Super Resolution of Airplane Target in Remote Sensing Images via A Multi-Degradation Model

Feng Cai,Keyu Wu,Hecheng Jia,Feng Wang

DOI: https://doi.org/10.1109/ICAIT56197.2022.9862621

2022-01-01

Abstract:In this paper, we argue that the degradation model assumed by existing work on super-resolution of remote sensing images deviates from those in real-world images. To address the above problem, this article proposes a super-resolution reconstruction method for unpaired airplanes in remote sensing images, which consists of a model that simulates the multiple degradations in real-world scenarios and a super-resolution generative adversarial network that generates high-resolution images. Specifically, the novel degradation model can cover a wide range of degradations in natural scenes, which contains diverse factors of additive Gaussian noise, Poisson noise, Brown Gaussian noise, Gaussian blur, etc. Experiments are conducted on the airplane targets in the Gaofen challenge dataset, and the results demonstrate the superiority of our method compared to state-of-the-art methods, especially in raw remote sensing images with multiple noise and blur, which may be preferred in other practical satellite applications with harsh circumstances.