Open MalcolmMielle opened 7 years ago
My apologies, apparently I made a mess with the versions, please use the branch "cleaning_files" while I fix it. You will need another ROS node to choose the map you want to segment, i used this (up means next and down segment)
Thanks a lot for the quick feedback !
I get corrupted double-linked list
errors when runing NBL and lab_intel and a seg fault still on lab_ipa. I don't know if this has to do with the cleaning version and will be solved when the master get cleaned but I thought I would let you know :).
The corrupted double-linked list error comes from the library of Dual Space Decomposition, I didn't modify their code. I've tried the rest of the maps myself with no problem, maybe you should increase a little the threshold.
By the way, many of these problems occurs when segmenting static maps, the incremental version is more reliable for time varying maps.
OK I'll try that then :)
Hello,
I'm running this package under Ubuntu 14.04 using ros indigo.
I'm running the evaluation test using a concavity measure of 2.5 as in the paper using this command
rosrun inc_dude evaluation 2.5
. I experience segmentation faults when running the algorithm on NLB:As well as Freiburg101_scan although works with the default parameters:
Freigbrug52 worked though with 2.5m but the result looks weird
It seg fault with default:
So I guess I'm doing something wrong while using this. What do I need to do to run this exact same tests as in the paper published at ICRA2017 ?
Thanks a lot,