在上篇博客特征点检测学习_1(sift算法) 中简单介绍了经典的sift算法，sift算法比较稳定，检测到的特征点也比较多，其最大的确定是计算复杂度较高。后面有不少学者对其进行了改进，其中比较出名的就是本文要介绍的surf算法，surf的中文意思为快速鲁棒特征。本文不是专门介绍surf所有理论（最好的理论是作者的论文）的，只是对surf算法进行了下整理，方便以后查阅。

该代码的作者给出的主函数实现了6中功能，包括静态图片特征点的检测，视频中特征点的检测，图片之间的匹配，视频与图片之间的匹配，特征点聚类等6中功能。本次实验就简单的测试了图片的检测，匹配和特征点聚类3个功能。并加入了简单的界面。

　　开发环境为：opencv2.4.2+Qt4.8.2+open surf+windosxp-_1 (Sift algorithm) brief in a classic sift algorithm detected feature points in the previous blog to learn and sift algorithm is more stable, more feature points detected, its largest to determine the higher computational complexity. Followed by many scholars to its improvement, one of the well-known surf algorithm, this article described the Chinese meaning of surf speed feature. This article is not devoted to surf all the theory (the best theory is that the author' s thesis), but the surf algorithm under collate for later reference. Main function given by the author of the code to achieve the 6 functions, including static image feature points detection, detection of feature points in the video, pictures match, the match between the video and pictures, feature point clustering 6 in function. This experiment a simple test image detection, matching and feature point clustering of three functions. And by adding a simple interface. Development environment: opencv2.4.2+Qt4.8.2+open sur