Incorporating texture information into polarimetric radar classification using neural networks

Most of the recent research on polarimetric SAR classification focused on pixel-based techniques using the co-variance matrix representation. Since multiple channels are inherently provided in polarimetric data, conventional techniques for increasing the dimensionality of the observation, such as texture feature extraction, were ignored. In this paper, we have demonstrated the potential of texture classification through gray level co-occurrence probabilities (GLCP), and proposed an unsupervised scheme using the self-organizing map (SOM) neural network. The increase in separability of the feature space is shown via the Fisher criterion and also verified by increased classification performance. Compared to the Wishart classifier, promising classification results are obtained from the Flevoland data set.