中文摘要：

研究了图像匹配与图像角点匹配之间的关系．并在此基础上提出一种基于角点检测的图像匹配算法，成功地应用于图像拼接中．该算法将角点作为图像的特征点，并通过角点值、邻域角点数、角点间距及参数一致性等4个指标对角点集进行逐级筛选，有效地剔除了不匹配的角点，保证了匹配精度，同时避免了传统算法中进行模板匹配的繁重计算，大大提高了匹配速度．图像拼接实验验证了本文算法的快速、准确和稳定的特性．

英文摘要：

A new image matching algorithm based on corner detection was proposed and applied successfully to image stitching, The new image matching algorithm was achieved by a four-step corner matching,i, e. ,the corner value matching,the corner number matching,the corner distance matching,and the consistency checking, The unmatched corners were eliminated from the corner sets, and a precisely matched corner subset was obtained. The computational complexity of the proposed algorithm is lower than that of other algorithms with template matching. Results of the image stitching experiments showed that the proposed algorithm is fast,accurate,and robust.