Inverse Polynomial Required in SVO_PRO

[Subodh Mishra]

Hello Dr. Scaramuzza,

I am using svo_pro in my research and I wanted to test the difference in algorithm performance when using fisheye (pinhole-equidistant) camera model vs the omnidirectional camera model, that this toolbox proposes. 

I looked into the code base of svo_pro to guess what the omnidirectional camera model expects. 

Referring the model described here: https://sites.google.com/site/scarabotix/ocamcalib-omnidirectional-camera-calibration-toolbox-for-matlab and the relevant papers if I am correct 

*for principal point we need to enter [xc', cy'], 

*for distortion we need to provide: [c, d, e], and 

*for polynomial, we need to provide: [a_0, a_1, a_2, a_3, a_4].

What I do not understand is the inverse polynomial parameter that the camera model expects. Where and how can I obtain this?

I will be glad if you can provide this information to me.

Thanks,

Subodh.

[Subodh Mishra]

Now, I understand that it is available in the txt file generated by the toolbox. However, I want to understand how is it calculated? So that I can do it myself for the result obtained from the fisheye calibration method described here: https://www.mathworks.com/help/vision/ug/fisheye-calibration-basics.html 

Which uses the same model as this toolbox, only with a little change in conventions.

Subodh.

[Subodh Mishra]

Okay. I figured this out. One needs to use the file/function: findinvpoly.m .

Subodh.