Is the distance shown at the control point mm, cm, m ?

[jm p]

Hi, good morning.
Among your programs...

Average control toe distance: 1.178942
Minimum control point distance: 0.006924
Maximum control point distance: 12.701439

The unit of distance here is
I'm very curious to know what it means in the millimeter (mm), centimeter (cm), meter (m), or anything else.

And when I updated the version from 12.24 to 12.26, I can see some parts of the panorama that are not well created.
Is it just me?
I'm going to erase the 12.26 version
It is being used in 12.24 version.
I would appreciate it if you could reply.
Thank you.

[Erik Krause]

Am 06.07.2024 um 13:47 schrieb jm p:
> The unit of distance here is
> I'm very curious to know what it means in the millimeter (mm), centimeter
> (cm), meter (m), or anything else.

Read the screen: it says "measured in source image pixels".

--
Erik Krause
http://www.erik-krause.de

[PTGui Support]

Indeed, it's calculated by projecting the second control point back into
the first image and measuring the distance in pixels.

An average below 2 is considered very good, it means that there is no
significant parallax and there are no control points on moving objects
(apart from a few outliers perhaps).

Kind regards,

Joost Nieuwenhuijse
www.ptgui.com

[John Houghton]

On Sunday, July 7, 2024 at 9:15:31 AM UTC+1 PTGui Support wrote:

An average below 2 is considered very good, it means that there is no
significant parallax and there are no control points on moving objects
(apart from a few outliers perhaps).

2 would be a good optimization figure, provided that there is an adequate number of control points points that are widely spread over the areas of overlap.  You might be able to improve the optimization figures by simply deleting the points with the largest control point distances, but that might be cheating!  If those points were accurately positioned on matching features, then you may well have parallax or subject movement issues to investigate.  However, if there are no visible stitching errors or distortions visible, then it's probably a waste of time vainly trying to lower the optimization figures.

John