Erschienen in:IEEE 875-878 (2019). doi:10.1109/IGARSS.2019.8899831 ; [Proceedings] - IEEE, 2019. - ISBN 978-1-5386-9154-0 ; [Proceedings] - IEEE, 2019. - ISBN 978-1-5386-9154-0 IGARSS 2019 - 2019 IEEE International Geoscience and Remote Sensing Symposium, Yokohama, Japan, 2019-07-28 - 2019-08-02
Sprache:
Englisch
DOI:
https://doi.org/10.1109/IGARSS.2019.8899831
Entstehung:
Anmerkungen:
Diese Datenquelle enthält auch Bestandsnachweise, die nicht zu einem Volltext führen.
Beschreibung:
The classification of land-cover classes in remote sensing images can suit a variety of interdisciplinary applications suchas the interpretation of natural and man-made processes onthe Earth surface. The Convolutional Support Vector Machine (CSVM) network was recently proposed as binary classifier for the detection of objects in Unmanned Aerial Vehicle (UAV) images. The training phase of the CSVM isbased on convolutional layers that learn the kernel weightsvia a set of linear Support Vector Machines (SVMs). Thispaper proposes the Multi-scale Convolutional Support VectorMachine (MCSVM) network, that is an ensemble of CSVMclassifiers which process patches of different spatial sizes andcan deal with multi-class classification problems. The experiments are carried out on the EuroSAT Sentinel-2 dataset andthe results are compared to the one obtained with recent transfer learning approaches based on pre-trained ConvolutionalNeural Networks (CNNs).