(1)　　读取n副连续有重叠部分的图像，在n副图像中检测SIFT特征，并用SIFT

特征描述子对其进行描述。

(2)　　匹配相邻图像的特征点，并根据特征点向量消除误匹配。

(3)　　使用RANSAC方法，确定变换参数。

(4)　　图像融合

-(1) Read n successive overlapping sub-part of the image, the image of the n sub-SIFT features detected and characterized using SIFT descr iptors be described. (2) adjacent to the image feature point matching, and according to the feature vector to eliminate false matching points. (3) using RANSAC approach to determine the transformation parameters. (4) image fusion