ICE-BA integration with vilib

[Ilyass Taouil]

Dear @baliika @foehnx

I recently read your publication, thanks for making the VIO framework available.

I am trying to re-create the results you got by combining the VIO with the ICE-BA backend bundle adjustment, but I am having some doubts on how you combined your framework and ICE-BA.

My current understanding is that there is a three-step approach to it:

1. Run the detection and tracking using vilib (and somehow store these data)
2. Create dat files for each frame using the images as well as the stored feature points.
3. Finally run the back-end bash script using the .dat files generated from 2.

If so, how did you manage to create the .dat files without using the ice-ba executable?

Is this correct approach? If not could you point me in the correct direction?

Best,
Ilyass