auto estimation of control points - enhancement request

[paul womack]

When creating control points manually, I note that auto-estimate
is not very clever (albeit better than nothing).

From the documentation:

> auto estimate Tries to estimate the position of the second point by
> estimating the translation between the two images. This is very
> crude and probably only works for single row panoramas
> created from Rectilinear Projection images.

Given that hugin has a handy panotools library available,
might I suggest that a quick 'n' dirty YPR model be applied
to the two windows?

For the first CP, only YP would be optimised, but after that YPR is possible.

I *think* that YPR is general enough (and the relationship between
adjacent images of a pano restricted enough) that this would give very good results.

BugBear

[T. Modes]

Am Freitag, 29. Mai 2015 22:46:00 UTC+2 schrieb bugbear:

Given that hugin has a handy panotools library available,
might I suggest that a quick 'n' dirty YPR model be applied
to the two windows?

From the release notes of 2015.0
* The fine-tune and auto-estimate functions in the control point tab have been made projection aware.
 

For the first CP, only YP would be optimised, but after that YPR is possible.

That's exactly what is implemented in 2015.0

[Frederic Da Vitoria]

FWIW, in the few panos I assembled with 2015.0, I did have a feeling auto-estimate was indeed working better, especially when one (or both) of the images are blurred.

--

Frederic Da Vitoria
(davitof)

Membre de l'April - « promouvoir et défendre le logiciel libre » - http://www.april.org

[Paul Womack]

Is this YPR the "global" optimised one, or a custom YPR for the image

pair whose CPs are being edited/created?

 BugBear

[T. Modes]

Am Sonntag, 31. Mai 2015 10:24:41 UTC+2 schrieb bugbear:

Is this YPR the "global" optimised one, or a custom YPR for the image

pair whose CPs are being edited/created?

The optimisation is done only for image pair which is currently edited. The result of the optimisation is only used internally to determine the position of the new cp. The positions of the images in the pano remains unchanged.
For more detail read the source code.

[Paul Womack]

Is this YPR the "global" optimised one, or a custom YPR for the image

pair whose CPs are being edited/created?

The optimisation is done only for image pair which is currently edited. The result of the optimisation is only used internally to determine the position of the new cp. The positions of the images in the pano remains unchanged.

 

Wonderful! Thank you.

 BugBear