K. N. Joshi
B.T. Patil
Abstract
Machine vision systems have the potential of replacing traditional methods used for inspection of surface quality. Surface finish is an important characteristic from operational, ergonomic and aesthetic aspects of quality. Machine vision systems can effectively extract texture of surface under inspection and evaluate its surface roughness. This non-contact type of measurement technique is efficient, reliable, fast, robust and cost-effective in capturing surface quality. In this paper, recent advances in machine vision based evaluation of surface roughness using texture analysis techniques and predictive modelling methods are reviewed. Image texture analysis techniques used by researchers for this purpose are mainly categorised into statistical approaches and filter based approaches in spatial and frequency domain. The image textures extracted using these techniques are then used for evaluating surface roughness using various prediction models. A comprehensive study of these techniques is presented in this paper that throws light on the current state-of-technology.
Keywords- Machine Vision, Surface Roughness, Texture Analysis.