Efficient gan-based remote sensing image change detection under noise conditions

Abstract

Efficient GAN-based remote sensing image change detection model under noise conditions is studied in this research work. Based on the multi-scale segmentation remote sensing change detection, this research work proposes an optimal method for the remote sensing image change detection with stable features, which provides an active way to increase the accuracy of the change detection. The target object selected in this article is relatively flat, but most target objects are different in practical applications. Therefore, the change detection method based on stable feature point removal based on multi-scale change detection is closely related to the multi-scale change detection, so the result is not avoiding the presence of “salt and pepper” noise. The GANs are integrated to pre-process the images, and the de-noising work is enhanced for the high-resolution images. The experiment has proved its effectiveness.