Super-Resolution for Remote Sensing Imagery via the Coupling of a Variational Model and Deep Learning

TL;DR

This paper introduces a novel super-resolution framework for remote sensing images that combines a learned gradient prior with non-local total variation, improving image quality by effectively preserving edges and structures.

Contribution

It proposes a gradient-guided multi-frame super-resolution method integrating deep learning-based and model-based priors for enhanced remote sensing image reconstruction.

Findings

Outperforms several state-of-the-art SR algorithms in quantitative metrics.

Produces visually pleasant high-resolution remote sensing images.

Effectively preserves edges and suppresses artifacts in reconstructed images.

Abstract

Image super-resolution (SR) is an effective way to enhance the spatial resolution and detail information of remote sensing images, to obtain a superior visual quality. As SR is severely ill-conditioned, effective image priors are necessary to regularize the solution space and generate the corresponding high-resolution (HR) image. In this paper, we propose a novel gradient-guided multi-frame super-resolution (MFSR) framework for remote sensing imagery reconstruction. The framework integrates a learned gradient prior as the regularization term into a model-based optimization method. Specifically, the local gradient regularization (LGR) prior is derived from the deep residual attention network (DRAN) through gradient profile transformation. The non-local total variation (NLTV) prior is characterized using the spatial structure similarity of the gradient patches with the maximum a posteriori (MAP) model. The modeled prior performs well in preserving edge smoothness and suppressing visual artifacts, while the learned prior is effective in enhancing sharp edges and recovering fine structures. By incorporating the two complementary priors into an adaptive norm based reconstruction framework, the mixed L1 and L2 regularization minimization problem is optimized to achieve the required HR remote sensing image. Extensive experimental results on remote sensing data demonstrate that the proposed method can produce visually pleasant images and is superior to several of the state-of-the-art SR algorithms in terms of the quantitative evaluation.