Remote sensing image enhancement and restoration method and network based on reference image without high resolution

Abstract

The invention discloses a remote sensing image enhancement and restoration method and network based on a reference image without high resolution, and belongs to the technical field of remote sensing image super-resolution enhancement. A first stage is a simulated degradation stage, a simulated degradation network based on a cyclic generation network is constructed, and a sub-network learns degradation distribution of a real remote sensing image by using the low-quality real degraded remote sensing image; a simulated remote sensing degradation image is generated under the same content from the high-resolution non-degradation remote sensing image through a simulated degradation network; and a second stage is a remote sensing enhancement stage, a remote sensing enhancement network is built, and end-to-end training is performed by using the high-resolution remote sensing image and the corresponding remote sensing degraded image pair. In order to enable the network to fully learn a remote sensing degraded image enhancement model, a dual-input enhanced generative adversarial network is innovatively used to realize a super-resolution stage network; and finally, the learned remote sensing image enhanced generative adversarial network is applied to any remote sensing degraded image to generate an enhanced reconstructed remote sensing image corresponding to the remote sensing degraded image.

Description

technical field [0001] The invention belongs to the technical field of super-resolution enhancement of remote sensing images, and more specifically relates to super-resolution enhancement processing of low-resolution remote sensing images under the supervision of no paired high-resolution reference images. Background technique [0002] In the field of remote sensing imaging, people usually further analyze and utilize the acquired remote sensing data, such as military target monitoring, environmental early warning and disaster research. High-resolution remote sensing images have finer texture information, which provides a more favorable basis for further applications and research related to remote sensing. [0003] Due to hardware degradation of remote sensing imaging equipment and complex environment and other factors, the spatial resolution of remote sensing images collected by them is not enough, resulting in low image quality, which seriously restricts the application of ...

AI Technical Summary

Problems solved by technology

However, most methods need to collect real high-resolution remote sensing images corresponding to low-quality remote sensing images, which is very difficult

Therefore, the lack of real data sets and no paired training data seriously restrict the processing effect and generalization ability of these methods.

Images