Learning Approaches for Remote Sensing Image Classification

The latest generation of aerial- and satellite-based imaging sensors acquires huge volumesof Earth's images with high spatial, spectral and temporal resolution, which open the doorto a large range of important applications, such as the monitoring of natural disasters, theplanning of urban environments and precision agriculture. In order to fully exploit thepotential offered by these sensors, there is a need to develop accurate and time-efficientmathematical models and algorithms for spectral-spatial analysis of the recorded high-resolution data.The main goal of my research is to develop learning approaches, which would helpto automatically interpret, or classify, remote sensing images. This manuscript presentsseveral strategies I have explored for this purpose, varying from the use of strong shapepriors to detect objects, regularization of classification probabilities on the image graphs,and up to the use of convolutional neural network models capable to learn deep hierar-chical contextual features.The experimental results on diverse benchmarks of images and image time series showthe competitiveness of the developed methods when compared to the state-of-the-art ap-proaches. In particular, we have recently created large-scale classification benchmarkof aerial images and have demonstrated that the modern deep learning-based methodssucceed in generalizing to the dissimilar urban settlements around the Earth. This opensnew exciting perspectives towards designing systems which would be able to automati-cally update world-scale maps from remote sensing data.