search

uzh-rpg / rpg_trajectory_evaluation

Toolbox for quantitative trajectory evaluation of VO/VIO
MIT License
1.06k stars 348 forks source link

The aligned plot is not being created #42

Closed javrtg closed 2 years ago

javrtg commented 3 years ago

Thanks a lot for the library :)

After executing the example:

python2 scripts/analyze_trajectory_single.py results/euroc_mono_stereo/laptop/vio_mono/laptop_vio_mono_MH_05

I got the following plot:

image

After seing the output from the example I guess that the grountruth trajectory visually matches the aligned one. But I am not sure if the aligned trajectory has been created and taken into account to compute the error metrics. Is there a way to check this?

Thanks in advance!

neliudochka commented 1 month ago

I have the same issue. Did you manage to understand if it is correct result?

javrtg commented 1 month ago

Hi @neliudochka, I believe that the module uses the label 'Estimate' to refer to the aligned curve, not the (original) estimated curve. This can be seen e.g. in:

https://github.com/uzh-rpg/rpg_trajectory_evaluation/blob/8c8ceec55c5c5094a6494208cfc5f54afe0bbc4d/scripts/analyze_trajectory_single.py#L173-L177

Concretely, the label 'Estimate' is assigned to the curve plot_traj.p_es_aligned. This curve results from aligning the (original) estimated curve p_es with the ground-truth curve p_gt. Here’s how it’s done:

https://github.com/uzh-rpg/rpg_trajectory_evaluation/blob/8c8ceec55c5c5094a6494208cfc5f54afe0bbc4d/src/rpg_trajectory_evaluation/trajectory.py#L250-L258

So to sum up, yes that plot is a correct (expected) result. It just has to be taken into account that the curve labeled as "Estimate" actually refers to the "Aligned" one.

I hope this helps clarify things for you!