A New Algorithm for Speckle Suppression using Mathematical Morphology and Adaptive Weighted Technique

In this paper, a new morphological image-cleaning algorithm that preserves thin features while removing speckle noise is presented and analyzed. In this new algorithm, the multi-scale top-hat transformation and bottom-hat transformation are added into the conventional multi-scale morphological opening and closing filtering. It differs from previous morphological filters in which it uses residual images to extract and smooth the features. The coefficients of the multi-scale top-hat transformation and bottom-hat transformation are optimized by adaptive weighted technique. This paper also shows that the new filtering algorithm preserves thin features significantly and reduces the speckle index greatly. The performance of the new filtering algorithm is superior to that of the conventional filtering methods.