Abstract
Mapping the relativistic jets emanating from AGN requires the use of a deconvolution algorithm to account for the effects of missing baseline spacings. The CLEAN algorithm is the most commonly used algorithm in VLBI imaging today and is suitable for imaging polarisation data. The Maximum Entropy Method (MEM) is presented as an alternative with some advantages over the CLEAN algorithm, including better spatial resolution and a more rigorous and unbiased approach to deconvolution. We have developed a MEM code suitable for deconvolving VLBI polarisation data. Monte Carlo simulations investigating the performance of CLEAN and the MEM code on a variety of source types are being carried out. Real polarisation (VLBA) data taken at multiple wavelengths have also been deconvolved using MEM, and several of the resulting polarisation and Faraday rotation maps are presented and discussed.
Highlights
Mapping the relativistic jets emanating from AGN requires the use of a deconvolution algorithm to account for the effects of missing baseline spacings
Where H is the entropy of a model map of the source, χ2 is a measure of the difference between the model and the observed visibilities, α and β are the Lagrangian optimisation parameters and other conditions can be included to represent additional constraints, such as the positivity of the intensity in the model map
The Gull and Skilling entropy is maximum for an unpolarised source that is identical to the bias map and this is the source that Maximum Entropy Method (MEM) will produce in the absence of any data that forces it to make a more complicated model
Summary
The Maximum Entropy Method (MEM) is an alternative deconvolution method to the CLEAN algorithm. By iteratively maximising J in Eqn (1), the MEM method develops a model of the source which maximises the Gull and Skilling entropy of the model (the model has lowest possible polarisation, and looks as much like the bias map as the data allows), while reproducing the observed data to within noise levels. This results in a balance between entropy (representing the effects of unsampled visibilities and thermal noise) and fidelity to the observed data. Monte Carlo simulations are being performed to test how well it performs on model sources (where the details of the source are known at sub-Nyquist resolutions)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
