Multilabel Remote Sensing Image Classification with Capsule Networks

As a result of the developments in remote sensing technologies, the classification of remote sensing images according to user needs has gained great attention in recent years. Deep learning techniques are also known to increase the classification performance of remote sensing. In this study, image classification is made with capsule networks, which are deep artificial neural network model, on the multilabel datasets -Ankara Hyperspectral Image Archive and UC Merced Land Use. The performance of the classification is measured with various performance metrics.