Revision history [back]

Hi! I guess the only problem for you is not to think of 3D mapping but to establish the tf for your platform in the first place. If you have a single RGB-D sensor mounted on a tripod like in my case, I would write a launch file establishing all related tf frames and also using to load in a blank map or using gmapping package to render a map to complete the tf tree as well as the necessary topic for octomap.

So, at this point, your tf tree should look like: "map --> odom --> base_footprint --> camera_link".

Then call in octomap_server in your launch file, make sure to remap the "cloud_in" with your 3D camera's depth topic. Also, set the parameter "latch" to false if you don't have any map to load in. Don't forget to set the parameter "frame_id" to /map, If you succeeded in doing all these procedures, then you can render octomap with Marker Arrays on Rviz.

This is just my case, if your are using other sensors, modify corresponding topics according to your needs. Good luck !

Hi! I guess the only problem for you is not to think of 3D mapping but to establish the tf for your platform in the first place. If you have a single RGB-D sensor mounted on a tripod like in my case, I would write a launch file establishing all related tf frames and also using map_server to load in a blank map or using gmapping package to render a map to complete the tf tree as well as the necessary topic for octomap.

So, at this point, your tf tree should look like: "map --> odom --> base_footprint --> camera_link".

Then call in octomap_server in your launch file, make sure to remap the "cloud_in" with your 3D camera's depth topic. Also, set the parameter "latch" to false if you don't have any map to load in. Don't forget to set the parameter "frame_id" to /map, If you succeeded in doing all these procedures, then you can render octomap with Marker Arrays on Rviz.

This is just my case, if your are using other sensors, modify corresponding topics according to your needs. Good luck !