Sergei Golyshev (AFK during workdays)
Laboratorium Solaris: Workshop "Math vs Physics"
Acqusition time: 22.08.2016 around 09:25 MSK
TIS DMK 23U274 on Coronado PST
140 out of 1000 frames were stacked in AS!2, deconvolved in AstraImage 3.0 PRO (Richardson-Lucy aggressive, Cauchy-type, 0,3 pixels, 11 iterations). Contrast enchancement and masking-blending were done in PS...
What we get is on the left. What is obvious is the brightness gradient from the center of the solar disk towards the edges. I know at least two version of why this happens with PST, and while this effect gives the Sun "spherical look", I don't like it. What can be done?
To all intens and purposes this is vignetting. Vignetting can be supressed either by removing the components which cause it (physics), or by flatfield correction (math). Modding PST is a possibility, but I don't wholeheartedly embrace it. So I went the path of math. I have taken my already processed image into ImageJ and made a copy of it. The copy was subjected to Gaussian blur with radius of 15 pixels. And the I DIVIDED the source image by blurred image using ImageJ Image Calculator function. The result was 32-bit image with negative and fractional brightness values :). It looks ugly but has all the information preserved*. I had adjusted the histogram of to include all the usefull signal and converted it into cozy 16-bits. I have taken it back to PS and placed it on top of my procesed image. Some contrast tweaking and masking-blending resulted in the image on the right. It looks pancake-flat , which may not be pleasing, but it has even and high contrast across the face of the Sun.
Opinions are welcomed ;)
*Close inspection had revealed some pixel-level junk bits, but they are tolerable.
Laboratorium Solaris: Workshop "Math vs Physics"
Acqusition time: 22.08.2016 around 09:25 MSK
TIS DMK 23U274 on Coronado PST
140 out of 1000 frames were stacked in AS!2, deconvolved in AstraImage 3.0 PRO (Richardson-Lucy aggressive, Cauchy-type, 0,3 pixels, 11 iterations). Contrast enchancement and masking-blending were done in PS...
What we get is on the left. What is obvious is the brightness gradient from the center of the solar disk towards the edges. I know at least two version of why this happens with PST, and while this effect gives the Sun "spherical look", I don't like it. What can be done?
To all intens and purposes this is vignetting. Vignetting can be supressed either by removing the components which cause it (physics), or by flatfield correction (math). Modding PST is a possibility, but I don't wholeheartedly embrace it. So I went the path of math. I have taken my already processed image into ImageJ and made a copy of it. The copy was subjected to Gaussian blur with radius of 15 pixels. And the I DIVIDED the source image by blurred image using ImageJ Image Calculator function. The result was 32-bit image with negative and fractional brightness values :). It looks ugly but has all the information preserved*. I had adjusted the histogram of to include all the usefull signal and converted it into cozy 16-bits. I have taken it back to PS and placed it on top of my procesed image. Some contrast tweaking and masking-blending resulted in the image on the right. It looks pancake-flat , which may not be pleasing, but it has even and high contrast across the face of the Sun.
Opinions are welcomed ;)
*Close inspection had revealed some pixel-level junk bits, but they are tolerable.