Hi,
I consider a difference in output when I want to export the poses. I use the following options for the GUI (database viewer):
RGBD-SLAM ID format -> Map's graph -> robot (base frame)
And for the script I use rtabmap-export --poses --poses_format 11 $DB
It seems that the script version does some optimization:
[ INFO] (2024-11-23 10:12:38.628) Rtabmap.cpp:5131::optimizeCurrentMap() Optimize map: around location 1611 (lookInDatabase=true)
[ INFO] (2024-11-23 10:12:38.629) Rtabmap.cpp:5140::optimizeCurrentMap() get 1579 ids time 0.000964 s
[ INFO] (2024-11-23 10:12:38.632) Rtabmap.cpp:5183::optimizeGraph() get constraints (ids=1579, 1583 poses, 3253 edges) time 0.002875 s
[ INFO] (2024-11-23 10:12:38.701) Rtabmap.cpp:5255::optimizeGraph() Optimization time 0.068538 s
[ INFO] (2024-11-23 10:12:38.701) Rtabmap.cpp:5143::optimizeCurrentMap() optimize time 0.072313 s
[ INFO] (2024-11-23 10:12:38.702) Rtabmap.cpp:5152::optimizeCurrentMap() Correction (from node 1611) xyz=-0.011847,-0.096141,-0.002846 rpy=0.000646,-0.000955,0.029272
Optimizing the map... done (0.782533s, poses=1583).
Create and assemble the clouds...
Create and assemble the clouds... done (4.187351s, 3020073 points).
Poses exported to
The outputs are different (global rotation) and tiny pose differences. In addition the GUI version export the landmark poses as well. I use the latest osx app (GUI) and latest source code (script).
Can you explain the difference? Is it deliberately?
Hi,
I consider a difference in output when I want to export the poses. I use the following options for the GUI (database viewer):
RGBD-SLAM ID format -> Map's graph -> robot (base frame)
And for the script I use
rtabmap-export --poses --poses_format 11 $DBIt seems that the script version does some optimization:
The outputs are different (global rotation) and tiny pose differences. In addition the GUI version export the landmark poses as well. I use the latest osx app (GUI) and latest source code (script).
Can you explain the difference? Is it deliberately?