SIFT特点提取.docVIP

摘 要 图像匹配在遥感、医学、计算机视觉等很多领域中有着广泛的应用，是提取和匹配图像特征是实现图像自动配准的一种重要方法。 SIFT(尺度不变特性变换，Scale Invariant Feature Transform)最初是作为一种关键点的特征提取出来的，其匹配能力较强，可以处理两幅图像之间发生平移、旋转、防辐射[4]等变换情况下的匹配问题，甚至在某种程度上对任意角度拍摄的图像也具备较为稳定的特征匹配能力。作为一种优秀的局部特征描述子，SIFT算子被广泛应用于目标识别、图像配准、室内场景定位等场合。 论文着重介绍了一种基于图像特征值的匹配方法(SIFT算法),包括SIFT算法的发展历史，算法特点，主要思想和实现步骤，然后实现SIFT特征点提取，以寻找一种优化算法，减少SIFT的算法复杂度。(介绍用什么实验,得到什么结论) 关键词：SIFT算法，特征匹配，特征点提取 Astract In remote sensing image matching, medical, computer vision many adjacent domain in a wide range of applications, Extraction and matching image characteristics is to realize the automatic image registration is a kind of important method . SIFT was originally as a key point for feature extraction of come out, its matching with the strong ability can handle two images occur between translation, rotation and anti-radiation such transformation under the circumstance of the matching problem, even in some extent of arbitrary Angle photographs of image and also has a relatively stable feature matching ability. As a kind of excellent local characteristics description son, SIFT operators are widely used in target recognition, image registration, indoor scene positioning and so on. This paper mainly introduces a method based on image eigenvalue of matching method (SIFT algorithm). Including SIFT algorithm development history, algorithm characteristics, main ideas and implementation steps and then realized SIFT feature point extraction to find an optimization algorithm to reduce SIFT algorithm complexity . (请不要直译) Keywords：SIFT algorithm, feature matching, feature point extraction, simplified algorithm. 目 录 TOC \o 1-3 \u 摘 要 PAGEREF _Toc280518007 \h I 第1章 绪论 PAGEREF _Toc280518008 \h 1 1.1 课题背景 PAGEREF _Toc280518009 \h 1 1.2目的意义 PAGEREF _Toc280518010 \h 1 1.3研究内容及结构安排 PAGEREF _Toc280518011 \h 2 第2章 智能优化方法 PAGEREF _Toc280518012 \h 3 2.1 基于蚁群的优化算法 PAGEREF _Toc280518013 \h 3 2.2 基于鱼群的优化算法 PAGEREF _Toc280518014 \h 3 2.3 基于神经网络的优化算法 PAGEREF _Toc280518015 \h 3 2.4 本章小结 PAGEREF