Research On Scatter Modeling And Detection Technology Of Optical Surface Damage

Surface damage detection is one of the important contents of optical surface quality evaluation.In recent years,a lot of research on automatic detection technology of surface damage has been carried out at home and abroad,there are two key technical problems in automatic detection at present.On the one hand,for the surface damage of micron scale in the detection standard,the scale is between the micro detection and macro highresolution imaging,which is a weak scattering process.At present,there is no efficient and flexible detection method for this level in the detection technology in the industry,which has the problems of excessive resolution and small field of view.On the other hand,in practical application,the samples exposed in the air are easy to be attached with dirt,such as dust.Due to the similarity of surface damage and dirt in the intensity image,it is difficult to distinguish them from the dirt that can be removed,resulting in high detection error rate.In this dissertation,based on the application background of automatic detection of surface damage in the current industrial production process in China,based on the establishment and later application of surface damage scattering model,the theoretical establishment,model verification and simulation analysis methods of surface damage scattering model are systematically studied.In this dissertation,BRDF(bi-directional reflection distribution function)is used to establish the surface damage scattering model.The diffraction effect of light source coherence on micron scale surface damage is considered.Based on non paraxial scalar diffraction theory,the diffraction BRDF model of partially coherent light with Gaussian filter kernel function as spatial weight is established.By describing the transfer function of surface damage,a BRDF model of surface damage diffraction is established,which is more suitable for describing the real physical process.In the verification of BRDF model of surface damage,this dissertation discusses the Fourier transform light scattering measurement method,analyzes the effective angle of single lens to achieve Fourier transform,uses the sub aperture scanning method to overcome the limitation of this angle measurement range,and uses the array detector to receive the scattered radiation to achieve fast and high-precision BRDF measurement.On this basis,the validity of the experiment and the model are verified respectively.The total integral scattering of the optical element surface,which is combined by the surface damage scattering and the surface roughness scattering,is studied.The usability of the model in proposing the detection standard and detection method is analyzed.In view of the distinction between surface damage and particle contamination,there are few studies on the distinction between surface damage and contamination from the perspective of models.This dissertation analyzes the whole image chains process of surface damage and particle contamination,and proposes a modeling method to study the difference in light scattering parameters between surface damage and particle contamination.On this basis,a polarization description method is proposed to increase the degree of freedom of scattering parameter analysis.Based on Fresnel transmission law and Mie scattering theory,a polarization transmission model of surface damage and surface contamination is established by combining with Mueller matrix.The main factors affecting polarization scattering are analyzed.In terms of engineering application of theoretical research,for the special detection target is a 20-inch photomultiplier glass bulb with large area and ellipsoid shape,a detection scheme suitable for the research object is designed based on model analysis.According to the target intensity distinguishing characteristics obtained by the simulation analysis of surface damage and particle contamination,using the one-to-one mapping relationship between the detector radiation dynamic range and the exposure time,the optimal exposure time of the detector is obtained through experimental optimization,and the particle contaminations are filtered out for imaging interference.Finally,the highprecision,full-coverage and rapid detection of components is realized by the method of processing first and then counting and splicing.The research content of this dissertation has laid a solid theoretical and technical foundation for the design of automatic detection systems for surface damage of optical components used in different fields.