Direkt zum Inhalt springen
Computer Vision Group
Faculty of Informatics
Technical University of Munich

Technical University of Munich

Menu

Submission form for automatic evaluation of RGB-D SLAM results

Groundtruth trajectory
Estimated trajectory
Evaluation options Offset: seconds (add to stamps of estimated traj.)
Scale: (scale estimated traj. by this factor)
Evaluation mode absolute trajectory error (recommended for the evaluation of visual SLAM methods)
relative pose error for pose pairs with a distance of (recommended for the evaluation of visual odometry methods)
relative pose error for all pairs (downsampled to pairs, alternative method for the evaluation of visual SLAM methods)

Runs the evaluation script on your data and displays the result. No data will be permanently saved on our servers. Alternatively, you can also download the evaluation script and perform the evaluation offline. Additional information about the evaluation options and the file formats is available. We also provide an example trajectory for freiburg1/xyz by RGBD-SLAM as well as instructions how to reproduce these trajectories.

Rechte Seite

Informatik IX
Chair for Computer Vision & Artificial Intelligence

Boltzmannstrasse 3
85748 Garching

info@vision.in.tum.de