search

pjrambo / VINS-Fusion-gpu

This repository is a version of VINS-Fusion with gpu acceleration. It can run on the Nvidia TX2 in realtime
GNU General Public License v3.0
501 stars 134 forks source link

KITTI GPS Fusion (Stereo + GPS) results in skew paths and one is faster than the other #5

Closed mmp52 closed 5 years ago

mmp52 commented 5 years ago

When I try to run the kitti gps fusion test datase t(which I have downloaded from your link) in my jetson tx2 , system results in a funny skew paths - gps based global path and VIO path are in different orientations. Also gps path is updated slower than the other. What can be the reason? screenshot_different_paths_kitti

And one final question, is there a script to calculate rmse between two paths , I also want to try that in euroc results for different configuration parameters. Thanks for all of this, Metin

pjrambo commented 5 years ago
  1. Different orientation is normal. Since VIO is in a local frame that is relative to the frame when the initialization procedure finishes. GPS based path is in the global frame (ENU). So in most of the case, their orientations are different.
  2. GPS path is generated by a global optimization method, which may cost more computing resources. The computing resources cost is rising with time goes by. I've never tried this part in TX2 platform, maybe TX2's computing resource is not enough.
  3. For calculating RMSE, you may find some code on Github. You need to align the two results and then calculate it.
mmp52 commented 5 years ago

Thank you for your response. How can I align them and where can I find RMSE calculating scripts on github? Also is there a ground truth in kitti vision series, when I try to calculate RMSE what should I use as my reference ?

Metin