Techniques such as flat fields and dark frames are called data reduction. The term calibration is often used, but data reduction is a more accurate description of what goes on. Calibration is actually the process of comparing your values to a known accurate reference, such as calibrating the brightness of a star in your image to its known magnitude.
Bias frame
A bias is an exposure of the shortest possible duration, taken with the shutter closed (dark). Used when scaling dark frames, applying them to images with a different exposure time than the dark frame.
Dark frame
A dark frame is an exposure taken with the shutter closed. A dark is normally taken with the same exposure duration and cooling temperature as the light frame to which it will be applied. The purpose of a dark frame is to record the system noise of the camera. It is subtracted from a light frame to remove that system noise. For best results, take multiple dark frames and median combine them (getting a 'master dark frame'). This will reduce the amount of random noise added to the image by the dark frames. Although a dark frame is used to remove system noise, it can add it a small amount of random noise at the same time.
Flat-field frame
Just as a dark frame removes the camera system noise from an image, a flat-field frame removes optical sources of noise. This includes dust that casts shadows on the CCD chip; vignetting in the optical systems; unven lighting from internal reflections; etc. When imaging under very dark skies with a non-vignetted optical system, flat-field frames are less necessary. The brighter the sky, the more critical it is that you take high-quality flat-field frames. Unlike darks, flat fields are not subtracted from an image. The flat is scaled to the background level of the image, and then divided into the image to remove the effects of optical noise.
See articles : Using a flat-field.
Reduction
The process of removing system noise and errors from your images. This is typically accomplished by taking and applying bias, dark, and flat-field frames. <!http://www.brianmwalsh.com/AutoPhotReduction/AutoPhotReduction.htm>
