Synthetic Aperture Radar Image Processing using the Supervised Textural-Neural Network Classification Algorithm

Abstract

Synthetic Aperture Radar (SAR) satellite images have proven to be a successful tool for identifying oil slicks. Natural oil seeps can be detected as elongated, radar-dark slicks in SAR images. Use of SAR images for seep detection is enhanced by a Texture Classifying Neural Network Algorithm (TCNNA), which delineates areas where layers of floating oil suppress Bragg scattering. The effect is strongly influenced by wind strength and sea state. A multi orientation Leung-Malik filter bank [1] is used to identify slick shapes under projection of edges. By integrating ancillary data consisting of the incidence angle, descriptors of texture and environmental variables, considerable accuracy were added to the classification ability to discriminate false targets from oil slicks and lookalike pixels. The reliability of the TCNNA is measured after processing 71 images containing oil slicks.

DOI: 10.1109/IGARSS.2008.4779960

Extracted Key Phrases

3 Figures and Tables

Cite this paper

@inproceedings{GarciaPineda2008SyntheticAR, title={Synthetic Aperture Radar Image Processing using the Supervised Textural-Neural Network Classification Algorithm}, author={Oscar Garcia-Pineda and Ian R. MacDonald and Beate Zimmer}, booktitle={IGARSS}, year={2008} }