Monitoring defects of a moving metallic surface through Tsallis entropic segmentation

Speckle patterns can be generated by the scattering of a laser light in a rough surface. Digital images of speckle patterns can be related with surface roughness. In this paper the Tsallis entropy was used to determine the threshold of the entropic segmentation on a digital image of speckle pattern. We show that it is a powerful tool to discern defects, which can be inferred as roughness variation, in a moving metallic sample. Furthermore, we investigated the results yielded by two laser wavelengths in order to verify the influence of this parameter. •We propose an alternative technique to detect defects of moving metallic surface.•Tsallis threshold is sensitive to abrupt roughness variation of a moving surface.•We compare the results for two laser wavelengths to verify their sensitivity.