A fast sequential algorithm for pixel grey level ranking in adaptive rank filters applied to digital images

Abstract

A fast sequential algorithm is presented for ranking the pixel grey levels which occur in the window of a digital filter as it is passed over a two-dimensional image. The algorithm is based on continually updating the frequency distribution of the grey levels in the window and ranking them in a one-dimensional array. The proposed algorithm is applied to images containing 256 grey levels and is faster than conventional sorting methods when more than a single ranked parameter is required.