Control Points Won't Save

[cwa...@oyez.org]

Hi all,

I've had a reoccurring issue with control points not saving properly. 

Here's the gist of it... After control points are added and optimized, I'll save the project. All is well, and it reopens just fine. However, after I generate a panorama and then save that exact same project, upon opening it again the control points change from having an average distance of 2.00 or less to being 15.00 or greater.

Any idea as to why the distance for the control points does not save properly after generating the panorama? Also, any help for how to avoid this would be much appreciated. (It's rather frustrating and time consuming.)

[Terry Duell]

Not sure what is causing your problem.
I have run through the process you describe and can't reproduce the
behaviour you are seeing, the project opens with the same fit statistics.
I am using hugin-2014.1.0.
Which OS and version of hugin are you using?

Cheers,
--
Regards,
Terry Duell

[dkloi]

The estimate of the control point error depends on size of the output panorama. The default size of the output may be much smaller than the eventual size you generate, hence the differing numeric values of the control point errors. Is it just the reported numbers changing or is the optimization worse upon reopening?

[cwa...@oyez.org]

Hi all,

Thanks for the response. I'm using Hugin-2013.0.0 for Mac OS X.

The reported distance for the control points changed, and when I ran the optimizer upon reopening, I received these results:

average control point distance: 21.215626

standard deviation: 16.718980

maximum: 104.639822

That compares to an average control point distance of 2 or 3, with a maximum of 6 prior to stitching. Is that what you were looking for, dkloi?

[dkloi]

Try changing the output size and then re-running the optimizer. You'll see that if you double the output size the estimated error will also double (assuming the same optimization). The initial numbers are for the initial default size for the output.