search

hku-mars / HBA

[RAL 2023] A globally consistent LiDAR map optimization module
GNU General Public License v2.0
469 stars 66 forks source link

I think the code and the contents of the paper are different. #5

Closed jaeseok4104 closed 1 year ago

jaeseok4104 commented 1 year ago

The code has a constraint between all the nodes in the window, and didn't make the constraint of the floor 2 layer.

Can you explain this?

image

image

samsdolphin commented 1 year ago

Hi @jaeseok4104, yes you are right. In short, we found this implementation has a faster convergence rate and a better precision. We thus open source this version of code instead of the very original one (paper version). However, you could modify the code structure by yourself. Hope this helps.

jaeseok4104 commented 1 year ago

thanks :) Have you plan to open about paper version code?

samsdolphin commented 1 year ago

Haha, this shouldn't be hard for you ;)

HaisenbergPeng commented 1 year ago

Hi, after the sliding window ba in the 1st layer， the optimized poses are not used for factor graph, instead, you used the initial poses and the ba hessians. Could you please explain the reason? thanks a lot! image

image

samsdolphin commented 1 year ago

Hi @HaisenbergPeng, in each local ba, we want the hessians to constrain the relative poses between the 1st layer poses. The optimized poses in each local ba are used to construct the next layer's keyframe.

HaisenbergPeng commented 1 year ago

okay, got it! Maybe I could try using the optimized poses of the bottom layer as factors.