How to compare rgbdslam and rtab map is any software available to test?

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

How to compare rgbdslam and rtab map is any software available to test?

dhagash
I have a task to build a 3D map for autonomous navigation I want to see which one of the two is better to perform the specific task and I cannot tell that this one is better without any results. Is there any SLAM evaluating software or code available if yes how to evaluate the two and if no then what are the points to consider to arrive at conclusion.
Reply | Threaded
Open this post in threaded view
|

Re: How to compare rgbdslam and rtab map is any software available to test?

matlabbe
Administrator
Hi,

You may use RGB-D SLAM Dataset and Benchmark to compare. To process these datasets in rtabmap, you can use rtabmap-rgbd_dataset tool (available when building rtabmap from source).

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

Re: How to compare rgbdslam and rtab map is any software available to test?

dhagash
Thanks for the prompt reply can you explain me in more detail how can do the evaluating from the tool provided by computer vision group I am not able to follow you exactly. I want to compare in numbers is it possible or not? I am not able to see the rgbddataset in here how to run it I am noob please if you can explain..
Reply | Threaded
Open this post in threaded view
|

Re: How to compare rgbdslam and rtab map is any software available to test?

dhagash
In reply to this post by matlabbe
I have downloaded the bag file from the computer vision group at TUM and have done it for rgbdslam now can you explain similarly what can be done for RTAB map
Reply | Threaded
Open this post in threaded view
|

Re: How to compare rgbdslam and rtab map is any software available to test?

matlabbe
Administrator
Hi,
Use this tool by giving path file for the selected dataset: https://vision.in.tum.de/data/datasets/rgbd-dataset/online_evaluation

Use absolute error, the lower the rmse is, the better the approach.

Cheers,
Mathieu