渐进式异源光学卫星影像SIFT匹配方法

针对异源光学卫星影像SIFT算法匹配率低的问题，论文基于SIFT匹配方法，以特征点相对主方向、相对尺度为约束条件提取初始匹配点，并利用初始匹配点构建Delaunay三角网，采用点扩散方式来获取局部几何约束模型，基于特征点主方向约束、特征向量欧氏距离及局部RANSAC误差剔除方法渐进式地实现了对异源光学卫星影像的高精度匹配。与其它匹配算法对比实验结果表明，在异源影像存在较大尺度、旋转差异的情况下，论文算法具有匹配数量大、精度高的优势。

Progressive SIFT Matching Algorithm for Multi-Source Optical Satellite Images

Considering the low matching rate of SIFT for multi-source optical satellite images, based on SIFT matching method, initial controlling points constrained by the relative scale and main orientation of SIFT features are extracted in this paper. Then, Delaunay triangulation is built by initial matching points. Local geometrical constraint model is built through point spreading. Finally, on the basis of the main orientation of features, minimum Euclidean distance criterion and local RANSAC algorithm, the matching of multi-source optical satellite images is progressively achieved with high accuracy. Compared with other matching algorithms, in the case of large scale and rotate images, the proposed algorithm has significant advantages in terms of matching amount and precision.