Yaw PItch and Roll Vs Hugin

[jose.mar...@gmail.com]

Greetings, I have done some tests with HUGIN trying to get mosaics, but I have some doubts about the Yaw/pitch/ roll parameters, the end result (mosaic) is the same either with the drone's  Y/P/R that when I put values of Y/P/R = 0. 

Then I ask:

How does hugin use the drone's YPR values ?

How does hugin get YPR values after alignment?

How can I manually calculate the values of YPR that gets hugin?

Thanks for your answers.

JMartinez

[Sean Greenslade]

If I recall correctly, the YPR values are
used only for regular nodal panos.
Mosaic uses a different set of parameters.

--Sean

[jose.mar...@gmail.com]

thanks; If that were the case, how does hugin calculate them?

JMartinez

[Sean Greenslade]

On Mon, Jan 23, 2017 at 06:19:13PM -0800, jose.mar...@gmail.com wrote:
<snip>

> > >How does hugin use the drone's YPR values ?
> > >How does hugin get YPR values after alignment?
> > >
> > >How can I manually calculate the values of YPR that gets hugin?
> > >
> > >Thanks for your answers.
> > >JMartinez
> >
> > If I recall correctly, the YPR values are
> > used only for regular nodal panos.
> > Mosaic uses a different set of parameters.
>

> thanks; If that were the case, how does hugin calculate them?

This is the best explanation I can find for how the mosaic projections
work:

http://wiki.panotools.org/Stitching_a_photo-mosaic

I was mistaken when I said y/p/r is not used for mosaic. Yaw and pitch
are generally not used for flatbed scan mosaics, but they would
definitely be used for stitching, say, aerial photographs.

I would love it if someone more familiar with the mathematics could give
an explanation of what Tpy and Tpr actually mean in the context of the
geometric image remapping. Are they intended to be used with mosaics?
The fast preview doesn't seem to think so.

As for how Hugin actually calculates the values of all the variables it
is optimizing, that is more of an abstract mathematical problem. You
have a many-variable, many-equation problem that is (in most projects)
overconstrained. The problem uses non-linear functions, so a
least-sqaures fitting process is used. If you want a general concept
overview, check this article:

https://en.wikipedia.org/wiki/Non-linear_least_squares

--Sean

[jose.mar...@gmail.com]

Thank you very much for your time and responses

J.