Revision history [back]

Sounds like you have some sort of SLAM package operational. However this is not enough to detect obstacles. Yes, you will get lidar points where there are obstacles, but you will also end up with noise or false positives. You need to process this further to detect obstacles. One of the most common and easy ways to do this is to cluster the points to filter out noise or points that aren't very dense. Then you will have to track these objects. If all of this sounds too overwhelming, there are several packages that perform this task. The very simple lidar_tracking package should help you with this. It returns a list of detected and tracked objects as markers where you should be able to pull out the positions and poses of the objects.