Weighted Stacking of Radio Images Affected by Noise and Interfering Signals

Abstract

We implement an algorithm based on the weighted stacking of astronomical images that can combine different observations of the same region of the sky removing the interfering signals. We develop a C++ code that takes as input a set of spectral cubes and computes the local weights of the intensity for each pixel of every channel. The weights are calculated as the inverse variance of the nearby pixels and are used to compute the weighted merge of the input files. Astronomical sources, present in all cubes, are preserved by the weighted average. However, interfering signals, present in specific cubes and a certain frequency range, are down-weighted in the average and removed from the output spectral cube. We present the results obtained by analyzing simulated spectral cubes containing astronomical sources, noise, and a random set of interferences of different intensities and spectral occupations. The performance of the algorithm is evaluated by comparing the output result with the input sky-model image. Finally, we present the results obtained by applying the method to a set of real data consisting of observations of the Andromeda galaxy, Messier 31 (M31), at 6.6 GHz obtained with the Sardinia Radio Telescope.