Re: [hugin-ptx] Control points - autopano-sift-c

[Carlos Eduardo G. Carvalho (Cartola)]

I have used autopano-sift in the past, but I think I've read around that it is kind of obsolete, not sure. Nowadays I am starting to use cpfind with some success. It is a little different, you need to create a pto file with the images loaded and then run it on the pto. In fact you can work with it in many different ways. You can mark as many images you want in the Images tab and run in through the "find control points" button in the bottom of the tab.

As in most of times I use a computer with little memory, I am having success creating the *key files in advance. I save the PTO file with the images, then run a cpfind for each image:

cpfind -k 0 file.pto
cpfind -k 1 file.pto
...

and it can be in a shell script or a BAT file, depending on your sistem. When you have the key files already generated the control point detection is much much faster. Besides this, it make the auto detection possible in machines less powerful. I wish hugin had a way to generate those key files one by one automatically. Is there?

Cheers,

Carlos E G Carvalho (Cartola)
http://cartola.org/360
http://www.panoforum.com.br/

2012/11/1 gilman <gilg...@gmail.com>

Hi All,

Ive been using Hugin for a little while not with nice success. But Ive been defining my own control points which can get tedious. Im trying now to get autopano-sift-c to work. But Im getting wxExecute errors telling me it cannot run the autosiftc executable i defined in the control points section under preferences.

When I investigate further and actually type in the autosiftc command it is trying to execute in a command line, I get an error back saying the image im trying to use is not a valid. Any idea how to fix this or what it means?

Log below:

Thanks so much!

-Gilman 

C:\hugin_ap\bin>autopano-sift-c.exe --maxmatches 10 c:\temp\ap_181.tmp c:\temp\ap_182.tmp

APSCpp, enhanced autopano-sift-c version 2.5.2 23July2009

Reading project file c:\temp\ap_182.tmp

  Default fisheye lens type is equal-area.

  Focal length will be computed from hfov.

  Stereographic projection enabled for hfov >= 65.0 degrees.

Filename C:\M1\m1_006.tif

We only support 3 or 4 samples per pixel

Syntax error: Not a valid image

--
You received this message because you are subscribed to the Google Groups "Hugin and other free panoramic software" group.
A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugi...@googlegroups.com
To unsubscribe from this group, send email to hugin-ptx+...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

[Bruno Postle]

On Nov 1, 2012 8:48 AM, "gilman" wrote:
>
> We only support 3 or 4 samples per pixel
> Syntax error: Not a valid image

This means that it only works with 8bit per channel RGB images with or without an alpha channel. So you can't use HDR or 16bit or greyscale images.

cpfind, which is part of recent hugin releases, is much better at this sort of thing than autopano-sift-c.

--
Bruno

[dex Otaku]

This means that it only works with 8bit per channel RGB images with or without an alpha channel. So you can't use HDR or 16bit or greyscale images.

Autopano-sift-c works fine with 16bpc images [or at least, it has here, on windows, for the past 4+ years].  The error is almost definitely not caused by 16bpc images.  

I would check that the TIFF images in use may have layers in them [that would result in more than 3 or 3+alpha colour channels], that they're not >16bpc, or that they're not CMYK or something.

FWIW, I find there are cases where autopano-sift-c does a way better job of detecting points, but cpfind is mostly caught up in terms of functionality [compared to 3+ years ago].  I tend to start with cpfind and then try alternatives if it doesn't work out, these days, because one [CP detector] may often work well in situations where the others don't.