De-speckling of Synthetic Aperture Radar Image using Hybrid Filtering Techniques

For SAR de-speckling, two hybrid filtering algorithms based on the bilateral filter (BF), edge-based adaptive mean filter (EAM) and non-local mean filter (NLM) are presented in this work. The non-local filtering principle is used to choose the neighborhood that will be used for the estimation of the filtered pixels. The suggested system performances are evaluated and contrasted with current SAR de-speckling techniques. According to the results, the proposed hybrid filters outperformed the Frost and Median filters in terms of reducing speckles and preserving edge detail.