[X] My Assistant

[Bjorn Jonsson]

I'm curious to know how people have been dealing with badly smeared images to reduce or eliminate the smear. I'm making 20000x10000 pixel global maps of Ganymede and Callisto. Lots of the highest resolution Voyager 1 images are badly smeared but I really want to use them in the maps if possible - and use them without smear (otherwise the maps get too ugly). I've been experimenting with the Smart Sharpen filter in Photoshop. I'm rather happy with the results I've had but before I start reprojecting these images and adding them to the maps I'm curious to know if anyone knows of something even better.

Here are two examples from Voyager 1 where I attempted to desmear smeared images using the PS Smart Sharpen filter. The original is shown together with a desmeared image. One is from a moderately smeared partial image of Callisto and another one is a badly smeared image of Ganymede. As indicated above I'm rather happy with the results but if something better is possible I'd like to know.

Callisto:



Ganymede:

[machi]

There are some programs specially designed to do that. I heard about Topaz InFocus or Huygens. Both aren't free and I haven't them so I'm not sure about quality (but Huygens has some good references in scientific literature).
Topaz and others are using blind deconvolution. Only freeware with this function is Image Analyzer by the Michael Vinther.
Unfortunately it doesn't work with images with more than 8 bits per channel which limits its capabilities.
Problem with blind deconvolution is that it very often fails in case of images with complex blur.
But there is the hard way of classic deconvolution with known PSF (point spread function).
If you don't know PSF, you can try to deduce it. Small bright craters are good guideline or you can even try them as PSF.
This is quick result of Wiener deconvolution with your blurred image of Ganymede.
(It's made in the "Swiss knife" of the image processing - ImageJ).

Attached thumbnail(s)