When the topic of imaging comes up in aperture synthesis we are usually referring to the process of transforming visibilities into a spatial domain representation (image) of the sky an the process of deconvolving the array point spread function (PSF) from the measured image to attempt to recover a 'true' sky image. As we have seen in the previous chapter, the resulting images are affected by the the array configuration and observation parameters (bandwidth, time length, pointing declination, etc.). These parameters are known and the resulting array PSF is independent of the sky, so the (reasonable) goal of deconvolution is to separate the effects of the instrument from the sky.
Colloquially, one will hear someone talk about creating a 'clean' image from a 'dirty' image. This means that a deconvolution algorithm, typically a variant of the CLEAN algorithm, is used to construct a more 'complete' image based on our knowledge of the array PSF and what types of sources we expect to see in the sky.
To understand the motivation for the types of deconvolution algorithms used in radio interferometry we will first discuss how sources are typically represented in a sky model. We will then build up a basic form of the CLEAN algorithms, and increase the complexity of these methods. Then, we will cover the issue of when to halt deconvolution, how image quality is described, and how a final image is constructed. Finally, we will discuss automated source detection for extracting sky source information.
Next: 6.1 Sky Models
In [ ]: