【作者】 石金明；

【导师】 陈於学；

【作者基本信息】 华中科技大学 ， 机械（专业学位）， 2022， 硕士

【摘要】 随着科技的进步,光学仪器在人们的生产生活中越来越重要。光学玻璃作为光学仪器中不可或缺的一环,其表面疵病对光学仪器的影响显著,因此对光学玻璃表面的疵病检测十分必要。目前主要通过人工法检测光学玻璃表面,不利于光学玻璃的大规模生产与检测,因此需要进一步研究光学玻璃表面疵病的参数化检测方法。本文采用线扫描技术获取光学玻璃表面图像,基于Open CV视觉库设计检测算法,实现了光学玻璃表面擦痕、麻点以及脱膜疵病的参数化检测,主要研究内容包括:(1)提出了一种光学平板玻璃表面轻擦痕边缘提取及其微分检测算法。首先,基于Canny算子识别轻擦痕边缘;然后,采用Hough变换算法消除虚假边缘,实现擦痕边缘分割;最后,针对弯曲的擦痕,采取基于改进的MER算法对其微分后获取擦痕参数,实现擦痕参数化检测。(2)提出了一种光学平板玻璃表面细麻点改进的Do G斑点检测算法。首先,采用线性变换增强麻点特征;然后,基于Do G斑点检测算法获得麻点增强区域;最后,采用灰度加权算法对增强区域收缩,实现麻点参数化检测。(3)提出了一种光学镜片表面脱模调制/解调检测法。首先,使用水平的光栅条纹调制脱膜疵病;然后,采用单向微分算法解调,分离脱膜疵病与背景光栅条纹;最后,采用OTSU算法筛选并提取脱膜疵病,实现对脱膜的参数化检测。论文设计并搭建了检测系统,对给定光学玻璃样品进行参数化检测,实验结果表明该检测系统具有准确性与可行性。更多还原

【Abstract】 With the progress of science and technology,optical instruments are becoming more and more important in people’s production and life.As an indispensable part of optical instruments,surface defects of optical glass have a significant impact on optical instruments,so it is very necessary to detect surface defects of optical glass.At present,the surface of optical glass is mainly detected by manual method,which is not conducive to the large-scale production and detection of optical glass.Therefore,it is necessary to further study the parameterized detection method of surface defects of optical glass.In this paper,line scanning technology is used to obtain optical glass surface images,and detection algorithm is designed based on Open CV vision library to achieve parameterized detection of optical glass surface scratches,digs and demulching defects.The main research contents include:(1)An algorithm for edge extraction and differential detection of light scratches on optical flat glass surface is proposed.Firstly,the edges of light scratches are identified based on Canny operator.Then,the Hough transform algorithm is used to eliminate false edges and realize the segmentation of scratch edges.Finally,for the curved scratches,an improved MER algorithm is adopted to obtain the scratches parameters after differentiating them,so as to realize the parameterized detection of scratches.(2)An improved Do G digs detection algorithm for fine pitting on optical flat glass surface is proposed.Firstly,linear transformation is used to enhance the pitting feature.Then,the spotty enhancement region was obtained based on Do G spot detection algorithm.Finally,the gray weighted algorithm is used to shrink the enhanced area to realize the pitting parameterized detection.(3)An optical lens surface demoulding modulation/demodulation detection method is proposed.First,a horizontal grating fringe is used to modulate the defect.Then,unidirectional differential algorithm is used to demodulate the defect and the background grating fringe.Finally,OTSU algorithm is used to screen and extract decoating defects,and parameterized detection of decoating is realized.In this paper,the detection system is designed and built,and the parameterized detection of the given optical glass sample is carried out.The experimental results show that the detection system is accurate and feasible.更多还原