Saturated RGB Pixels

[alaskacam]

I keep having this annoying problem with scattered saturated red, green and blue pixels in my processed photos. It only shows up when I'm really pulling hard on faint galaxies and I'm at the limit of what processing can handle.

I have attached a small example shot. It's cleaned up about as much as it can be with the limited number of subs on a faint object. You can see the scattered hot pixels and I don't know how to nuke them out or fix some flaw in my workflow. Here is a bit on my workflow...

I'm imaging with a SBIG ST-4000XCM color CCD. No filter is used at this point as I can't get a deep sky filter mounted in my optical train. At the beginning or end of the session I take my darks. Usually at least three 1x1 darks that are average combined into a master dark. My master flat is a from 100 lights and 20 darks, which I retake anytime I change the optical train. The lights are shot at 1x1 at the same time and temp as the darks. In the case of the example shot, it was 3 x 15 minute lights at -20C. All lights, darks and flats are taken with CCDSoft. I'm currently using Nebulosity to process the images. I stack with the Master Dark and Master Flat in an average combine. I don't always follow the same exact workflow with the stacked image to get my final, but if I pull hard on the faint contrast I get these dots. Will Nebulosity or other processing packages allow you to zap these pixels or would there be a weakness in my workflow somewhere? Thanks in advance..

Rob

[roundycat]

Your big mistake is to combine using an average output. Try some form of outlier rejection if your software is up to it. Failing that a hot pixel filter applied to each frame before combining might help. It will not do anything to hot pixels in pairs or other groupings.

Dennis

[alaskacam]

Thanks Dennis. I'll try a couple of other stacking options.

Rob

[RobH]

To add to what Dennis says, also you are not using anywhere near enough darks, and may well be adding noise there.

You need at least 20 darks, and these can be taken anytime as long as the temperature matches.

You also need at least 20 flats, and these can be bias subtracted....for bias frames, as they are so short, use at least 50.

Rob

[ollypenrice]

OSC seems susceptible to this. Mine certainly is. More darks, more flats, and run a hot pixel filter when combining your flats. Dithered guiding helps, too. Using Astro Art I run the hot pixel filter (default 20 percent setting) at the stacking stage when the light images are debayered. This is very, very effective and seems to have no adverse effects at all.

Olly

[alaskacam]

Thanks guys. Will experiment....