基于改进的 SI FT图像匹配方法研究

对SIFT（Scale Invariant Feature Transform）算法进行了研究。针对高斯差分金字塔阶数与特征点数关系问题，对算法进行了改进，通过控制金字塔的阶数来控制特征点数。实验数据分析表明：在图像存在不同程度的旋转变形和噪声影响时，改进后的算法鲁棒性更好，实时性更强。

Study on the Improved SIFT Image Matching Method

Study on the SIFT （Scale Invariant Feature T ransform ） algorithm ,to solve the problem of difference of Gaussi-an pyramid order and feature points relations ,improves the algorithm ,to control the feature points by controlling the or-der of pyramid .After analysis of the experimental data ,indicates when the image exists different degrees of rotation distor-tion and noise impact ,improved algorithm has better robustness and better real-time .