tested on the MATLAB R2019b on Windows 7 64-bit. This package depends on mexopencv library for keypoint processing, KLT tracking, and translation estimation. cv.* functions in this package cannot run without mexopencv install in the MATLAB environment. Please, build mexopencv in your OS first, and then run this package. Some of the functions such as estimateSurfaceNormalGradient_mex.mexw64 are compiled as MEX file to speed up the computation. You can use estimateSurfaceNormalGradient.m instead if you cannot compile MEX file.

3. Usage

• Download the ICL-NUIM dataset from https://www.doc.ic.ac.uk/~ahanda/VaFRIC/iclnuim.html, 'of kt3' is recommended.

• Or, Use the ICSLRGBDdataset/rgbd_dataset_302_09_square3/ included in this package.

• Define 'datasetPath' correctly in your directory at setupParams_ICSL_RGBD.m file.

• Run LPVO_core/main_script_ICSL_RGBD.m, which will give you the 3D motion estimation result. Enjoy! :)

4. Publications

The approach is described and used in the following publications:

• Linear RGB-D SLAM for Planar Environments (Pyojin Kim, Brian Coltin, and H. Jin Kim), ECCV 2018.

• Low-Drift Visual Odometry in Structured Environments by Decoupling Rotational and Translational Motion (Pyojin Kim, Brian Coltin, and H. Jin Kim), ICRA 2018.

You can find more related papers at http://pyojinkim.com/_pages/pub/index.html.

5. License

The package is licensed under the MIT License, see http://opensource.org/licenses/MIT.

if you use LPVO in an academic work, please cite:

@inproceedings{kim2018low,
  author = {Kim, Pyojin and Coltin, Brian and Kim, H Jin},
  title = {Low-Drift Visual Odometry in Structured Environments by Decoupling Rotational and Translational Motion},
  year = {2018},
  booktitle = {IEEE International Conference on Robotics and Automation (ICRA)},
 }