中文摘要：

将传统相似三角形匹配方法和快速二维聚类匹配方法进行融合，再利用基于灰度的方法对部分伪匹配三角形进行剔除，实现了一种新的抗旋转、缩放的特征点匹配算法．融合后的算法对有效点的要求降低，同时通过在复数向量空间中进行相似三角形检索及参数聚类，提高了算法的效率．

英文摘要：

A new feature points matching algorithm based on similar triangles and 2-D parameters clustering was proposed, which has the function of anti-rotation and scaling. A part of pseudo-match triangles was eliminated by using redefined MCD distance approach, which greatly reduced the demand on efficient points. Meanwhile, plural vector space was used in similar triangle searching and parameters clustering, which greatly decreased the cost of this algorithm.