Surface flaws detection algorithms for large aperture optical element

In this paper we consider the inspection of surface flaws in large aperture optical element. A high efficiency and precision instrument is proposed which contains two kinds of imaging systems. One is dark-field imaging system (DFIS) constructed by line scan camera with 10μm resolution. The other is bright-field imaging system (BFIS) constructed by microscope with 1μm resolution. Considering the small depth of field of DFIS, an adaptive scanning method based on collimation laser and several focusing points is proposed to keep the clarity of DFIS in large scope scanning. After the scanning, an image mosaic algorithm for the DFIS is presented based on SIFT features and clustering algorithm. Then, the feature extraction algorithm of flaws in DFIS and BFIS is designed. In order to check the flaws more precisely, a location algorithm that can guide the BFIS to inspect the same flaw in DFIS is introduced. Finally, the calibration method of two imaging systems is studied. Experiments show that the device can scan an optical element with size of 810mm×460mm in less than 6min without complex manual operation and the detection precision can reach 3μm satisfying the requirement of practical need.