RTABMAP Standalone Accessing GraphViewer.h file

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

RTABMAP Standalone Accessing GraphViewer.h file

simonfrfr
Hello,
I have been working with the RTABMAP Standalone, because I do not want to deal with ROS. I have everything running the way that I would like it to run right now except for one thing. This is accessing the Graph Viewer. I wrote my own method for accessing the graph data, but then I realized that RTABMAP can generate its own map data. My implementation relies heavily on PCL, and is slow when it comes to large Point clouds. I am also using my own implementation of Concurrent Dijkstra's Algorithm (CDA). I saw that RTABMAP also has its own A* implementation, so I was quite happy, even though CDA is faster. I was just interested to see if there was any support for outsiders of the ROS community. And I also wanted to compare execution times of the two methods.
Thanks,
Simon
Reply | Threaded
Open this post in threaded view
|

Re: RTABMAP Standalone Accessing GraphViewer.h file

matlabbe
Administrator
Hello Simon,

RTAB-Map uses A* only for planning purpose, it doesn't have any effect on the SLAM part. What do exactly you want to compare? Time for planning?

The GraphViewer is a QGraphicsView showing the RTAB-Map's graph. For convenience for those using only the standalone library, I've just moved all common headers to the public headers directory. GraphViewer is now accessible from outside the library as well as other useful widgets.

The graph (and cloud data) is sent over rtabmap::Statistics events (I highlighted in the link the getters of the graph). An example how parsing the event for building a 3D map can be found in processStatistics() method of the RGBD mapping example. The CloudViewer widget (wrapper of the PCL visualizer) is used for point cloud visualization.

cheers
Reply | Threaded
Open this post in threaded view
|

Re: RTABMAP Standalone Accessing GraphViewer.h file

simonfrfr
Thank-you for your Speedy reply, I just started school on the 24th, and just realized the reply. Thank-you I think you have answered my question, and I do want to compare time for planning if there is really too much of a statistical difference. (Your A* might be more optimized than my concurrent dijkstra algorithm)
Thanks,
Simon
(If I have any more questions I will ask :) Thanks again)