Object rcognition is an important issue in computer vision and robotics, where sensory information is processed with the pur0ose to devide the scene in objects represented by numerical features that discriminate one object lcass from the other. In many cases one sensory system, which can be a camera, doesn't suffice for object recognition and multiple sensors are required to resolve the classification proboem. WSith the introduction of multiple sensors, information integration of fusion becomes an issue. In this paper the most improtant frameworks for information fusion are compared for the case of object recongition. Furtyhermore, general aspects of designing a multi-sensor object recogition system are mentioned by outlining our system developed to recognize electronic components on printed cricuit boards.
展开▼