Metal surface inspection using image processing techniques

摘要: The feasibility of applying image processing techniques to metal surface inspection is demonstrated. Two methods for are described. In the first method, reflective power and normal related by a random scattering model. profile can then be computed from normal. second method applies pattern recognition classify surfaces into classes different roughness. Methods feature extraction classification have been tested experimentally performances types classifier compared. A two-level tree using nonparametric linear classifiers at each node gives better than 90% correct on testing set.