Re: Planet X: Alternative Explanation 2

Bill Nelson wrote:
> 
> In sci.astro The Small Kahuna <person@company.com> wrote:
>>>
>>> On what basis would you call such a thing "verified"?  What's to
>>> prevent you from claiming that the picture was really taken at
>>> some other time?  What's to prevent you from claiming that the
>>> object was really there but edited out of the image?
> 
>> Nothing really.  Nothing that is, but the self consistency of the data.
>> I have proposed regular pictures including differential processing.
>> Regular pictures would all be slightly different.  The noise
>> characteristics could be measured.  Anything other than a Really
>> Professional editing job would be patently obvious with some relatively
>> simple signal processing.  And believe me, if legitimate pictures were
>> posted, I would be processing them looking just for these artifacts.  I
>> would also be aligning the images and doing correlation calculations.  I
>> would be doing differential processing.  I would accuse the poster of
>> faking the images when I could prove it.  That's because there would be
>> data.
> 
> The trouble is - for the size and distance that Nancy has claime for this
> supposed object - it would be a point source. In other words, it would
> not appear diffuse - and would only cover 1 pixel on a CCD image. There
> is no way to differentiate such from noise - or detect if that one pixel
> had been edited.

Sure.  It would move.  Also, it would cover more than one pixel because
of several things: 1) diffraction limits and optic imperfections would
make the point spread function probably cover additional pixels 2)
anti-aliasing filters in the camera would add to that point spread
function 3) almost all color cameras are tiled with the Beyer pattern
and most amateurs have color cameras  (But not all).

Noise would be detectable from image to image.  The dark pattern could
be estimated given more than one image.  The characteristics of the
noise could be analyzed.  The point spread function could be estimated
and compared to the pixel in question.

The point is that it would be possible to fake an image, to be sure.  It
is also possible to detect most fakes.

Still, I have as yet to see an image, faked or otherwise, so my computer
sits idle. 

> 
> Bill Nelson   (billn@peak.org)

The Small Kahuna