Chihhsiong S. Shih
Department of Mechanical Engineering, Aeronautical Engineering and Mechanics, Rensselaer Polytechnic Institute, Troy, N.Y. 12180
Lester A. Gerhardt
Prof. of Electrical, Computer, and Systems Engineering, Prof, of Computer Science, and Associate Dean of Engineering Rensselaer Polytechnic Institute, Troy, N.Y. 12180
The best view based on visible edge length can usually not only minimize the number of rotations needed to inspect an entire object, but can also provide maximum amount of information for integration with non-uniform sampling techniques for a variety of parts and manufacturing processes. Optimizing view planning to achieve the goal of minimum occlusion and minimum rotation to completely inspect an object, not only satisfies the goal of optimizing view planning, but also automatically maximizes the number of surfaces to which adaptive sampling can be most fully applied. This new integration approach of first viewing then sampling establishes that the optimization criterion of maximum edge length is more robust than anticipated in it's original role for view planning alone. Experiments results strongly support the theory developed.