中文说明：SIFT（Scale Invariant Feature Transform）是一个很好的图像匹配算法，同时能处理亮度、平移、旋转、尺度的变化，利用特征点来提取特征描述符，最后在特征描述符之间寻找匹配。该算法主要包括5个步骤进行匹配： 请点击左侧文件开始预览 ！预览只提供20%的代码片段，完整代码需下载后查看 加载中 侵权举报
English Description：
SIFT (Scale Invariant Feature Transform) is a good image matching algorithm, and can handle the brightness, pan, rotate, scale changes, the use of feature points extracted feature descriptor, and finally find a match between the characterization character. The algorithm includes five steps to match: 1, to establish the scale of space, look for a feature point; 2, excluding unstable feature points; 3, extracts a feature point in the feature descriptor; 4, the use of feature descriptors to find a match point; 5, the calculation of transform parameters. For a more detailed understanding of the algorithm, you can see David G. Lowe wrote Distinctive Image Features from Scale-Invariant Keypoints. The article has a very detailed algorithm description.