CCD Image Processing From Novice To Expert

[Darryl]

I have been thinking about adding a CCD Image Processing forum for quite some time. So here it is. I for one find image processing just as important as taking the image if not more so. I can divide getting that great CCD image into two categories when it comes right down to it. Take the ccd image or images in the case of stacking, and process the image. Depending on your attention to detail or experience processing the image can just as long and be just as challenging as taking it, and can make or break the final result.

I don't have any type of personal work flow at the moment when processing an image because I haven't found one I'm happy with yet, and could still consider myself a novice when it comes to this part of Astrophotography or any other part for that matter. I am actually not sure when one could be called an expert anyway because almost every time I go out to take images and then do the processing I'm always learning and refining new techniques and working on perfecting ones I've already used.

[dmill1220]

HI Darryl,

I have developed a definitive work flow but only for the image acquisition, and that is very specific to the camera, software, and scope combination that I use

I do seem to have to a certain degree, a work flow when post processing, but like you I am still learning and I try different things all the time.
I am definitely a green horn in CCD imaging!

Since I use the Meade DSI 1st generation camera, it has its own proprietary image processing software, but I do augment it with other software such as photoshop7.0, Pix Insight LE, and Registax

Here is what I normally try to do with each image.

All my images are acquired in the fits format, that are broken down into L,R,G,B,
the Meade DSI, although a one shot color camera, automatically breaks down the image into LRGB components when I use the save every composite image function.

while the image displayed on my screen while acquiring, is automatically stacked and the dark frames are also automatically subtracted, when you go to where the fits images are stored they are broken up into the LRGB components already dark subtracted.

1) I start by opening all the blue fits files as a group,( they are already dark subtracted at this point automatically) and I examine each Blue fits sub, I have to prescale each separate blue fits sub as I examine them with the Autostar image processor, but I then throw out any blue sub that I dont like, before I stack them. why Blue first? I have an achromat so blue halos around brighter objects can be reduced before stacking and combining.

when I have removed all of the bad blue Fits subs that I dont like, I stack and align the blue fits images by using the two star method with the avg combine function selected. Once this is done I then perform a linear stretch, adjust the histogram and then perform a log function stretch. then I adjust the image brightness and contrast. the blue fits file is then corrected for any hot or cold pixels, and an unsharp mask is applied if needed, then the file is saved as the master blue fits file.

So basically I am selectively opening each color as a group except for the L channel, working with one color at a time and manipulating the data as I described above, and making a master for each. what I am left with at this point is a master fits file that has been adjusted, prescaled, linear and log stretched, for each RGB channel.

At this point I also take my master fits files for each RGB channel and make a copy/conversion to a BMP, a Tiff, and a JPEG.

2) Now at this point, still using the Autostar image processor and working with the Master fits files, I perform a Color merge where the software combines only the RGB channels and not the L channel, Each master RGB channel is then selected and again individually prescaled with a histogram before it is merged, this results in a RGB composite image for which I then perform a RGB alignment.

3) The next step I do is to adjust each color level in Autostar by changing the color weights or balance, I also adjust the gamma for the combined RGB composite to bring out more detail, and I adjust the contrast and brightness levels, at this point I save the RGB composite in a fits, tiff, and BMP format.

4) now I work with the L channel subs and perform all the steps with this channel that I did for the RGB channels, a master L channel is then saved in the Fits. Tiff, and BMP formats.

What I have left now at this point is a I have an RGB Composite image in fits Tiff and Jpeg, and a master Luminance channel also saved in the various formats.

It is at this point were I now use either photoshop 7.0 Pix Insight LE, or Registax to experiment with different adjustments.

I have discovered that with the DSI one shot to get the luminance channel it is better to take your subs as a mono during acquisition, so I have to image each object twice once in full one shot color mode, and one in Mono for the luminance channel.

Its all experimentation at this point, but I have found that photshop makes it easy to combine the L channel with the RGB composite. with the DSI Meade software, there is no simple way to accomplish the LRGB merge.

I have been heavily experimenting with Photoshop 7.0 I have figured out how to reduce star bloat and how to remove blue halos since I use a plain achromat for image acquisition. I got a lot of help[ with photoshop online here at this link, Tutorials

I plan on getting my grandson to help me by filming a complete imaging session and post processing procedures, for my setup.

I hope to show some of our imaging experts at my local astronomy club the film, in the hope that they can correct me during each process or offer advice or other methods of post processing.

when we are imaging at the observatory everyone is way too busy with their own session to have to teach anyone else.

Dennis