Separating Point Source Contaminants at the Visibility Level

Bright point source contaminants are an ever-present problem in astronomy. This issue is especially pronounced for radio interferometric telescopes. The calibration process for these telescopes is never perfect, and consequently even the smallest calibration errors on bright sources can cause interferometric artefacts that obscure weaker sources within the same field of view. This effect can strongly limit the dynamic range of images produced by the next generation of radio interferometric telescopes. Peeling, or subtracting modelled visibilities corresponding to the bright source in visibility space, is often employed to address this problem. However, this does not always remove all interferometric artefacts associated with the bright contaminating source. We present a visibility‑based method, Bluebild Imaging Plus Plus (BIPP), for the removal of contaminating point sources. BIPP is an interferometric imager developed to reconstruct the sky intensity matrix directly on the celestial sphere. It uses functional principal component analysis (fPCA), based on the intensity of sources, to separate input visibilities into eigen‑visibilities. These eigen‑visibilities are then reconstructed into eigenimages, which can be optimally combined into energy levels for source separation. Since the fPCA is performed directly on the visibilities, all energy levels contain only the artefacts from the sources present within each energy level. We use this method to separate a contaminating bright source at two different frequencies for simulated observations generated from a realistic sky model. In doing so, we demonstrate the potential of BIPP as a method for bright source separation. We also compare the BIPP output with WSClean, to analyse and contrast the results produced by the two algorithms.