Optical detection of mechanically damaged areas on manufactured metal parts by spatial frequency filtering

Abstract

An inspection of mechanically damaged areas on manufactured metal parts is necessary to produce high-quality products. A scanning probe on a sample is necessary for a conventional surface inspection system, which is time-consuming. We propose a novel high-speed detection method for defects on metal surfaces with rolling indentations. To obtain a large field of view in a measurement, we used a laser sheet that was expanded with a laser line generator and also used an expanded collimated beam, rather than a small laser spot as used in conventional techniques. Furthermore, we used an obliquely incident laser beam to suppress the effect of the rolling indentations surrounding defects, and also applied spatial frequency filtering to extract only defects. The spatial frequency filtering under oblique incidence is theoretically explained and defect extraction was investigated in experiments.