Breaking maps apart in tiny, more meaningful, bits

tl;dr: a method to segment maps in semantic regions such as rooms and corridors

Maps can be daunting. They can be big, complex, and it’s hard to talk about maps without saying the words rooms, corridors, kitchen… Effectively breaking down the complexity.

So being able to do exactly that would be very interesting for robots. And while we are at it, let’s not do that ony for robot built maps! We should be able to do it for any type of map we would like to use. So let’s take maps that someone drew, i.e. sketch maps.

 

Here is a sketch map where each region has a different color. Attention, the same color, but in different color regions, doesn’t mean they are the same type of region

And the same with a robot build map

I want to use this

The paper can be found on Arxiv and the code is available on Github under GNU licence.

If you use this work, please cite this paper and hopefully, it will be accepted to ICRA and I’ll update the bib ref:

@ARTICLE{2017arXiv170909899M,
 author = {{Mielle}, M. and {Magnusson}, M. and {Lilienthal}, A.~J.},
 title = "{A method to segment maps from different modalities using free space layout - MAORIS : MAp Of RIpples Segmentation}",
 journal = {ArXiv e-prints},
archivePrefix = "arXiv",
 eprint = {1709.09899},
 primaryClass = "cs.RO",
 keywords = {Computer Science - Robotics},
 year = 2017,
 month = sep,
 adsurl = {http://adsabs.harvard.edu/abs/2017arXiv170909899M},
 adsnote = {Provided by the SAO/NASA Astrophysics Data System}
}

 

Advertisement