Mueller matrix based classification of polarimetric SAR data

A modified version of the unsupervised classification algorithm is proposed. In this method, multiple pixels are considered for classifying each pixel in the image during the classification procedure. First, the scattering matrix is used to calculate the corresponding Mueller matrices for each pixel. The Mueller matrices of all pixels in a window are then calculated and used to make a decision as described in J. J. van Zyl (1989). If all pixels (Mueller matrices) or more than N/2 pixels (more than 50 per cent of pixels, where N is the number of pixels in a window) in a window are classified into one category, then the pixel of interest (the Mueller matrix of the center pixel within the moving window) will be classified into the same category. In addition, one criterion is added for diffuse scattering. If the pixel properties are completely random, this means that, if fewer than N/√N (or equal N/√N) pixel properties in a moving window are classified into the categories (if odd number of reflections =<N/√N, even number of reflections =<N/√N, and diffuse scattering=<N/√N), the pixel property of interest will still be placed in the diffuse scattering category. Note that all three scattering properties occur in a small area. Therefore, the pixel of interest will be classified into the diffuse scattering category. If there is no significant pixel property belonging to the categories mentioned above in a moving window, then the pixel property of interest is unclassified. The window is then moved, and this process is repeated