Abstract: A system and method of visual monitoring of a work implement (e.g., a welding torch) while a task is being performed (e.g., forming a welding joint) to train workers (e.g., apprentices, inexperienced workers) in proper welding technique, for example) and/or to evaluate the worker's use of a particular work implement (e.g., to determine if the welding torch was held in a desired relationship to the items being welded together, determine if the welding torch formed the joint at the current speed, etc.). In general, one or more cameras may acquire images of a target secured to and/or formed on the work implement. The images may be analyzed to provide feedback to the user, to be evaluated for weld integrity purposes; and/or may be used to compare the performance of a task (e.g., forming a welding joint) with a database of one or more profiles made by experienced and/or expert craftsmen.

Abstract: Disclosed herein are embodiments and methods of a visual guidance and recognition system requiring no calibration. One embodiment of the system comprises a servo actuated manipulator configured to perform a function, a camera mounted on the face plate of the manipulator, and a recognition controller configured to acquire a two dimensional image of the work piece. The manipulator controller is configured to receive and store the face plate position at a distance “A” between the reference work piece and the manipulator along an axis of the reference work piece when the reference work piece is in the camera's region of interest. The recognition controller is configured to learn the work piece from the image and the distance “A”. During operation, a work piece is recognized with the system, and the manipulator is accurately positioned with respect to the work piece so that the manipulator can accurately perform its function.

Abstract: A method for visual recognition of an object in an electronic image includes extracting unique points of an object to be learned and/or a target object. The unique points are obtained by cross-correlating the image with a structure. Generally, the structure and/or the size of the structure may vary to detect extremum information associated with the learned object and/or target object. An icon corresponding to each of the unique points is extracted. The size of the icon corresponds to the scale of the unique point. After extraction of the various icons, an object becomes a collection of icons. Each of these icons is un-rotated and normalized or resized to a constant size so it can be compared with other icons. One of the unique properties of these icons is their stability over scale and angle. Thus, this invention allows the recognition of an image(s) or object(s) from large number of trained images or objects very quickly.