Robust Blind Deblurring Under Stripe Noise for Remote Sensing Images

The blind image deblurring methods have achieved great progress for Gaussian random noise. A few works have paid attention to image deblurring under structural noise, which is a very common degradation in multidetector imaging systems. This article considers the practical yet challenging problem of blind deblurring in the presence of the line-pattern stripe noise for remote sensing images. To overcome this issue, we explicitly formulate the structural noise into a novel and robust blind image deblurring framework. We observe that the structural line-pattern stripe noise would deteriorate both the kernel estimation and nonblind deblurring and propose a three-stage restoration framework to progressively estimate the blur kernel and clean image. Specifically, we first estimate an intermediate blur kernel by getting rid of the negative influence of the stripe noise in the unidirectional gradient domain. Next, a learning-based kernel refinement network is introduced to rectify the missing details of the inaccurate kernel. Finally, a low-rank decomposition-based nonblind deblurring model is proposed to simultaneously estimate the clean image and stripe noise. Experimental results on real and synthetic datasets demonstrate that the proposed robust blind image deblurring under stripe noise (RBDS) method outperforms the state-of-the-art blind deblurring methods.

Automated summary

This article explores the effect of structural noise on blind image deblurring in remote sensing images and proposes a three-stage restoration framework to progressively estimate the blur kernel and clean image. By eliminating the negative influence of stripe noise and introducing a learning-based kernel refinement network and a low-rank decomposition-based nonblind deblurring model, the proposed robust blind image deblurring under stripe noise (RBDS) method outperforms the state-of-the-art blind deblurring methods.