Machine Learning with Remote Sensing Image Data Sets

Computer vision, as a part of machine learning, gains significant attention from researches nowadays. Aerial scene classification is prominent chapter computer vision with vast application: military, surveillance and security, environment monitoring, detection geospatial objects, etc. There are several publicly available remote sensing image datasets, which enable the deployment various aerial algorithms. In our article, we use transfer learning pre-trained deep Convolutional Neural Networks (CNN) within classification. networks utilized in research high-dimensional previously trained CNN on ImageNet dataset. Transfer can be performed through feature extraction or fine-tuning. We proposed two-stream method afterward handcrafted classifier. Fine-tuning was adaptive rates regularization label smoothing. The techniques were validated two datasets: WHU RS datasets AID Our obtained competitive results compared to state-of-the-art methods.