search

PRBonn / ir-mcl

IR-MCL: Implicit Representation-Based Online Global Localization https://arxiv.org/abs/2210.03113
122 stars 11 forks source link

About the [occumap.npy] file in the data set #3

Open fffzzzjjj opened 1 year ago

fffzzzjjj commented 1 year ago

Regarding the files [occumap.npy] contained in the various datasets in the project, how were they obtained? Was the environment mapped by gmapping or cartographers algorithm and then the images were converted to .npym format? I hope everyone can help me explain this mystery.

KuangHaofei commented 1 year ago

Hi,

Thanks for your question! Of course, you can also use gmapping or cartographer to get the map. But since our dataset already has ground-truth poses, we directly run the occupancy grid mapping algorithm to build the map.

You can refer to our SLAM course for more details about the algorithm: MSR1 coursres: https://www.ipb.uni-bonn.de/photo12-2021/ Occupancy Grid Mapping: https://www.youtube.com/watch?v=v-Rm9TUG9LA

I will also integrate the grid mapping code into tools in the future.

Thanks!