Releases: uzh-rpg/rpg_trajectory_evaluation
Releases · uzh-rpg/rpg_trajectory_evaluation
Improve support for analyzing multiple trajectories and multiple runs
This version mostly improves the support for analyzing multiple algorithms on multiple datasets:
- now the algorithms and datasets are specified with a
yaml
configuration file, should be easier than the old way (#18) - sub-trajectory lengths can be specified in the
yaml
file, using either a fixed set of lengths or the percentages of the trajectory.
Support for analyzing the results from multiple runs is also added, for both analyzing single and multiple trajectories.
Other notable improvements changes are:
- differentiate different estimation types (real-time poses, pose graph and bundle adjusted poses) to have different file names
- allow to specify the start and end time of the estimate to consider
- support using the algorithm/dataset order in the configuration file or sort them alphabetically (#15)
- diverse changes of the plotting/saving options in analyzing multiple trajectories