Chapter

Advanced Concepts for Intelligent Vision Systems

Volume 6915 of the series Lecture Notes in Computer Science pp 103-114

Fuzzy Cognitive Maps Applied to Synthetic Aperture Radar Image Classifications

  • Gonzalo PajaresAffiliated withCarnegie Mellon UniversityDpto. Ingeniería del Software e Inteligencia Artificial, Facultad de Informática, University Complutense of Madrid
  • , Javier Sánchez-LladóAffiliated withCarnegie Mellon UniversityRemote Sensing Laboratory (RSLab), Signal Theory and Communications Department, Universitat Politècnica de Catalunya (UPC)
  • , Carlos López-MartínezAffiliated withCarnegie Mellon UniversityRemote Sensing Laboratory (RSLab), Signal Theory and Communications Department, Universitat Politècnica de Catalunya (UPC)

* Final gross prices may vary according to local VAT.

Get Access

Abstract

This paper proposes a method based on Fuzzy Cognitive Maps (FCM) for improving the classification provided by the Wishart maximum-likelihood based approach in Synthetic Aperture Radar (SAR) images. FCM receives the classification results provided by the Wishart approach and creates a network of nodes associating a pixel to a node. The activation levels of these nodes define the degree of membeship of each pixel to each class. These activations levels are iteratively reinforced or punished based on the existing relations among each node and its neighbours and also taking into account the own node under consideration. Through a quality coefficient we measure the performance of the proposed approach with respect to the Wishart classifier.

Keywords

Fuzzy Cognitive Maps Wishart classifier Synthetic Aperture Radar (SAR) Polarimetric SAR (POLSAR) classification