Red noise in image - any way to get rid of it

[mspruce]

Hi all!

I'm new to Pixinsight but I have already done a couple of images with my Atik 314L+ and am blown away by what I can do. Needless to say I'm sold on it.

However I have a problem that I have been unable to solve. I decided in a moment of madness to try me DSLR and do a couple of OSC images with it. However once I had processed them to generate the initial integrated image to begin cleaning up, stretching, etc. I found that there was what looked like concentric red circles in the image. I have tried using several techniques to try to mask it out  and stretch the object I want to make it more visible and keep the red circles down and had some success, but not enought. I wondered if anyone had any suggestions, techniques, etc. that I could use to eliminate it better or completely. Also does anyone have any clue as to what may be causing the effect? I took two objects that night and they both have the same pattern. I'm not using an LP filter but I would have expected LP to be more even across the image.

I have attached a jpg (stretched to show it all up) of the initial image before cleaning up if you want the un-stretched xisf file to have a play with then just shout.

Many thanks for your help

[cdavid]

Hi!

Those look like they may be internal reflections?  Did you flat calibrate the images?  Did you use DBE during the processing prior to stretching?

-C

[mspruce]

Hi, Yes I did use DBE I tried both subtraction and division. Division does a slightly better job but washes the colour out of the nebulosity more. Neither get rid of the issue completely although I have not really played with the settings that much. Do you think that DBE should be able to get rid of most of this problem if I have a play with the settings. If so any clues what may help?

I think you must be right about internal reflections. I just need to work out where they are coming from and how to eliminate them. do you think an LP filter might help?

[cdavid]

I think a couple of runs of DBE with division and then subtraction might help.  Perhaps post a link to the image and I can see what settings might work.  In terms of the reflections- these are most commonly from the adapters between the focuser and your camera.  If they are internal from the scope- they will be more difficult to deal with.  What I have found in dealing with reflections from my adapters, is painting them with black camouflage paint or lining them with black felt will eliminate them.

[mspruce]

I think I have worked out where the red light source is coming from. The camera, before and after shifting the mirror to expose the CCD shines a red led to illuminate the viewfinder grid. I think I'm getting internal reflections from this and a few photons are managing to get to the sensor somehow, once a load of images are stacked then this effect builds up and you get the red cast that I'm experiencing. So now to experiment with the settings to prove/disprove/stop this.

Unfortunately it does not solve my problem with the current images. If there is no way to minimise/remove the issue then I will simply resort to having to live with it and put it down to experience. Its certainly a good challenge anyway.

[mspruce]

I think a couple of runs of DBE with division and then subtraction might help.  Perhaps post a link to the image and I can see what settings might work.  In terms of the reflections- these are most commonly from the adapters between the focuser and your camera.  If they are internal from the scope- they will be more difficult to deal with.  What I have found in dealing with reflections from my adapters, is painting them with black camouflage paint or lining them with black felt will eliminate them.

Thanks for the suggestion. I will give it a try when I get chance. If I get really stuck I will need to find somewhere to post the image for you to have a play with as its big (1.21GB !!!, the down side to using a 36 mega pixel camera) I will give it a try and let you know. 

[jkmorse]

There is a quick, way to fix the problem, though it is not entirely true to the data.  But it will save your image.  First, make a clone of your image, then convert it to greyscale and invert it so that the clone shows the background white and the target and stars as dark.  You are likely going to need to use Histogram Transform to get it right.  Once you have it set, apply the clone to your main image as a mask.  Now you are protecting the stars and target and can work on the background alone.  Open Curves Transformation and choose the saturation button on the far right (labelled "S").  Now just lower the curve to desaturate your background.  It may take a couple of iterations but in the end you will have a nice grey background.

Jim 

[mspruce]

Hi Jim,

Many thanks for the suggestion, I will give it a whirl when I get chance. I'm not too interested in being true to the data at this point. I just want to try to salvage something decent from it. We get few enough nights that are clear these days so I want to get all I can from what I have.

The other advantage with having a problem image like this is that its the best opportunity to learn. If you processed all your images using the same techniques all the time then you learn nothing. So I'm looking on this as a great learning opportunity.