|As Digital and CCD imagers become more affordable, more and more people are using them for astrophotography. As
these cameras are used more often, different imagers have developed their own methods of image processing that work
well for them and acheive consistantly excellent results. Advanced astroimagers like Russell Croman and Ron Wodaski
have even published books discussing their methods and allowing others to try them out. One area that seems to be
lacking, though, is a good description of how to process images captured by an SBIG One-Shot color camera like the
ST-2000 XCM. This is not an exhaustive end-all processing tutorial, but it should help those of you new to using One-Shot
SBIG cameras to acheive excellent results with a little effort.
For this tutorial, I will use the raw frames of an NGC 253 image that I took from the OPT San Pasqual observing site on the night of
August 26, 2006. Click on any of the images to see them full size. Many of the steps will have a corresponding image - hopefully
making it easier to understand what is going on. I use Photoshop version CS2 and CCDOPS version 5, but most steps will work
equally well in Photoshop CS and CCDSoft or MaximDL.
Part I - CCDOPS
1. Step 1 is to acquire an image. I use SBIG's CCDOPS version 5, which came bundled with the ST-2000 XCM, but you can also
CCDSoft (also bundled) or Maxim DL 4. I won't go into discussion of polar alignment or autoguider calibration very much here, as
there are many other manuals and books that cover those topics (Ron Wodaski's The New CCD Astronomy is an excellent
example). I will just say that you will want to make sure your scope is aligned and balanced, your cooling system is running and stable
(no more than 85% capacity), any filters (IR blocking, IDAS, H-Alpha, etc) are in place, and your autoguider is calibrated. The picture
at right shows a single, raw (unstretched) exposure of 15 minutes at -10 degrees C acquired in CCDOPS. The picture is not yet in color, but
we'll get to that later. Save the image as a .sbig file (I create a new folder for every month to keep my images semi-organized). This
image will look much brighter and noisier in CCDOPS, because CCDOPS shows only part of the data at one time - and tends to emphasize
2. Now CCD images, even SBIG cameras with the cooling running full blast, have some inherent noise due to heat (amp glow),
misbehaving hot or cold pixels, and lack of signal in the darkest areas of the image. This requires the subtraction of a "Dark Frame",
which is simply an equivalent exposure (same time and temperature) shot with the camera's shutter closed. This allows you to remove
the first two types of noise easily, while the third type will be covered later. Here is a "stretched" (brightened to show irregularities) 15
minute Dark Frame shot at -10 degrees C acquired in CCDOPS. (Dark Frames can be shot before the night of your image, but should be
re-shot every 3-6 months to deal with "migrating" hot or cold pixels). To take a Dark Frame, select [Grab] from the menu, and set the
exposure time to equal the length of your intended "Light" Frame (eg 900 seconds = 15 minutes). Save this image as a raw .sbig file.
3. The way to use a Dark Frame to improve a raw image is simple. First, load the aquired image in CCDOPS. Then, using the menu
top, select [Utilities] -> [Dark Subtract] and choose the correct dark frame file. To make this easier, save your dark frames in a
folder called "Darks" and name them by time and temperature (such as "900_sec_-10C_Dark1.sbig") Here is the original image with the
dark frame subtracted. The picture doesn't look different now, but if I had applied filters like [Curves] in Photoshop CS2 before
subtracting a dark frame, I would have had an image looking like a galaxy shot through a snowstorm. In CCDOPS, the difference
between a raw image and a dark subtracted one is readily apparent, though. Either save this over the raw image, or save it as a new .sbig
4. Now, I am ready to deal with the color data. Unlike a monochrome CCD camera, the SBIG ST-2000 XCM acquires all of its
color data in a single shot through the use of Red, Green, and Blue filters layered over each pixel in a checkerboard pattern (each 2 x 2
square has 1 Red, 2 Green, and 1 Blue sensitive pixel). The raw image in CCDOPS will look as if it was shot through a screen door
due to the differing color sensitivities of adjacent pixels, but once the image is color-converted, the software fills in the gaps in each
color channel. In CCDOPS, go to [Utilities] -> [Single Shot Color] -> [Color Process] to convert an image to RGB color. You can
pick either RGB or DDP mode (read the One-Shot Color addendum to the SBIG manual for information about the differences between the
two). Here is the 15 minute image converted to color using DDP mode. Save this image as an 8 bit .tif file.
5. CCDOPS can only convert images to 8 bit color (256 brightness levels per chanel) RGB images, though, so in order to keep the image in
much smoother 16 bit (65,000+ brightness levels per channel) workspace, you also have to extract a 16 bit monochrome luminance
image (generated synthetically in CCDOPS) for each image file. To do this, go to [Utilities] -> [Single Shot Color] -> [Extract Color
Channel] and choose [Luminance] for the option. This file will look nearly identical to the raw image, except the "screen door" effect will
now be gone, and the noise will be smoothed out somewhat. Save this image as a 16 bit .tif file.
6. For most images, you will have several frames shot of the same target. "Stacking" multiple images of the same target in Photoshop
helps increase detail in the darker areas of an image, and helps decrease "random" noise by averaging the values of all used frames
together. Remember to extract both an 8 bit RGB color and a 16 bit Luminance image for every frame you shot of the target and save them
all as seperate .tif files. In this case, I shot three 15-minute exposures of NGC 253, so I have three RGB files and three Luminance files.
