Vision Based Computer Aided Evaluation of Roughness of Flat Surfaces Part I

In almost all manufacturing processes, it is always observed that some irregularities are always present on the surface and a surface finish plays an important role for the quality of the product. Hence, an evaluation of the surface roughness is very essential. A number of methods are widely used in practice. However, all these methods have some limitations of measurement. In this paper, a comparison-based method for measurement of roughness of flat surfaces using a low cost 'Computer Vision System' consisting of a flat bed optical desktop scanner connected to a personal computer (PC) is proposed and developed. The machined surface under test is viewed by the scanner and its two-dimensional image is obtained. The image is then digitised and stored in the computer memory, which is further processed using digital processing techniques and then analysed.
It is observed that the values of the R, G, B colour components and Gray scale intensity at each point on the digital image depends upon the configuration of the surface at the corresponding point of the test surface. Thus, the parameters such as mean, standard deviation etc. for the intensity levels measured on the digital image is an indication of the surface configuration i.e. surface roughness. The values of the parameters thus obtained for a surface under test are then compared with the corresponding values of the standard (reference or master) surface obtained by a carefully monitored machining process. To compare the roughness of the test surface with that of the standard surface, additional statistical parameters such as Coefficient of Variance, Kari Pearson's Coefficient of Correlation, Equality Index based on Fuzzy Sets etc. are also computed. This method of comparison between the parameters evaluated for digital images of test surface and standard surface obtained from computer vision system is less expensive, yet more effective option for the evaluation of roughness of very smooth machined surfaces (even in the range of nano-meters) for finished products used in general, as well as areas of micro-electronics, bio-medical and surgical equipments etc.