Finding half boundaries and junctions in images

摘要

The theoretical advantages of using image junctions and corners in higher vision algorithms are well understood. However, due to inadequacies in current edge finding techniques the full potential of exploiting these constraints in practice has yet to be demonstrated. This paper describes a new technique for finding boundaries and junctions, the half boundary finding algorithm, which uses a combination of morphological feature detection and a novel boundary following procedure to overcome the criticisms of previous edge finders. Results are presented on real images, and both the limitations of the current algorithm and some enhancements are discussed.