The CCD camera is much more sensitive than the eye, and the images are captured in digital format that can be manipulated to enhance subtle features. The process starts with spending time to make sure the telescope optics are collimated and in focus. A good raw image may look like this:

The CCD camera does not have an equivalent response for each pixel, and with some optical setups dust can produce diffuse dark blotches (such as here). This can be dealt with by taking a "flat field" image. This is simply an exposure of a white background and from which the response across each pixel of ccd can be determined. A flat field might look like this (yes, I know this looks like crap, and yes, I have tried to clean the CCD optical window. It used to be a lot worse, believe me):

If the pixels in the original image are divided by the corresponding pixel intensities in this flat field image, it will normalize for the non-linear response of each pixel. (Note: the dark blotches are primarily dust on the optical window of the CCD).

For long exposures there is also noise associated with stray cosmic rays and other weirdness that produces noise even if the image is a pure black background. This situation can result in noise such as the following "dark field" image:

If the pixels in this dark frame are subtracted from the original image it will eliminate the effects of this "dark noise". (Note: the dark frame is subtracted before the flat field compensation).

Here is the effect of the dark frame and flat field compensation on the original image:

The atmosphere is constantly in motion and distorts any individual image. Furthermore, the stars and planets are moving relative to the earth due to the earth's rotation. Although the telescope mount can move to compensate for this, it also can introduce distortions due to jiggling. A solution to this problem is to take multiple images and average them. Here is the effect of averaging 12 images in comparison to a single image:

Noise and distortions within a single image are now minimized in the averaged image. The averaged image is a little fuzzy. If we had an image of just the "fuzz", we could subtract it from this image and leave only the "sharp" part of the image. The question of obtaining an image of the "fuzz" gets back to the subject of our original images. Although 12 images were averaged in the above picture, far more than 12 images were collected. In fact,120 images were collected and only about 10% were reasonably sharp. The other images were slightly blurred, and some were extremely blurred. If we take the worst images from our 120 (and average them) we will have an image that describes fairly well the nature of the "fuzz" (this is termed an "unsharp mask"):

We can substract this image from our averaged image to remove the "fuzz" (actually we will subtract about 40% worth of this image to avoid excess noise in the result)

Here is a comparison of our original single image with the averaged image minus the unsharp mask:

There are various other sophisticated digital processing steps beyond these simple ones, and the final results can be impressive:

This is truly amazing (particularly when you realize that the automated focusing software, the image stacking software, and the image deconvolution software are free), but leads to some disappointment when people think that such final images are what you will see when you look through the telescope. Some people never actually observe through their telescope, preferring instead to eke the best image out of their digital photos.

Links to software:

© 2003 Michael Blaber