Image synthesis for the GeoSTAR array

A new approach to the GeoSTAR image inversion is presented below which solves several practical problems which are encountered with large interferometer arrays. The approach is a variation of the G-matrix solution that uses the measured spatial response of the array-e.g. as measured on the antenna range-as the basis for image synthesis. But rather than attempting inversions of the entire array-a process very sensitive to errors-a new approach is presented which corrects the visibility responses of the array piecewise. The technique starts by assuming that the response of a given visibility sample in a STAR array will deviate from that of an idealized model in a manner that can be characterized by just a few low-order spatial harmonics. This presumes that imperfections of the antenna elemental radiation patterns and of the interferometric fringe frequencies are dominated by errors which are largely contained within the physical area of the elemental antenna apertures and of their immediate surroundings-due to mechanical tolerances, mutual coupling, scattering, etc.. As such, it should be possible to correct anomalies in the measured response of a given visibility sample by combining it with its nearest neighbors on the U-V plane.