Superresolution Interferometric Imaging with Sparse Modeling Using Total Squared Variation --- Application to Imaging the Black Hole Shadow

Kazuki Kuramochi,Kazunori Akiyama,Shiro Ikeda,Fumie Tazaki,Vincent L. Fish,Hung-Yi Pu,Keiichi Asada,Mareki Honma

DOI: https://doi.org/10.3847/1538-4357/aab6b5

2018-03-13

Abstract:We propose a new superresolution imaging technique for interferometry using sparse modeling, utilizing two regularization terms: the $\ell_1$-norm and a new function named Total Squared Variation (TSV) of the brightness distribution. TSV is an edge-smoothing variant of Total Variation (TV), leading to reducing the sum of squared gradients. First, we demonstrate that our technique may achieve super-resolution of $\sim 30$% compared to the traditional CLEAN beam size using synthetic observations of two point sources. Second, we present simulated observations of three physically motivated static models of Sgr A* with the Event Horizon Telescope (EHT) to show the performance of proposed techniques in greater detail. We find that $\ell_1$+TSV regularization outperforms $\ell_1$+TV regularization with the popular isotropic TV term and the Cotton-Schwab CLEAN algorithm, demonstrating that TSV is well-matched to the expected physical properties of the astronomical images, which are often nebulous. Remarkably, in both the image and gradient domains, the optimal beam size minimizing root-mean-squared errors is $\lesssim 10$% of the traditional CLEAN beam size for $\ell_1$+TSV regularization, and non-convolved reconstructed images have smaller errors than beam-convolved reconstructed images. This indicates that the traditional post-processing technique of Gaussian convolution in interferometric imaging may not be required for the $\ell_1$+TSV regularization. We also propose a feature extraction method to detect circular features from the image of a black hole shadow with the circle Hough transform (CHT) and use it to evaluate the performance of the image reconstruction. With our imaging technique and the CHT, the EHT can constrain the radius of the black hole shadow with an accuracy of $\sim 10-20$% in present simulations for Sgr A*.

Instrumentation and Methods for Astrophysics,High Energy Astrophysical Phenomena