/glores

Globally Optimal Registration based on Fast Branch and Bound

Primary LanguageC++GNU General Public License v3.0GPL-3.0

glores - Globally Optimal Registration based on Fast Branch and Bound

Copyright (C) 2019 Luca Consolini, Mattia Laurini, Marco Locatelli, Dario Lodi Rizzini.

OVERVIEW

Library glores implements the globally optimal registraion method using two lower bound estimations. It has been kept to a minimal design.

If you use this library, please cite the following paper:

L. Consolini, M. Laurini, M. Locatelli, D. Lodi Rizzini. A Second-Order Lower Bound for Globally Optimal 2D Registration. under review, arXiv arXiv:1901.09641.

@misc{consolini2019glores,
  author={Consolini, L. and Laurini, M. and Locatelli, M. and Lodi Rizzini, D.},
  title = {A Second-Order Lower Bound for Globally Optimal 2D Registration},
  Year = {2019},
  Eprint = {arXiv:1901.09641},
}

DEPENDENCIES

The software depends on the following external libraries

  • Boost
  • Eigen 3.0

If Matlab (we tested version R2018a=9.4) is available, it also compiles the glores_mex extension. Other dependencies are placed in directories named thirdparty/.

QUICK START

After compiling glores library and the related applications, you can test the algorithm calling:

${glores_ROOT}/build/testPointSetGlobalRegistration -cfg ../data/glores_scan_example_0180.txt

Similarly,

${glores_ROOT}/build/testPointSetGlobalRegistration -cfg ../data/glores_scan_example_1440.txt

Configuration files glores_scan_example_0180.txt and glores_scan_example_1440.txt sets the parameters of the algorthm.