Part II - Photoshop CS2 - Luminance
7. In Photoshop, open the first of the Luminance files and zoom the view in to 100%. The image will look dark like images 1 and 3 at this
stage because it is unstretched - i.e. showing the entire range of brightnesses. All 65,000+ Values are in this image, although Photoshop
only displays 100 (in Grayscale mode) or 256 (in RGB mode) values at one time. Next, open up the second Luminance frame and
stack it on top of the first one at 50% opacity, making sure to line up the stars. THe third frame is added at 33% opacity. Continue
this until all Luminance frames are stacked before merging down the image into a single composite frame. For a much more detailed
discussion on the how's and why's of stacking, see Steps 3 - 7 in my article Intro to Astronomical Image Processing in
Photoshop. Make sure to save this file as a seperate 16 bit . tif file with a name labeling this file as the Master Luminance frame for
your image (like "NGC253_3x15min_MasterLum.tif").
8. Before you get to the color data again, you will need to use [Levels] and [Curves] to pull out the hidden detail in the Master
frame. These tools are found under the [Image] -> [Adjustments] frame along the top menu bar of the image window. Multiple
iterations will almost certainly be needed in order to get the most out of the data - the majority of which is low in the "shadows" part of
the image. You will also use [Levels] to set the "Black Point" of the image - the value below which everything is displayed as black.
See my previous article or read The New CCD Astronomy by Ron Wodaski for more information on how to do this. Here is the
image after setting the Black Point and adjsting the contrast once in [Levels]. Notice how much better the image looks already.
9. Continue making adjustments to [Levels] and [Curves] until you get the image looking as bright and contrasty as you can. Be careful
not stretch the image too much, as you can begin to show too much low level "Read Noise" from the camera to mess up the shadows,
and the image may look artificial. At this point, I would convert the file to 16 bit RGB mode (which may change the look of the image
slightly - usually making it look a little smoother) and use [Filter] -> [Noise] -> [Noise Reduction] to clean up some of the graininess in
the darker areas. You may need to use the Clone Stamp tool to clean up dust specks or residual hot pixels the the Dark Frame
removal missed (there are always a few). Save the image as a new 16 bit .tif file such as "NGC253_3x15min_MasterLum_Adj1.tif".
Here is the image after several more iterations of [Levels] and [Curves] in Photoshop, plus noise reduction and hot pixel / dust cleanup.
10. I highly recommend buying Noel Carboni's Astronomy Tools for Photoshop. I use several of these actions everytime I process any
of my astrophotography images. I used the [Deep Space Noise Reduction] and [Local Contrast Enhancement] actions on the image in the
last step, and more of the actions will be used later when the color data has been added in.
Part III - Photoshop CS2 - Color
11. Repeat the stacking process with the RGB color frames as explained in Step #7 in the last section and save the flattened image as a
new .tif file such as "NGC253_3x15min_MasterRGB.tif" denoting the image as the master color data. Again, see my previous article
for a more in-depth discussion of the ins-and-outs of stacking images. Make sure to line up all frames on the first image in the stack, so
that your Luminance and RGB images will not be shifted relative to each other when they are combined. The color data wil be much
noisier than the Luminance image, so I like to apply a Gaussian Blur filter - [Filter] -> [Blur] -> [Gaussian Blur] - at this time with a
Radius of around 1-1.5 pixels.
12. Copy the Master RGB image and paste it on top of the earlier Master Luminance image. This should place the slightly blurred
color data into a second layer on top of the monochrome luminance image. In the [Layers] window, change the blending more from
[Normal] to [Color] and you should see the combined colorized image. You may need to use [Levels] to set the Black Points of each
color channel and [Curves] to stretch to color data before before flattening the image. Once the image looks halfway decent, flatten it
down to a single LRGB layer and save it as something like "NGC253_3x15min_LRGB.tif" Keep the image in 16 bit and uncompressed
for now. Here is the newly combined LRGB image of NGC 253.
13. If you check the Red, Green, and Blue channels in the [Channels] window of Photoshop, you will notice that the background for
channel will look quite noisy. This noise comes parly from the checkerboard-pattern Bayer matrix overlaying the CCD sensor on
color cameras and partly from uncertainties and noise inherent in the shadow area of almost all astronomical images - due mainly to
exposure (especially compared to the much greater exposure data captured with normal daylight images). Here is a close-up of the color
in the background of the LRGB image.
14. This color noise must be cleaned up before final processing can be done. I recommend using the [Remove Color Blotching] action
from Noel Carboni's Astronomy Tools - mentioned previously in Step #10 above. A similar but less efficient and lower quality effect can
also beacheived by using a [Gaussian Blur] filter of 1-3 pixels and then using [Edit] -> [Fade Gaussian Blur] and changing the blending
to [Color] - but this often causes fainter stars to lose their color, so be careful. Here is the same close-up image of NGC 253 after using
the [Remove Color Blotching] action.
15. From this point onward, what you do to the image to improve its appearance is completely optional and open to interpretation. Another
iteration or two of [Curves] may help to balance the highlights better or bring just a little more detail out of the shadows. You also may
need to reset the Black Point in [Levels] slightly if you do any further noise reduction to the image. I usually boost the color saturation
in [Image] -> [Adjustments] -> [Hue/Saturation] to bring out the star colors and contrast a bit more, but the amount of actual
adjustment depends on the individual characteristics of the image in question. Lastly, I save a final 16 bit .tif file of the image before
converting to 8 bit RGB mode and creating compressed .jpeg images in different sizes for posting online.
Here is the final image of NGC 253 after all processing is complete.