Multi-sale morphological filtering method for preserving the details of images

A multi-sale morphological filtering method for preserving the details of images (MMFPD) was applied by adding the multi-scale top-hat transformation and bottom-hat transformation to the conventional multi-sale morphological opening and closing filtering. The two added transforms were used to extract and smooth the features which are smaller than the current scale. It was found that the smaller features have greater possibility to contain noise particles. Accordingly, the coefficients of top-hat transformation and bottom-hat transformation were modified. Simulation results on the standard gray-level images show that the method can effectively remove noise and preserve the details of images completely, and demonstrate better performance than the conventional filtering methods.