Posted by
Frederic V. Hessman on
URL: http://imagej.273.s1.nabble.com/Extensible-Persistent-Properties-for-ImageJ-Images-tp5023228p5023245.html
Indeed, this is exactly what astronomers try to do, but with a big difference: they (usually) know that the underlying thing they're trying to image is something simple - usually a point source or several point sources. This makes the problem mathematically MUCH simpler. If you're trying to un-distort a complex image, there's still a method, but it's extremely complex, mathematically.
If you assume that the image is distorted by shifting alone - probably not too bad an approximation - you can use the a simple shift-and-add algorithm suggested by Gabriel. Amateur astronomer make fantastically detailed images of planets using video cameras on small telescopes using this assumption (after throwing out most of the images that are badly distorted). Here, you want to do much larger images, so for each sub-image (must be on a scale that is larger than the distortion variations in your image, here the wavelength of the waves), measure the 2-D shift of each image relative to an average image, correct for the shift, and then co-add. In order to combine the sub-images, you may have to dither by using overlapping sub-images.
One effect you can't easily remove using this method is the presence of the dark and light bands, cause by the refraction of the sun's light by the waves themselves.
Rick
> On 19 Apr 2020, at 12:43, Gabriel Landini <
[hidden email]> wrote:
>
>> This morning whilst walking the dog around the bay I stopped to stare into
>> the sea, as I often do. It's so hypnotically beautiful.
>> Here it is:
http://ellisnet.uk/Rants/Water1080p.mov>> Then I started to wonder. Is there an image processor that can remove the
>> distorting effect of the water such that it produces an image of the sea bed
>> as if the water were not there?
>
> Hi, Interesting question. It sounds similar to what astronomers do with
> adaptive optics/mirrors to get rid of the effects of observing through
> atmospheric turbulence.
>
> I wonder if by doing an average projection of the (registered) stack on a long
> enough sequence, could minimise the flucutations a bit.
>
> Cheers
>
> Gabriel
>
> --
> ImageJ mailing list:
http://imagej.nih.gov/ij/list.html--
ImageJ mailing list:
http://imagej.nih.gov/ij/list.html