Current machine vision-based detection methods for metal surface roughness mainly use the grey values of images for statistical analysis but do not make full use of the colour information and ignore the subjective judgment of the human vision system. To address these problems, this paper proposes a method to measure surface roughness through the sharpness evaluation of colour images. Based on the difference in sharpness of virtual images of colour blocks that are formed on grinding surfaces with different roughness, an algorithm for evaluating the sharpness of colour images that is based on the difference of the RGB colour space was used to develop a correlation model between the sharpness and the surface roughness. The correlation model was analysed under two conditions: constant illumination and varying illumination. The effect of the surface textures of the grinding samples on the image sharpness was also considered, demonstrating the feasibility of the detection method. The results show that the sharpness is strongly correlated with the surface roughness; when the illumination and the surface texture have the same orientation, the sharpness clearly decreases with increasing surface roughness. Under varying illumination, this correlation between the sharpness and surface roughness was highly robust, and the sharpness of each virtual image increased linearly with the illumination. Relative to the detection method for surface roughness using gray level co-occurrence matrix or artificial neural network, the proposed method is convenient, highly accurate and has a wide measurement range.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.