Remote Sensing Image Deblurring Based on Differential Lorentzian PSF Model

Image deblurring is a technique employed to reduce or eliminate degradation resulting from the effects of impulse response and atmospheric turbulence during remote sensing image acquisition. However, the issue of low recovery accuracy due to inadequate point spread function (PSF) matching is a significant challenge in current deblurring methods. In this letter, a remote sensing image deblurring method based on the differential Lorentzian PSF model is proposed. First, a differential Lorentzian PSF model is proposed, which is able to model the line spread function (LSF) of a real image as a series of pulsewidth variable Lorentzian functions and their differential function combinations. Subsequently, a model parameter estimation algorithm based on the improved zeroing filter is proposed. This algorithm is able to reconstruct the target parameters by utilizing the edge information of the actual image, with the objective of eliminating the matching error of the PSF model in the actual complex scene and improving the parameter estimation accuracy. Finally, a remote sensing image deblurring method is proposed by using the differential Lorenz point diffusion function model combined with the L-R algorithm. The results of the simulation experiments demonstrate that the method proposed in this letter is more effective than existing remote sensing image deblurring methods in terms of image recovery accuracy.