Complex-Valued Convolutional Neural Networks for Object Detection in PolSAR data

Conference: EUSAR 2010 - 8th European Conference on Synthetic Aperture Radar
06/07/2010 - 06/10/2010 at Aachen, Germany

Proceedings: EUSAR 2010

Pages: 4Language: englishTyp: PDF

Personal VDE Members are entitled to a 10% discount on this title

Hänsch, Ronny; Hellwich, Olaf (Computer Vision and Remote Sensing, Technische Universität Berlin, FR3-1, Franklinstr. 28/29, 10587 Berlin, Germany)

Detection methods for generic object categories, which are more sophisticated than pixel-wise classification, have been rarely introduced to polarimetric synthetic aperture radar (PolSAR) images. Despite the great success in other computer vision applications, the transfer to PolSAR data has been delayed due to the different statistical properties. This paper provides a first investigation of Complex-Valued Convolutional Neural Networks (CC-NN) for object recognition in PolSAR data. Although an architecture with only one single convolutional layer was used, the results are already superior to those obtained by a standard complex-valued neural network.