基于SIFT特征点改进搜索算法的图像匹配技术研究

Image Matching Technology with Improved Search Algorithm Based on SIFT Feature Points

本文研究一种改进的近邻搜索算法的图像匹配技术。本文采用基于特征的图像匹配方法，利用SIFT算法提取特征点。在特征点匹配的过程中，为提高搜索样本特征点的最近邻和次近邻特征点的速度，本文采用一种基于二叉检索树算法改进的近邻搜索算法，该算法用最近邻与次近邻比值来进行特征点的匹配。用MATLAB语言实现该算法并运用到图像特征匹配中，实验证明优于原算法并具有较高实时性。

The paper studied the image matching technology with an improved search algorithm for k-nearest neighbors. Adopting featurebased image matching method, this paper used SIFT algorithm to extract the feature points. In the features matching process, in order to speed up searching the nearest neighbors and the next nearest neighbors of the sample feature points, based on binary search tree algorithm, this paper puts forward an improved search algorithm for k-nearest neighbors in which the ratio of the nearest neighbors to the next nearest neighbors was used to match the feature points. Experiments show that this algorithm, realized by MATLAB language to match image features, is superior to the former one and has better real-time performance.