Compressed sensing Fourier single pixel imaging algorithm based on joint discrete gradient and non-local self-similarity priors

Abstract

Fourier single-pixel imaging (FSI) has been proven capable of acquiring excellent image quality when is sampled the fully information in Fourier domain. However, when the sampling measurements is limited, image reconstruction by applying inverse Fourier transform algorithm would result in the severe ringing effect and the loss of image details. In this report, we propose a new algorithm for FSI reconstruction based on compressed sensing theory, which utilizes joint discrete gradient and non-local self-similarity priors, thus substantially using the prior knowledge of natural images in reconstruction process. Both the results of computational simulations and experiments demonstrate the efficacy of the proposed algorithm.