Research on accuracy optimization of coastal zone remote sensing images based on deep learning

The coastal zone, as the boundary between the ocean and the land, is one of the most intensively active areas on Earth. However, its complex geographical environment, such as suspended sediments, tortuous coastal zones, and extensive tidal flats, has brought sudden challenges to the spatiotemporal dynamic analysis of the coastal zone. Accurately extracting coastal information monitoring and its spatiotemporal changes has become an important and urgent task. To address this problem, we designed multiple scenarios and combined Deep Learning methods to improve the accuracy of remote sensing images in coastal zone to obtain more detailed ground object information. Experimental results show that our method can achieve up to 42.60 and 0.9858 in structural similarity (PSNR) and peak signal-to-noise ratio (SSIM), which proves that our method can effectively improve the accuracy of remote sensing images.