基于特征点的快速匹配算法  

A Fast Image Matching Algorithm Based on Characteristic Points

作　　者：甘进[1] 王晓丹[1] 权文[1] 

机构地区：[1]空军工程大学导弹学院,陕西三原713800

出　　处：《电光与控制》2009年第2期64-66,共3页Electronics Optics & Control

基　　金：陕西省自然科学研究计划项目(2007F19)

摘　　要：为了实现快速、高精度的图像匹配,将基于灰度的图像匹配技术与基于特征的图像匹配技术相结合。定义了一种灰度值特征,并提出一种基于图像灰度值特征点的快速匹配搜索算法。该算法利用得出的灰度值特征点作为搜索依据,改变了传统匹配方法遍历性质的搜索策略。由于特征点的选取对噪声和灰度值的线性变化具有一定的"免疫性",因此对灰度值不均匀的图像匹配问题同样适用。该算法在大大提高匹配速度的同时兼顾了匹配精度,实现了高效率、高精度的图像匹配。通过实验,证实了该算法的快速性和准确性。In order to accomplish fast and precise image matching, the image matching technique based on image grey-scale value and the technique based on characteristic points are combined together. An image grey-scale value characteristic is defined, and a fast matching algorithm based on image grey-scale value characteristic point is proposed. The algorithm uses the obtained grey value characteristic point as the searching foundation, which is different from the traditional ergodic search strategy. Because the selection of the characteristic point has certain ＂immunity＂ to the noise and the linear change of the grey value, this algorithm is also applicable to the asymmetrical image matching problem. The algorithm can improve the matching speed greatly while keeping the matching precision, and accomplishe image matching with high efficiency and high precision. Experiments verified the rapidness and precision of the algorithm.

关 键 词：图像匹配 搜索策略 特征点 灰度值 图像处理 

分 类 号：V271.4[航空宇航科学与技术—飞行器设计] TP391[自动化与计算机技术—计算机应用技术]