Quality grading of painted slates using texture analysis

This paper details the development of an automated vision-based solution for identification of paint and substrate defects on painted slates. The developed vision system consists of two major components. The first component of the system addresses issues including the mechanical implementation and interfacing the inspection system with the sensing and optical equipment. The second component involves the development of an image processing algorithm that is able to identify the visual defects present on the slate surface. The process of imaging the slate proved to be very challenging as the slate surface is darkly coloured and presents depth non-uniformities. Hence, a key issue for this inspection system was to devise an adequate illumination system that was able to accommodate challenges including the slates' surface depth non-uniformities and vibrations generated by the conveying system. The visual defects are detected using a novel texture analysis solution where the greyscale (tonal characteristics) and texture information are embedded in a composite model. The developed inspection system was tested for robustness and experimental results are presented.