Thanks for the great work and detailed user guidance!
The multi-session mode indeed has great practical values and solves the long term mapping issues.
However, I'm wondering does LTAOM supports the manual input of initial pose (i.e., transformations) when running the multi-session mode? I found the function addInitialPoseValues and the variable pose_prior, and guess they maybe related to the initial pose.
In some challenging degenerated scenarios, LTAOM may fail to merge the results from two mapping sessions. So the manual input of initial pose maybe of help in such scenarios.
Hi @zuhaozou2017,
Thanks for the great work and detailed user guidance! The multi-session mode indeed has great practical values and solves the long term mapping issues.
However, I'm wondering does LTAOM supports the manual input of initial pose (i.e., transformations) when running the multi-session mode? I found the function
addInitialPoseValuesand the variablepose_prior, and guess they maybe related to the initial pose.In some challenging degenerated scenarios, LTAOM may fail to merge the results from two mapping sessions. So the manual input of initial pose maybe of help in such scenarios.