![]() |
|
|
|
|
|
||||||||||
|
| User list | Rules | You are not logged in.
I have a 'multiple viewpoint' image set that was entirely taken by one camera system with identical settings in all aspects for all images (focal length, focus, aperture, exposure time, ISO, WB...). Nevertheless, APG 3.0 computes focal lenghts that are far off from what was actually used (~300mm instead of 51mm).
In more detail: Since the EXIF data was missing, I manually entered the 51mm equivalent for all images and also set Focal and Distortion calculation to 'force identic'. I do get a somewhat reasonable stitching result, but global RMS is poor with a value of 8.8. The layer editor reports under 'Computed focal' high focal lenghts (>300 mm, Fov ~ 4.5) that also vary for each image.
Now my question is: Why can't I force APG to use those 51mm I initially specified?
Thanks for any insights
Martin
Offline
Why is the EXIF missing?
Anyway, there is another way to set focal length if EXIF is missing, I don't know whther it wll work any better for you, but it's worth a try?:
Last edited by mediavets (2012-12-06 15:54:25)
Offline
EXIF is not entirely missing, but for unknown reasons, that (very expensive) camera did not include sensor size data. So I manually need to add that.
I tried your way of setting the EXIF data, but got basically the same results. It appears as if 'multiple viewpoints' will override 'force identic'.
Martin
Offline
What camera are you using?
Why do you need to use the multiple viewpoints feature?
What sort of image set is this? Aerial photos? Handheld?
Last edited by mediavets (2012-12-06 17:02:43)
Offline
These are indeed aerial images, taken with a phaseone iXA180.
Offline
maddin wrote:
These are indeed aerial images, taken with a phaseone iXA180.
This seems to be the data relevant to thatr camera?:
Resolution 10320 x 7752 (80 MP)
Dynamic range >72 db
Aspect ratio 4:3
Pixel Size 5.2 micron
CCD size effective 53.7 x 40.4 mm
Lens factor 1.0
So what values are you entering manually in image settings window - below?
Offline
aereal images? use 1000mm focal length! could be a way...
see: http://www.kolor.com/forum/t12895-creat … aerophotos
search forum too for 1000mm ...
Georg
Offline
mediavets wrote:
So what values are you entering manually in image settings window - below?
Since APG detects the 80mm lens correctly (but does not know about the sensor), I adjust the conversion factor (= 0.64) until I end up with 51mm in the 35mm-equivalent value, which is what the vendor tells me for that lens.
I have just tried to change the initial focal to 1000 mm; full optimization brings that back down to around 680-780 mm, with a global RMS of 9.37
Martin
Offline
maddin wrote:
Since APG detects the 80mm lens correctly (but does not know about the sensor), I adjust the conversion factor (= 0.64) until I end up with 51mm in the 35mm-equivalent value, which is what the vendor tells me for that lens.
The PhaseOne web site data - quoted in my earlier post states that the 'Lens factor' is 1.
http://www.kolor.com/forum/p108279-yest … 46#p108279
But this table states that the crop factpor is 0.64, as you have written:
http://en.wikipedia.org/wiki/Image_sens … _and_sizes
Last edited by mediavets (2012-12-07 12:21:06)
Offline
mediavets wrote:
maddin wrote:
Since APG detects the 80mm lens correctly (but does not know about the sensor), I adjust the conversion factor (= 0.64) until I end up with 51mm in the 35mm-equivalent value, which is what the vendor tells me for that lens.
The PhaseOne web site data - quoted in my earlier post states that the 'Lens factor' is 1.
Yes, but what does that mean? The sensor, as you correctly found out, has a size of 53.7 * 40.4 mm (quite a bit larger than the normal 36 * 24 mm sensor on full-frame SLRs), and the lens has a true focal length of 80 mm. According to the vendor, that translates into a diagonal FOV of 47 deg, which is equivalent to a 51mm lens on a 35mm SLR. As I understand APG, it just deals with the final focal length, regardless of how it was calculated (80 * 0.64 * 1, or 80 * 1 * 0.64, or directly 51 mm).
Martin
Offline
maddin wrote:
As I understand APG, it just deals with the final focal length, regardless of how it was calculated (80 * 0.64 * 1, or 80 * 1 * 0.64, or directly 51 mm).
Martin
That's my understanding too.
So if you are getting extremely different computed focal lengths, and a high RMS, I believe that indicates that APG has created false links from incorrectly identified control points.
Have you had a look at the control points using the Control Point Editor?
...............
Did you check out the wiki topic on stitching aerial photos?:
http://www.autopano.net/wiki-en/action/ … ial_photos
Last edited by mediavets (2012-12-07 12:29:55)
Offline
Hi Martin
You are shooting with the Schneider-Kreuznach LS AF 80 mm f/2.8 ?
Then you can start with these settings, see below.
The P65 is in the sensor database (cameras.txt) with a crop factor of 0.666.
For the iXA180 it should be 0.644.
Last edited by lumelix (2012-12-07 14:02:30)
Offline
And if you like to start with the correct settings, you could copy the file cameras.txt in the User-folder
/user/data/... and edit it with the values for the iXA180:
Last edited by lumelix (2012-12-11 11:04:31)
Offline
lumelix wrote:
And if you like to start with the correct settings, you could copy the file cameras.txt in the User-folder
/user/data/... and edit it with the values for the iXA180:
Great advice, thank you!
Found my cameras.txt under
%localappdata%\kolor\Autopano Giga 3.0\Kolor\Data\
and added the line
Phase One;iXA 180;0.644
Martin
Offline
maddin wrote:
lumelix wrote:
And if you like to start with the correct settings, you could copy the file cameras.txt in the User-folder
/user/data/... and edit it with the values for the iXA180:Great advice, thank you!
Found my cameras.txt under
%localappdata%\kolor\Autopano Giga 3.0\Kolor\Data\
and added the line
Phase One;iXA 180;0.644
Martin
That doesn't work.... the file cameras.txt gets overwritten with the old content when APG starts...
So I need to create a new file under
%localappdata%\kolor\Autopano Giga 3.0\User\Data\
That seems to work.
Martin
Offline
Powered by PunBB
© Copyright 2002–2005 Rickard Andersson
|
CHOOSING KOLOR Why choose Kolor? Which solution to choose? Download a trial Where can I buy? Education |
SOFTWARE Autopano Pro Autopano Giga Panotour Panotour Pro XnView |
ACCESSORIES Training DVD Panobook PROJECTS Paris 26 Gigapixels Yosemite 17 Gigapixels |
COMMUNITY Forums YouTube channel Google+ |
COMPANY Blog About Kolor Resellers Contact Visit us |
PRESS Press center Press review TOOLS My account |
