Hello Nils,
> I am quite new to octomap and I a have a few doubts.
>
> I am using a monocular SLAM algorithm in ROS, which outputs the map as
> a sparse pointcloud, about 400 points at start and I would like to
> build an occupancy grid map using this point cloud as the input cloud
> to octomap.
>
> Once I remapped the pointcloud topic in the octomap launch file, I
> launched it in ROS and I got the following warning: Nothing to
> publish, octree is empty.
octomap_server prints this warning when the octee is empty, that means
there were no sensor readings integrated. Does it constantly appear or
only at the beginning? Are there any other error or debug messages for
octomap_server? (check rqt_console, you might have to change the log level)
Another thing to check are the frame IDs of sensor messages and world
frame, and if there are tf transforms published between them.
Some parameters such as maximum range and pointcloud_[min|max]_z could
also result in an empty map.
>
> I have seen that octomap has been used lot with denser pointclouds
> such as the ones created from laser scans and RGBD cameras. May I be
> getting this error because the pointcloud from the monocular SLAM
> algorithm is too sparse? Or would I have to change some of the
> parameters in order for it to work?
Sparse point clouds should not be a problem in general, but certain
options like ground filtering should be disabled.
Best regards,
Armin