Hello All:
I am trying to improve a little project I am working on. Let's just say by the end I hope to have a tool to help automatically extract a PSF estimate for terrestrial images.
In the meanwhile I have a little problem. Assume you have an image with a multitude of spacial "frequency" components, i.e. large and small structures. Imagine then applying a filter, specifically a convolution over the image. Depending upon the particular filter certain spacial components will be attenuated (averaged out) or effectively "removed" from the image.
Knowing a rough idea of the spacial size distribution in the original, and knowing the exact filter applied, IF someone then added noise to the image... is there a way to isolate and remove the noise given the knowledge of the "frequency" content that is in the actual image versus that which arguably could not have existed in after filtering with the known filter?
I'm wondering if there is a friendly way to filter an image... with knowledge of the PSF ... if the noise was added after. Seems to me that knowing the limits of an image after it has gone though a filter should give a huge hint that something is noise added after...
Just pondering...
Could one theoretically take a slightly out of focus image on purpose to better isolate CCD errors? I know that's pushing it...
cheers
C
