Ship-Iceberg Discrimination with Convolutional Neural Networks in High Resolution SAR Images

Konferenz: EUSAR 2016 - 11th European Conference on Synthetic Aperture Radar
06.06.2016 - 09.06.2016 in Hamburg, Germany

Tagungsband: EUSAR 2016

Seiten: 4Sprache: EnglischTyp: PDF

Persönliche VDE-Mitglieder erhalten auf diesen Artikel 10% Rabatt

Bentes, Carlos; Frost, Anja; Velotto, Domenico; Tings, Bjoern (German Aerospace Center (DLR), Germany)

The application of Synthetic Aperture Radar (SAR) for ship and iceberg monitoring is important to promote maritime safety in Arctic waters. Although the detection of ships and icebergs in SAR images is well established using adaptive threshold techniques, the discrimination between the two target classes still represents a challenge for operational scenarios. This paper proposes the application of Convolutional Neural Networks (CNN) for ship-iceberg discrimination in high resolution TerraSAR-X StripMap images. The CNN model is compared with a Support Vector Machine (SVM), and the final results indicate a superior classification performance of the proposed method.