基于SIFT算法的图像匹配方法研究+MATLAB程序

摘要：随着时代不断的发展与进步，获得信息的手段已经从文字向图像转变，图像已经成为了人们的首要信息而图像匹配也成为一项不可或缺的技术，很多图像分析技术都是以它为基础。图像匹配可以让信息具有高度集成性的特点，有利于储存与传送。图像匹配作为视觉技术的基础，有大量的学者对图像匹配这项技术进行了研究与改良，在提出的各种算法中，SIFT算法是其中比较成功的算法之一。SIFT算法就是尺度不变特征变化法，该算法于1999年提出，2004被加以完善。本文基于SIFT算法的图像匹配方法进行讨论并通过MATLAB软件平台对该算法从光照明暗、尺度缩放和角度旋转变换3个方面进行实验观察。通过理论性的知识和实验效果，说明了基于SIFT算法的图像匹配具有着优越性、适应性与准确性，虽然还存在不足，但是此算法未来的发展前景还是非常广阔。68394

毕业论文关键词：图像匹配，SIFT算法，匹配方法

Abstract：With the development and progress of The Times,the means of accessing to information is changing from the text to image, image has become the most important for information and image matching has become an indispensable part of technology.Many image analysis technology are based on it.Image matching can make information has the highly integrated features.It is good for storage and transmission.Image matching As the basis for the visual technology,there are a large number of scholars has studied and improved the technology of image matching.In all kinds of algorithm,SIFT algorithm is very successful of them.SIFT algorithm is scale invariant features change method.The algorithm was proposed in 1999 and were improved in 2004.Discussed in this paper, image matching method based on SIFT algorithm by MATLAB software platform for the algorithm, from the Angle of light and shade,scale zoom and rotation transformation three aspects to do experiment and observe.By the theoretical knowledge and experimental results,explain that the image matching based on SIFT algorithm has the advantages,adaptability and accuracy.Although it has many shortcomings,this algorithm also has much broader development in the future.

Keywords: image matching, SIFT algorithm, matching method

1 绪论 4

1.1 研究的背景、意义及其研究现状 4

1.2 本课题研究的内容和目标 4

2 图像匹配 5

2.1 图像匹配的概念 5

2.2 图像匹配技术的背景和意义 5

2.3 图像匹配技术的方法 5

2.3.1 以灰度为基础的图像匹配方法 5

2.3.2 以特征为基础的图像匹配方法 6

2.3.3 两类匹配方法的比较 6

2.3.4 匹配的关键要素 7

2.4 总结 7

3 SIFT算法 8

3.1 SIFT算法的概念 8

3.2 SIFT算法的原理与步骤 8

3.2.1 对尺度空间极值点进行检测 8

3.2.2 精确定位极值点的位置 9

3.2.3 找出极值点的方向 10

3.2.4 对极值点的描述 10

3.3 总结 11 基于SIFT算法的图像匹配方法研究+MATLAB程序:http://www.751com.cn/tongxin/lunwen_72997.html