Methods in the spatial deep learning: current status and future direction

A deep neural network (DNN), evolved from a traditional artificial network, has been seamlessly adapted for the spatial data domain over years. Deep learning (DL) widely applied number of applications and variety thematic domains. This article reports on systematic review methods in major DNN with remote sensing published between 2010 2020 aiming to understand application area, framework model development prospect DL analysis. It found that image fusion, change detection, scene classification, segmentation, feature detection are most commonly used areas. Based publication these areas, generic devised guide using based followed past. Finally, recent trends prospects terms data, method, discussed. The finds while DL-based approaches have potential unfold hidden information, they face challenges selecting appropriate methods, parameterizations which may hinder performance. increasing trend is expected leverage its strength at optimum research frontiers.