0
$\begingroup$

I was looking for an indoor SLAM algorithm to run on a robot and came across SVO 2.0 which looks very impressive: https://github.com/uzh-rpg/rpg_svo_pro_open

But from a brief look it doesn't seem to have a way to save or load the map or creates - does this mean the map needs to be created new every time even in the same environment? Thanks

$\endgroup$
1
  • $\begingroup$ In future questions, please make it clear which version of ros you are using e.g. ros-noetic. $\endgroup$
    – Mike973
    Sep 6, 2023 at 11:22

2 Answers 2

1
$\begingroup$

It is common for SLAM nodes to not provide this feature, because it is already available from the map_server ros package. map_server can subscribe to the map topic and write the necessary files to disk. For ros1, the command looks like:

rosrun map_server map_saver -f my_map

On success, this will create my_map.pgm and my_map.yaml in the current directory. There are more options described on the wiki page.

$\endgroup$
1
  • $\begingroup$ Thank you for the response. I will install the library and attempt this. $\endgroup$
    – JassD
    Sep 8, 2023 at 6:17
0
$\begingroup$

I also see no way to save or load a map into it.

If you want to export the pointcloud as a file to save and do something with it later, you could get the data with a python node and a subscriber. Just subscribe to the topic that is used in RVIZ to show the points, I assume.

That way you could use open3D in python to create a pointcloud object or other mesh file.

$\endgroup$
1
  • $\begingroup$ Thanks for providing an answer! But, per the question, it appears that this package is not relevant to pointcloud or mesh outputs, it is performing SLAM and odometry. The output of this process would be a 2-dimensional top-down occupancy grid. $\endgroup$
    – cst0
    Sep 6, 2023 at 15:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.