A Novel Approach for Scene Classification in Remote Sensing using Deep Learning

Scene classification in remote sensing plays an important role in various applications like disaster management, traffic control, etc., and has received remarkable attention. Researchers have made significant efforts to develop various datasets like UC Merced, RSSCN7, etc., and present a variety of approaches for scene classification, such as CNN, Transfer learning, etc., in remote sensing images. In this research paper, various scene classification approaches are discussed, and a novel scene classification approach transfer learning, using a pre-trained neural network on the satellite images for doing Scene-level classification, is proposed.