Sparse MapCloud

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

Sparse MapCloud

yousaf.farrukh
Hey there!
I'm new to ROS and found RTAB-Map useful for my project of quadcopter automation. The issue I am have in using RTAB-Map is; when I run the standalone application I get very precise map with good odometry. But, when I run RTAB-Map in rviz through ROS, I get a very sparse MapCloud. The point cloud disappears as the kinect is moved around, leaving a few points on the previously scanned frame.
Thanks in advance
Reply | Threaded
Open this post in threaded view
|

Re: Sparse MapCloud

matlabbe
Administrator
Hi,

Can you list the topics you are referring? Are there any errors/warnings in the terminal?

cheers,
Mathieu
Reply | Threaded
Open this post in threaded view
|

Re: Sparse MapCloud

yousaf.farrukh
Hi,
Sorry I was away so could not reply you.

No there are no errors/warnings in the terminal. However I'm getting the odometry quality of around 250 - 350

The topics I'm subscribing are as follows

PointCloud2 -> /voxel_cloud
MapData     -> /rtabmap/mapData

Thanks in advance
Reply | Threaded
Open this post in threaded view
|

Re: Sparse MapCloud

matlabbe
Administrator
Hi,

The voxel_cloud is just the current cloud of the Kinect with a lot of points. MapData contains the data from nodes in the map. In RVIZ, mapData's point clouds are decimated by 4 to save some memory and rendering time. You can change the MapCloud plugin parameters in RVIZ to decrease decimation to get more dense clouds.

cheers,
Mathieu