Identification of surface defects on glass by parallel spectral domain optical coherence tomography.

Abstract

Defects can dramatically degrade glass quality, and automatic inspection is a trend of quality control in modern industry. One challenge in inspection in an uncontrolled environment is the misjudgment of fake defects (such as dust particles) as surface defects. Fortunately, optical changes within the periphery of a surface defect are usually introduced while those of a fake defect are not. The existence of changes within the defect peripheries can be adopted as a criterion for defect identification. However, modifications within defect peripheries can be too small to be noticeable in intensity based optical image of the glass surface, and misjudgments of modifications may occur due to the incorrectness in defect demarcation. Thus, a sensitive and reliable method for surface defect identification is demanded. To this end, a nondestructive method based on optical coherence tomography (OCT) is proposed to precisely demarcate surface defects and sensitively measure surface deformations. Suspected surface defects are demarcated using the algorithm based on complex difference from expectation. Modifications within peripheries of suspected surface defects are mapped by phase information from complex interface signal. In this way, surface defects are discriminated from fake defects using a parallel spectral domain OCT (SD-OCT) system. Both simulations and experiments are conducted, and these preliminary results demonstrate the advantage of the proposed method to identify glass surface defects.