Surface roughness evaluation using machine vision approach and hurst operator

It has been proved by researchers that computer vision has real potential when applied to the automated measurement of surface roughness of engineering components. The present study considers the detailed examination of surface roughness using 2D-image information. First the image is smoothed out using a conventional low pass filter to filter the roughness information from other erroneous information accounting to lighting, waviness, from errors and other effects. A mean value mask is applied over the image a number of times to generate the reference intensity surface. This reference intensity surface is then subtracted from the original image to reveal the surface roughness information. The Hurst operator is then applied over this intensity information to estimate the optical roughness parameter. A correlation graph has to be plotted to relate this optical roughness value with the value (Ra) obtained from the stylus profilometer. In this procedure the workpieces used were of mild steel and they were machined using different processes like milling, shaping and grinding to ensure the repeatability of results. In this process the orientation of workpiece doesn't effect the process and the lighting effect is not considered, and is maintained and assumed that it is identical for all the images. © Society of Photo-Optical Instrumentation Engineers.