Large Optical Component Surface Defects Detection

In view of the existing large optical component surface detection system,Defects resolution is low,splicing speed is slow,splicing precision is low,can not microscopic measurement of the main problem.To launch a systematic research,design a set of experimental system for large optical component surface detection system based on optical scattering imaging.The system is macro and micro tested for flaws.The main research results are as follows:(1)In the image acquisition system,an image acquisition scheme with large field of view and small field of view is designed.Using the CCD and 3D motion control platform to collect the sub aperture image of the optical element to be measured in the large field of view and with the help of image stitching technology to obtain full size image.Small field of view through the microscope positioning and auto focus to complete the microscopic clear defects image acquisition.This scheme can not only guarantee the macroscopic high clear obtain defect image,and can be used for subsequent analysis of the defects in the micro.(2)A fast and accurate image registration algorithm is proposed in Image stitching.Firstly,the algorithm uses the adaptive threshold to extract the target information of the overlapping area.Secondly,the canny edge extraction algorithm and the Graham algorithm are used to obtain the minimum external rectangle and according to the relationship between the shortest edge of the minimum circumscribed rectangle and the horizontal axis obtain the registration angle.Finally,the difference is the defects target centroid determining registration translation relations,so as to complete the registration process.experimental results show that the proposed algorithm has a significant improvement in accuracy and speed compared with the typical algorithm.(3)In order to overcome the influence of the sparse texture of Micro defects on the focusing results in the process of auto focus,a kind of focusing evaluation function based on regional gray is proposed.The experimental results show that the algorithm can effectively overcome the effect of sparse texture on the focusing results to a certain extent,and has high applicability and reliability.