Width measurement of metal surface cracks based on image analysis technique

Width measurement of cracks has become more important in metal surface inspection. This paper deals with a quantitative image analysis technique for width measurement of cracks in metal surface. Using three procedures including image acquisition, image processing and image analysis, the maximum width of an incident crack can be obtained. Employing a CMOS image sensor, a crack image can be acquired. The processing and analysis of the crack image are accomplished with a self-developed VC++ software running on a microcomputer. Based on Ferret algorithm, the widths of cracks are defined and quantified. The experiment results indicate that the software is convenient and could run quickly, and the method is adaptable. The results obtained can provide highly precise information for the quantification of cracks on metal surface.