学术文献库

The desire for wide-field of view, large fractional bandwidth, high sensitivity, high spectral and temporal resolution has driven radio interferometry to the point of big data revolution where the data is represented in at least three dimensions with an axis for spectral windows, baselines, sources, etc; where each axis has its own set of sub-dimensions. The cost associated with storing and handling these data is very large, and therefore several techniques to compress interferometric data and/or speed up processing have been investigated. Unfortunately, averaging-based methods for visibility data compression are detrimental to the data fidelity, since the point spread function (PSF) is position-dependent, i.e. distorted and attenuated as a function of distance from the phase centre. The position dependence of the PSF becomes more severe, requiring more PSF computations for wide-field imaging. Deconvolution algorithms must take the distortion into account in the major and minor cycles to properly subtract the PSF and recover the fidelity of the image. This approach is expensive in computation since at each deconvolution iteration a distorted PSF must be computed. We present two algorithms that approximate these position-dependent PSFs with fewer computations. The first algorithm approximates the position-dependent PSFs in the $uv$-plane and the second algorithm approximates the position-dependent PSFs in the image-plane. The proposed algorithms are validated using simulated data from the MeerKAT telescope.