Model-based despeckling and information extraction from SAR images

Basic textures as they appear, especially in high resolution SAR images, ar e affected by multiplicative speckle noise and should be preserved by despe ckling algorithms. Sharp edges between different regions and strong scatter ers also must be preserved. To despeckle images, we use a maximum a posteri ori (MAP) estimation of the cross section, choosing between different prior models. The proposed approach uses a Gauss Markov random held (GMRF) model for textured areas and allows an adaptive neighborhood system for edge pre servation between uniform areas. In order to obtain the best possible textu re reconstruction, an expectation maximization algorithm is used to estimat e the texture parameters that provide the highest evidence. Borders between homogeneous areas are detected with a stochastic region-growing algorithm, locally determining the neighborhood system of the Gauss Markov prior. Smo othed strong scatterers are found in the ratio image of the data and the fi ltering result and are replaced in the image. In this way, texture, edges b etween homogeneous regions, and strong scatterers are well reconstructed an d preserved. Additionally, the estimated model parameters can be used for f urther image interpretation methods.