Successful use of the new Normalized Scale Gradient script with CMOS Data - M101

Roger Smith

I did not think it was possible my data set was useable. My raws have varying sky gradient backgrounds, almost no visible image in a sub, and tons of noise. The best background was .168, and anything above .3 I rejected. No, these are not flats, but raw images. The arms in the best image are about .173 With a normal screen stretch you cannot see M101, well  maybe a smudge. With STF controls I can barely see it. .

But I saw Adam's video on the pre-release NSG script (now found in Scripts/Batch Processing) I dug in and started my work. I re-watched Fundamentals videos and others. DBE was the most difficult thing after NSG. Eventually I separated the RGB Channels and attacked each one with variations in sample points. I got it as best I could, and knocked down the background color.

The data also had red fringing on brighter stars, not due to focus differences because this is a color camera. I have ZenithStar ZS103 APO. So maybe chromatic aberration. No matter the cause, I had it. And I had to get rid of it. So my new friends in PI are all masks! Game masks, star masks, and Pixel Math. For me, I learned to do all this from Adam's videos. Thanks.

To effectively work on the calibrated, aligned, integrated image (using NWEIGHT) and PCC, I extracted the Luminance and processed it separate from RGB image. Deconvolution work well (first time it ever worked well for me) on the luminance. 

I used History Explorer several times to go back to the integrated image and start over. It was a learning experience. In the end I maintained most of the History in the final image. Learning from Adam's Fundamentals helped me to understand it and how to use Pixel Math to write the history onto the main image.

Attached are a couple of .jpeg's of the final (I think) image.  One is uncropped. One is zoomed at 1:2, which is as brave as I can be.

Roger