[ptx] vignetting function estimation

[Pablo d'Angelo]

JD Smith schrieb:
> 
> On Feb 5, 2006, at 3:32 PM, Pablo d'Angelo wrote:
> 
>>
>> 2. Graphical display of vignetting correction polynomial.
>> 3. Feature to optimize this polynomial based on the overlapping  images.
>>    This function can only estimate the polynomial for the quotient 
>> approach
>>    right now.
>>
>>    I will post a tutorial on this later (in a week or two).
> 
> 
> Very cool.  I look forward to trying it out.  One possibility for 
> corrections which should be stable, especially these vignetting 
> correction features, is to average many many results from different 
> panos.

Hmm from my experience, the solutions are quite stable if the images are
"good". I don't think one needs another average step after that. I suspect
that white balanceing might also add some variance, especially if gamma
corrected data is used.

I have made some tests with photographs of my monitor displaying a
uniformly white image. My TFT monitor seems to be quite evenly illuminated,
even when looking from slightly different different angles.

Estimating the falloff from a few rotated views work quite well.

The resulting curves look quite ok when compared to a flatfield image (using
the same monitor as uniform light source, which is probably less accuate,
due to changes in viewing angle over the whole image.

I'll need a really high quality flatfield for comparison. Or does anybody
know a good site which has measured light falloff curves for one of these
lenses: EF 50/1.8 II, Sigma 18-128, Carl Zeiss Jena Sonnar 135 3.5 or Peleng
8 mm Fisheye?

ciao
  Pablo