Publisher's Synopsis
With the rapid growth in remote sensing imaging technology, vast amounts of remote sensing data are generated, which is significant for land-monitoring systems, and agriculture, etc., for Earth, Mars, etc. In recent decades, deep learning techniques have had a significant effect on remote sensing data processing, especially in image classification and semantic segmentation. However, several challenges still exist due to the limited annotations, the complexity of large-scale areas, and other specific problems, which make it more difficult in real-world applications. Therefore, novel deep neural networks combined with meta-learning, attention mechanisms, or other new transformer technologies need to be given more attention in remote sensing. It is also necessary to develop lightweight, explainable, and robust networks. Moreover, this Special Issue aims to develop state-of-the-art deep networks for more accurate remote sensing image classification and semantic segmentation, which also aims to achieve an efficient cross-domain performance through a lightweight network design.
