Closed astef3 closed 2 months ago
To be more specific, because I might have used a little confusing vocabulary, by drift in the previous question I meant the transformation between frames. :)
Closed it due to finding explanation in: https://github.com/MichaelGrupp/evo/issues/421
Hi,
I have a question related to the output of the RPE metric. I have ran the evaluation that and I got the result presented in the figure below:
Now as far as I understand (please correct me if I am wrong, I am quite new in the field of evaluation) from the explanation of the metric available at https://github.com/MichaelGrupp/evo/blob/master/notebooks/metrics.py_API_Documentation.ipynb , the drift in interval delta for the reference trajectory is calculated, then the drift in the estiamted trajectory is calculated and then the difference between the drifts is calculated. So, in the output I provided above, the first half of the trajectory is understandable for me as I saw the correspondance markers and it seams okay, however in the second half I have a trouble with understanding the calculated metric, as the RPE seems quite high, however when I check the corresponding poses in some interval delta (zoomed image bellow) it seems that the drift is not that huge.
The command I am using for metric estimation is:
evo_rpe bag poses_rtabmap_minus3.bag /tf_to_pose /rtabmap/odom_custom_timestamp -va --delta 1 --delta_unit m -p --plot_mode xy
I would be really grateful, if you can provide me with some information about whether I have some mistakes in understanding and interpreting the results. In case you need more data from me about this issue I would be happy to provide it. Thank you in advance for your help and time.