摘要超光滑表面元件在很多电子工程与光学技术领域都有很广泛的应用，这些领域主要以大规模集成电路基片，微光元件或激光器谐振腔等器件为主。这些元件的表面质量对整个系统的运行会产生很大的影响。通常对这些元件要进行波前畸变和粗糙度检测以排除干扰，除此之外，元件表面的缺陷，擦伤，麻点，划痕等各种表面疵病将使入射光发生不同程度的散射，所以对超光滑元件表面疵病的辨识和标记有很大必要。利用光学显微散射暗场成像技术能够得到元件表面的图像，但由于精度和分辨率等要求，所使用的CCD靶面并不能覆盖整个元件表面，得到的是很多幅有重叠区域的疵病图像，得到这些图像后需要对其进行仿形拼接，得到一幅完整的表面疵病图像。本毕业论文是在得到多幅有重叠区域的疵病图像的基础上对这些图像进行仿形拼接，利用Matlab软件对拼接过程进行仿真，尽量做到运算速度较快，拼接效果较好。23236
关键词 疵病图像 仿形拼接 程序仿真
毕业设计说明书（毕业论文）外文摘要
Title Principle research of defect surface image’s copying mosaic
Abstract
Super-smooth surface components have a very wide range of applications in many areas of electronic engineering technology and optics technology. Those areas mainly based in large scale integrated circuit substrates, shimmer devices or laser cavity devices. The surface quality of these components will have a huge impact on the whole system. Generally these components need to perform wavefront distortion and roughness detection to eliminate interference, and scratch, pitting, scratches and other surface flaws on the surface of the components will cause incident light occurs varying degrees of scattering. So for super-smooth component surface’s flaws have a great need to identify and mark. Using optical microscopy scatter dark field imaging techniques can get images of component surface. However, due to the accuracy and resolution requirement, CCD target surface can not cover the entire component surface, a lot of pieces of flaw images with overlap region can be get, then copying mosaic these images, get a complete picture of surface flaws. This article is mosaic multiple image with overlap regions, using Matlab software simulate mosaic process and try to make computing faster, mosaic better.
Keywords flaws image Copying mosaic Simulation
目次
1. 绪论 5
1.1 课题来源及研究背景 5
1.2 图像拼接技术的研究背景及研究现状 6
1.3 本文的主要工作 7
2. 仿形拼接的原理和算法 9
2.1 仿形拼接的基本原理 9
2.1.1 仿形拼接的总体思想 9
2.1.2 仿形拼接的实现方法 10
2.2 仿形拼接的算法 11
2.2.1 图像匹配算法 11
2.2.2 求解变换矩阵 14
3. 程序主要步骤及Matlab仿真结果 16
3.1 实现仿形拼接的主要步骤 16
3.1.1 特征点检测匹配 16
3.1.2 得到变换矩阵 17
3.1.3 仿形拼接 17
3.1.4 主函数（实现任意幅图像拼接） 18
3.2 无伸缩变形情况下拼接的Matlab仿真结果 19 Matlab表面疵病图像的仿形拼接原理研究:http://www.751com.cn/tongxin/lunwen_16206.html