基于Ly特征提取算子的高精度配准方法

A High- precision Registration Method Based on Ly Feature Extraction Operator

本文探讨了一种基于Ly特征提取算子的遥感影像的高精度配准方法。该方法利用Ly算子快速提取特征点，同时采用抑制局部非最大的格网技术控制特征点均匀分布，在少量准确控制点的基础上，以相关系数测度及两个影像灰度值差的绝对值和测度相结合的多重判据为匹配准则，采用松弛法匹配策略整体匹配得到同名点。并通过实验证明了该方法特征提取速度快，基于Ly算子得到的配准点可靠，配准精度高。

A method for remotely sensed image registration is presented in this paper, based on Ly feature extraction operator. The feature points are quickly extracted with Ly operator, and uniformly distributed by restraining local non - maximal grid. According to the matching rule of multi - criterion composed of correlative coefficient measure and the sum of absolute value of image gray difference measure, the conjugated points are obtained by total matching with the matching strategy of the relaxation method. The experiment results have shown that the method of feature extraction with Ly operator is high effective, and the conjugated points are more reliable and accurate.