Processing color and complex data using mathematical morphology

Abstract

Mathematical morphology has been used extensively to process binary and grayscale images. The use of mathematical morphology to process color images, however, is quite limited due to technical problems with ordering vector quantities. This paper describes three morphological techniques for processing color images that have appeared in the literature, in addition to presenting a new technique based on vector projections. The usefulness of these techniques is then illustrated on the task of extracting an object from a color image. Techniques for morphologically processing color images can be also applied to images and signals containing complex data.