Geometric search technique for surface roughness evaluation using machine vision

The importance of geometric search approach using computer vision techniques to inspect surface roughness has been reported in this paper. The surface images of the components are first acquired using a machine vision system and then the features of the surface images are enhanced. A new geometric search technique is applied to enhance the quality of images for quantification of surfaces, which can cope up with the process variations and adverse conditions in a CIM environment, and then the performance of these images were compared with the application of other commonly used filters. The two methods were used for evaluating the surface finish of components generated using shaping process viz. Regression analysis and Polynomial networks. The surface finish values obtained using these two methods after applying geometric search technique for images are compared with the estimated roughness values using images without applying geometric search. The estimated roughness values using machine vision images and that measured using stylus approach were finally compared and analyzed in this work.