A method for nonrigid optimization of 2D map alignments and correctiion of sensor map deformity.

The following article presents the method in detail.

• Saeed Gholami Shahbandi, Martin Magnusson, Karl Iagnemma. Nonlinear Optimization of Multimodal Two-Dimensional Map Alignment With Application to Prior Knowledge Transfer, in IEEE Robotics and Automation Letters, vol. 3, no. 3, pp. 2040-2047, July 2018. doi: 10.1109/LRA.2018.2806439. URL

Download, install the package and its dependencies. Most dependencies are listed in requirements.txt, except for OpenCV which needs to be installed with its python wrapper.

# Download
git clone https://github.com/saeedghsh/Map-Alignment-Nonrigid-Optimization-2D.git
cd Map-Alignment-Nonrigid-Optimization-2D

# Install dependencies
pip install -r requirements.txt

# Install the package [optional]
python setup.py install

NOTE: This repository is dedicated to the implementation of the nonrigid optimization of the alignment. This work relies on the 2D Map Alignment With Region Decomposition for the initial guess of the alignment. That package and its dependencies have to be installed too.

Run this:

python demo.py --img_src 'map_sample/F5_04.png' --img_dst 'map_sample/F5_layout.png' --hyp_sel_metric 'fitness' -visualize -save_to_file -multiprocessing

And you should see this:

A collection of layout and sensor maps from four different environments are available in this repository.

• Clean up optimize_alignment.optimize_alignment.py. For instance, put all the methods for region segmentation and region segmentation transfer into a separate module. And complete the documentation of all methods
• Add more demos, like for region segmentation and knowledge transfer.
• Include the 3D plots? hmmm, not sure.

Distributed with a GNU GENERAL PUBLIC LICENSE; see LICENSE.

Copyright (C) Saeed Gholami Shahbandi