renatojmsdh / motion-from-normals-spherical-view

Project page for motion estimation from normal surface vectors of wide FOV images

Code of "An Efficient Rotation and Translation Decoupled Initialization from Large Field of View Depth Images"

Renato Martins, Eduardo Fernandez-Moral, Patrick Rives

Email: renatojmsdh AT gmail.com

INRIA Sophia Antipolis, 2018

DOC:

• Please see the documentation and dependencies of this module in the "doc/" folder, by opening the index.html in your brownser

Running instructions:

• Open Matlab and select your current working directory as the one with the "main.m" function, open this function and run it.
• The salience function uses a compiled mex file for selecting the most informative information for the translation estimation. The code was compiled in Ubuntu 16.04 and should run in a similar PC architecture. If not, please compile the file "mex_saliency_geometry.cpp".

Short description:

This is an example for computing the pose initialization from normals using wide FOV sensors as in the IROS 2017 Paper:

• An Efficient Rotation and Translation Decoupled Initialization from Large Field of View Depth Images R Martins, E Fernandez-Moral, P Rives IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS'17

If this code is useful for you, please cite our paper in your research as:

 @inproceedings{martins17,
  author    = {Renato Martins and
             Eduardo Fernandez{-}Moral and
             Patrick Rives},
  title     = {An efficient rotation and translation decoupled initialization from
             large field of view depth images},
  booktitle = {{IEEE/RSJ} International Conference on Intelligent Robots and
             Systems, {IROS}, Vancouver, Canada},    
  year      = {2017}
  }