From Artifact Removal to Super-Resolution

Deep-learning-based super-resolution (SR) methods have been extensively studied and have achieved significant performance with deep convolutional neural networks. However, the results still suffer from the ringing effect, especially in satellite image SR tasks, due to the loss of image details in the satellite degradation process. In this article, we build a novel satellite SR framework by decomposing a high-resolution image into three components, i.e., low-resolution (LR), artifact, and high-frequency information. Specifically, we propose an artifact removal network with a self-adaption difference convolution (SDC) to fully exploit the structure prior in the LR image and predict the artifact map. Considering that the artifact map and the high-frequency map share a similar pattern, we introduce the supervised structure correction (SSC) block that establishes a bridge between the high-frequency generation process and the artifact removal process. Experimental results on satellite images demonstrate that the proposed method owns an improved tradeoff between the performance and the computational cost compared to existing state-of-the-art satellite and natural SR methods. The source code is available at https://github.com/jiaming-wang/ARSRN.

I. Introduction

As an emerging means of Earth observation, satellite platforms have drawn increasing attention in both military and civil fields, such as surface feature segmentation [1], [2], environmental monitoring [3], [4], satellite mapping [5], and object detection [6], [7]. Limited by the under-sampling effect of imaging sensors and complexity in the degradation process, captured satellite images (especially those from geostationary satellites) may fail to meet the demand of many applications that require high-precision measurement. Image super-resolution (SR) is a software-level technology that aims to increase the spatial resolution of low-resolution (LR) images without introducing additional hardware costs. With years of development, satellite image SR has become an essential step for many real-time and high-precision applications.