An improved remote sensing image blind deblurring algorithm

Abstract

The remote sensing image acquisition may become blurred and the texture is unclear due to the remote sensor s jitter, defocusing and other factors. Therefore, deblurring remote sensing images is a major problem in remote sensing research. In order to improve the quality of remote sensing images, an improved blind deblurring algorithm for remote sensing images based on dark channel prior is proposed. First, the blur kernel estimation based on the dark channel prior is performed, and then the image deconvolution is performed using the estimated blur kernel to achieve clear image restoration. A joint bilateral filter is introduced to maintain the image detail and enhance edges for the image ringing effect produced during the original clear image restoration process. The experimental results show that the peak signal-to-noise ratio (PSNR) of the objective evaluation index of this method is improved by about 1.5dB, and the structural similarity (SSIM) is improved by about 0.04. Compared with the existing image deblurring algorithm, this method can estimate the blur information more accurately, so that the restored result image can achieve the effect of maintaining the edge to remove the ringing.