I was reading the following paper on how you can incorporate variable time offset for a camera in VIO:


As far as I understand the paper is optimizing camera-image time offset and not IMU time offset.

What I am confused on is how the jacobian of the image residual is computed with respect to time.

I'm clearly missing something as normally the image jacobian should be time-independent(IE SFM does not consider time as a variable as there isn't a motion model associated with it).

So what's going on? I guess I don't understand what they are doing.



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.