摘要本文主要研究了图像拼接技术的一些方法。由于距离的限制，某些超大尺寸的物体无法仅用一张照片拍摄下来。当相机的视野小于人类的视野时，人们自然考虑到将多个照片拼接成一幅大的照片。于是，图像拼接技术应运而生，并成为近年来图像处理领域中的一个非常活跃的领域。73355

首先，本文介绍了图像拼接技术的研究现状以及整体流程。其次，对图像拼接技术中的两大方法——基于频域的图像拼接和基于特征点的图像拼接展开论述。对基于频域的图像拼接，介绍了相位相关算法的基本理论，在平移的基础上引入了旋转角度和尺寸缩放等因素；对基于特征点的图像拼接，主要研究了基于尺度不变特征点（SIFT）的算法，并提出了简化的特征描述向量构建方法，最后通过确定图像之间的投影变换在柱面坐标系下实现全景图像拼接。

针对这两种方法，本文还在Matlab平台上进行了相应的仿真实验，实验结果充分证实了每种方法对于各自处理条件下的图像拼接的合理性以及有效性。

毕业论文关键词  图像拼接  图像配准  相位相关   尺度不变   SIFT算法

毕业设计说明书外文摘要

Title  Research and Implementation of Algorithm for image Mosaic               

Abstract

The main methods of image mosaic are studied in this paper。Because of the limit of distance,the huge object can not be displayed in one picture。When the camera’s visual field is smaller than human’s,image mosaic would be considered to deal with the issue。Therefore,the image mosaic technology emerges and becomes a very active method in image processing field。

Firstly, the paper introduces the current situation and the whole process of the image stitching technology。Secondly,image mosaic based on frequency domain and image mosaic based on feature are discussed。For image mosaic based on frequency domain,the basic theory of phase correlation algorithm is introduced,and on the basis of the translation,the rotation angle and re-sizing are brought into expand the theory research。The image mosaic based on feature mainly researches the algorithm based on scale invariant feature points(SIFT),and the construction method of simplified feature vector is suggested。Finally,a panoramic image mosaic is achieved in cylindrical coordinates system by determining the projection transformation between the images。

For both methods, the corresponding simulation experiments also have been carried out  on the Matlab platform。The results proved their rationality and validity to image mosaic under special conditions。

Keywords:Image Mosaic; Image registration; Phase-Correlation; Scale-invariant;SIFT

目   录

1  引言（或绪论） 1

1。1  研究的背景与意义 1

1。2  研究现状 1

1。2。1  国内外发展现状 1

1。2。2  图像拼接方法分类 3

1。3  本课题研究的主要内容 3

2  图像拼接的实现过程 5

2。1  基本流程 5

2。2  图像获取 5

2。3  图像配准 5

2。3。1  基本概述 5

2。3。2  配准策略 6

2。3。3  评价函数