ABSTRACT
Image processing methods can be broadly grouped under six principal types: measurement-space-guided spatial clustering, single-linkage region growing; variant of the preceding which takes into consideration not only the value of the pixel, but also its neighbourhood and is known as hybrid-linkage region growing spatial clustering; centroid-linkage region growing; and split and merge methods which use contour reduction. Texture of an image is defined as a combination of textural elements. A textural element is a group of resolution elements which have the same value of and which are connected. Structure of an image is a combination of structural elements. A structural element is defined by repeated relationships existing between the textural elements. Hence a processing language has developed based on mathematical, statistical and computer languages. Segmentation is the action of dividing an image into classes without knowing what these classes thematically represent. The ultrametric parameter corresponds to choice made for comparing a pixel with group and group with another group.
