A Generalized Morphological Filtering Approach

摘要

In order to decrease the output statistical bias of the open-closing and clos-opening filters, we present a new class of morphological filters called the generalized open-closing and generalized clos-opening filters in this paper. The generalized morphological filters (GMFs) improve the performances of traditional morphological filter because of using the different sized structuring elements. To sufficient understand the noise attenuation properties, we analyze the statistical output properties by using the expression of stack filters, and compare the filtering performances between GMFs and traditional morphological filters by computer simulation. Furthermore, we suggest a filtering algorithm by using the generalized morphological filters and the linear multiple structuring elements. The simulation results have shown that the algorithm has better performance in noise-suppressing and detail-preserving.