Ideology and Terminology

Some fundamental concepts that are used by Aegean.

Images are considered to be a combination of three things:

• A background contribution which is on large angular scales. Usually positive, but could be negative.
• A noise contribution which is a combination of thermal noise, instrumental defects, calibration problems, and imaging artefacts.
• A signal contribution which is assumed to be a series of compact peaks that can be well characterised by one ore more elliptical Gaussian components.

The Aegean source finding algorithm does a lot of work on the source finding and characterisation, but does a rather quick and dirty characterisation of the background and noise (however this can be done using BANE).

Aegean uses the background and noise models to classify all pixels in an image as either significant or not significant. Significant pixels that are contiguous are grouped together to form Islands, and each Island is modelled as a series of Gaussian components.

The important distinction is that Aegean finds islands (hence the name) and then characterises them as components.

Aegean has three different models for things that it can find in an image:

• Island - A collection of contiguous pixels that are all above a given threshold. Assumed to contain one or more components.
• Component - Some part of and Island, modelled as an elliptical Gaussian.
• Priorized fit - Replace the source finding stage with an input catalogue and move right to the characterisation stage. There are various constraints that can be placed on the fitting that is done, see Priorized Fitting