【作者】 刘鹏；

【导师】 潘永强；

【作者基本信息】 西安工业大学 ， 武器系统与运用工程， 2012， 硕士

【摘要】 全反射棱镜是激光陀螺仪中的关键零件,其表面微缺陷的存在严重影响了激光陀螺仪的质量和功能。长期以来光学元件表面缺陷的检测主要依靠人工目视检测,存在检测效率低下和检测效果一致性差等问题,尤其是不能实现精密度量。检测手段亟需改善。课题基于缺陷对光线的散射特性,以光束斜入射棱镜待测面,使散射光成像于显微系统,得到暗背景上的亮缺陷像,非常适合数字图像处理,实现可视化的缺陷高精度测量。光学元件表面微缺陷的检测主要针对元件表面光斑轮廓、麻点和划痕缺陷,具体研究情况包括以下几个方面：1)根据课题中所针对的缺陷对象的检测要求,分析了光学元件表面微缺陷的光学散射特性,应用显微成像法并通过搭建光学显微成像系统实现了对全反射棱镜三个工作面的微缺陷图像的采集。2)分析比较了缺陷图像滤波处理中的四种滤波方法,结果表明中值滤波法和形态学滤波法对缺陷图像具有较好的滤波效果；通过比较分析阈值分割图像法和边缘提取法,采用阈值分割法实现表面光斑轮廓图像的二值化处理,应用Canny算子边缘提取法实现麻点和划痕缺陷图像的分割处理。3)通过提取光斑轮廓处的灰度值特征和计算灰度梯度,将光学元件表面质量划分为四个质量等级。4)经过大量分析缺陷的几何特征数据,确定以圆度和短长径之比做为麻点和划痕的识别依据,其中圆度的特征值为0.055,短长径之比的特征值为0.550；分别确定出麻点和划痕的几何像素尺寸和灰度值大小。本课题是应用显微散射成像技术在光学元件表面微缺陷检测方面做出的探索与尝试,为更加完善的光学元件表面微缺陷的专家检测系统的研制打下了基础。经过实验论证,课题中提出的方法和算法都有很好的工程应用价值。更多还原

【Abstract】 Total reflection prisms are the key elements in the laser gyroscope, the surface defects such as pits and scratches cause great damage to the prism and laser gyroscope. The detection for defects of optical element uses visual method for a long time, it is imperative to improve the detecting method because the visual method is inefficient and discrepant. This project can acquire bright image of defects in the dark background through laser beam oblique incidence on the element surface and scattering light imaging in the microscope system. The bright image is very easy to digitized and then the visualized high-precision detecting is realized.The test objects contain pits, scratches, flare outline. The research is mainly concentrated on the following aspects:1) In line with the requirement in the defects’examination, the scattering properties of micro-defects can be analyzed, and then images collecting about three working faces of the total reflection prism can be achieved by the microscopic imaging optical system which is putted up on the basis of the microscopic imaging method.2) It is indicated that the methods of median filtering and morphological filtering are better for filtering about the images of defects after comparatively analyzing four filtering methods; the threshold method is used to segmented images of flare outline and the method of edge extracting of Canny operator is used to segmented images of pits and scratches after comparatively analyzing the methods of the threshold and the edge extraction.3) The surface quality of optical elements can be divided into for levels though extracting gray level of flare outline and calculating the gray gradient.4) After a lot of analysis about geometrical characteristic data, circular degree and ratio of the minor axis and the major axis are used to sorted pits and scratches, and the eigenvalues are 0.055 and 0.550; the gray value of defects can be obtained, and the size of defects can also be measured using the number of pixels.The research is the trial and exploration on detecting surface defects of optical components using microscopic scattering imaging technique. It is useful for building of the detecting system of expert. The algorithm and method are valuable to engineering application proved by the testing.更多还原