离散特征点探测算法

Dispersed Featured Point Detection

在建立新的特征点响应函数的基础上,提出了一种特征点探测算法,算法以图像分块为基础,通过计算每一分块内的特征点,进行特征点连通集约简获得离散特征点集,实验证明,算法效率得到了较大提高,解决了使用全局特征响应阈值造成的特征点聚集和不敏感特征点被错误剔除的问题。

Based on setting up a new feature point response function, we propose a new feature point detection algorithm. By image segmentation, computing feature point in each block and continuous feature points set reduction, the dispersed feature point set is reached. The experiments prove that the detection efficiency is improved, while the feature points mass due to traditional global response threshold is avoid and most of the insensible feature points are preserved.