Poisson noise image restoration method based on variational regularization

In image processing problems, when the photon-counting imaging technology is used to obtain the target image, it is usually interfered with by Poisson noise, which causes the problem of image degradation and reduces the resolution of the image. The integer-fractional-order total variational regularization model proposed in this paper not only considers the relationship between the adjacent pixels of the image but also establishes a connection with the pixels farther away. Therefore, it has strong adaptability to remove noise in the image. In addition, by introducing auxiliary variables, an Alternating Direction Method of Multipliers (ADMM) algorithm for the I-FOTV model is deduced, which solves the constrained optimization problem of the I-FOTV model. Through numerical simulation experiments, the results show that the image restored by the I-FOTV model proposed in this paper not only has a certain improvement in visual quality but also improves the peak-signal-to-noise ratio (PSNR) by 0.18 dB–2 dB.