Computational Methods

Astronomical image processing applies a large variety of numerical methods to extract scientific results from observed data. The standard computational techniques used in this process are discussed with special emphasis on the problems and advantages associated with them when applied to astronomical data. It has not been the aim to give a full mathematical description of the methods; references to more detailed explanation of different techniques are given for further study. A general discussion of digital image processing can be found in e.g. Pratt (1978), Bijaoui (1981), and Rosenfeld and Kak (1982) while Bevington (1969) gives a good introduction to basic numerical methods.

The methods are presented in the order in which they are applied in a typical reduction sequence. A number of standard techniques are used at different stages of the reductions in which case they are treated only at the most relevant place. The general reduction sequence has been divided into three main parts. In Section 2.2 the transformation of raw observed data into intensity calibrated values is discussed while general image processing techniques are reviewed in Section 2.3. The evaluation of the resulting frames and the extraction of information from them are considered in Section 2.4 whereas Section 2.5 deals with the statistical analysis of the results. The terminology in this paper has been based of two dimensional image data although most of the techniques can equally well be applied to one dimensional data. Techniques which relate to special detectors or observational methods (e.g. speckle or radio observations) have not been considered.