Experiencias en la implementación de las operaciones morfológicas de erosión y dilatación para imágenes binarias empleando vecindades adaptativas

mathematical morphology is a powerful tool in image processing. it is asserted that the spatially-variant morphology produces better results than the invariant one in many cases. there are several spatially-variant morphologies and they differ in the form of assigning the structuring element to each pixel. one of these is the adaptive neighborhood morphology, which uses the concepts of criterion mapping and homogeneity tolerance. these concepts allow that the structuring element assigned to each pixel is adjusted to the characteristics of the pixel neighborhood and vary in form and size. the criterion mapping assigns to each pixel a real value that can represent: luminance, contrast or curvature among others. in this paper the operations of erosion and dilation of the adaptive neighborhood morphology are implemented using matlab. the images are processed with the matlab conventional operations and with the operations implemented in this paper to observe similarities and differences.