基于Laplace谱的点模式匹配方法研究

Research on the point pattern matching method based on the Laplace spectral

点模式匹配涉及诸多研究应用领域,是一个重要而基础的问题,对点模式匹配问题研究的技术方法也多种多样。利用特征点的空间信息,研究基于Laplace谱的点模式匹配方法,首先对特征点构建的图进行Laplace矩阵构建,通过对Laplace矩阵进行的奇异值分解获取匹配矩阵,实现图像匹配。实验研究表明,该匹配算法能够减轻计算负担,缓解传统的基于图像局部灰度信息匹配速度慢、效果不理想的问题,具有很好的匹配效果。

Point pattern matching is a basic and important problem involving many applied research fields. The methods which focus on the problem of point pattern matching remain various. This paper uses the spatial information of feature points, and studies the point pattern match method based on Laplace spectrum. First, it constructs the Laplace matrix of the figure constructed by feature points, then, does the singular value decomposition of Laplace matrix to obtain the matching matrix, and achieves the image matching. The experiment results show, the matching method not only can reduce the computational burden, but alleviate the problems of slow speed and the ineffective image matching based on local gray level information, which has good matching effect.