Calibration is the very first step in astronomical image processing. Astronomical images are calibrated to eliminate as much as possible the unwanted effects produced by a CCD camera and light polution. These effects are:

Base background level:

Thermal effects add unwanted intensity to the image. This is why CCD cameras are cooled. However, unless you can cool your camera under -50º centigrades, thermal effects will always show up.

Additionally, camera electronics produces noise when reading and transmitting the image to the computer. This is the readout noise.

One or several dark frames are taken to compensate for above effects. Dark frames are taken preventing light from entering the camera. This is usually acomplished by placing the protective cap in your scope. CADET requires that dark frames are taken with an exposure time equal to that used for the astronomical images being calibrated.

It must be noted that calibration will compensate for average additional intensity added by thermal and readout noise. It will not eliminate the fluctuations. To compensate as much as possible for these fluctuations, which are the real noise, the only solutions are either take longer exposures or take several images of the same object and add or average them. The first solution is preferred. However, the second one must not be discarded, as it allows us for unguided imaging within the limitations of the scope mount. Both methods improve the signal to noise ratio.

When substracting dark frames from the raw image, the noise in darks is added to the noise in the image. Due to this, it is better to average several dark frames to get a master dark frame. CADET allows to specify several dark frames during calibration. They will be automatically averaged.

Difference of sensitivity between pixels:

Current technology does not allow for perfect CCD cameras. This means that sensitivity will vary along the pixels in our camera.

Moreover, there are additional causes to produce a non uniform image, main ones being dust in the main mirror or lenses and vigneting - specially if a focal reducer is used in a cathadioptric scope.

To account for this effect, one or several flat fields are taken. Flat fields are just images of an uniformly iluminated background. Dawn or dusk horizon light is usually used for this task, although I have used light polution from a big city like Madrid with relative success.

Flat fields must be taken with an exposure time such that the image is midway between complete darkness and complete saturation.

The same as the dark frames, flat fields noise increases raw image noise. Again, the best solution is to average several flat fields to get a master flat field. CADET allows to specify several flat images during calibration. They will be automatically averaged.

Base background level in flat fields:

Flat fields are affected by thermal and readout noise the same way as a normal image. To compensate for this, one or several dark frames for flat fields are taken. Exposure time must be the same as the exposure time used for flat fields.

Of course, it is better to average several dark frames. CADET allows to specify several dark frames for flat fields during calibration. They will be automatically averaged.

Dark frames and flat fields are calibration images. To use them with CADET they must be taken following above guidelines.

There are several additional ways to acomplish calibration, for instance taking bias frames. Bias frames are just dark frames with an exposure time as short as possible and accounts for readout noise. This way we can somewhat use dark frames with a standard exposure time, different from that used in the astronomical image. Basic principles are the same as before, however in my opinion this technique is a bit more complex.

Deffective pixels:

Last, due to CCD technology not being perfect, several pixels may exist that saturate prematurely or don't produce any signal at all. Those "hot and cold" pixels are eliminated by means of a median filter. A median filter is a statistical operation that compares the pixel value with the neighbors. Pixels are corrected if their value differs too much from the median value of neighbor pixels.

Background luminosity:

Specially in big cities - or near them - you will get additional background luminosity in your images. This is due to light polution. You can account for this background luminosity in two ways:

First, the background luminosity can be substracted from the image to eliminate the effects of light polution. This is optionally done in a separate step.

The same visual effects may be achieved by adjusting the background and range, however this way the pixel values will not be modified, only the way the image is displayed changes. By doing so, the deconvolution operation will not be as effective as possible.

Additional considerations:

Several images may be averaged to account for short a exposure time. Of course, it is better to use a long exposure time, however averaging images must not be discarded. By using a shorter exposure time you can take unguided images (depending on the mount you are using), prevent image saturation and blooming.

When calibrating, CADET will automatically align the images with no user intervention at all. This is called image registration. Registering an image is shifting it so that all images properly align. Forget marking a reference star and this kind of stuff. This is not necessary with CADET. In any case, registering on a reference star and even not registering at all are also available options.

CADET will allow you to specify several images to be averaged when calibrating. It must be noted that CADET works internally using floating point numbers, so there is no real difference between adding and averaging images. Maximum accuracy may be achieved when saving the file by using the Optimize resolution option in the File save dialog.

For a nice explanation of image calibration and astronomical image processing in general, visit  Al Kelly's CCD Astrophotography Page .