VISUAL is an object-oriented and declarative language for locating objects automatically. The language is designed according to the principle of closeness and the principle of object orientation. A VISUAL program describes the geometric relationships among components of an object. The inference engine of VISUAL opens the corresponding databases and N × N 2-D maps of the components, then applies chromatographic search for unification. The inference engine outputs a database and a 2-D map of the object which could be a component of another object to be recognized. Hence VISUAL is a tool to build up a hierarchical model-based vision system.
關聯:
Journal of Systems Architecture vol. 43, no. 1-5 pp.327-335
