Darkfield Polarization Imaging Method for Classification of Dimples and Dust Defects on Optical Surfaces

Abstract

The invention discloses a dark field polarization imaging method and device for classifying hard spots and dust defects on optical surfaces. According to the device and method, a polarization device is loaded into a microscopic scattering dark field imaging detection light path for optical element surface defects to ensure that the light path has an ability of obtaining target polarization characteristics. When hard spots and dust are classified, the device is adjusted to an optimum measurement polarization state and acquires a plurality of polarization images; through gray value division, feature vectors which are only related to defect polarization characteristics are extracted; and through simple matrix operation, classification of the hard spots and the dust is realized. According to the method and device, differences of the hard spots and the dust on polarization characteristics are sufficiently utilized, the obtain feature description discrimination degree is higher, the classification correctness is apparently enhanced when being compared with traditional method, the development and application of automatic surface defect detection systems are promoted, and a powerful way isprovided for improving an advanced ultra-precision machining technology in optical fabrication and researching various ultra-precision machining processes.

Description

technical field [0001] The invention belongs to the technical field of machine vision detection, and in particular relates to a dark-field polarization imaging device and method for classifying pitting and dust defects on optical surfaces. Background technique [0002] The digital quantitative detection of surface defects is an important part of the sustainable development of advanced optical manufacturing ultra-precision processing technology. When the surface shape and roughness are well controlled, surface defects are increasingly restricting the ultra-precision processing technology and level of advanced optical manufacturing. The digital quantitative detection of surface defects of optical components will provide a powerful means for improving advanced optical manufacturing ultra-precision processing technology and researching various ultra-precision processing technologies. At present, most of the automatic detection of surface defects of optical components is based on...

AI Technical Summary

Problems solved by technology

Currently, the most commonly used defect classification method is pattern recognition. This method is mainly based on the small differences in pitting and dust features on dark field images, using sample training to build a classifier, and finally realizes the distinction between the two, so the classification accuracy is limited. limit

Although pockmarks and dust are indistinguishable in dark field images, due to completely different material and shape properties, there are obvious differences in their modulation effects on the polarization state of light

Images