图像拼接技术作为获取全景图像的工具,已经受到越来越多学者的关注.为了得到良好的拼接图像,减少图像空间几何畸变,便要对图像进行严格的配准.Surf(Speed-Up Robust Feature)算法是近年来图像配准的研究热点之一.对传统surf配准算法进行理论分析并在此基础上提出改进算法,特征点检测时,利用设立阈值来去除多余的特征点,并用MATLAB软件对改进算法进行仿真分析.实验结果表明,改进后的算法虽然减少了特征点个数,但增加了配准的成功率,前后性能相当且速度更快.%As a tool of getting the panoramic image, Image mosaic technology technology has received more and more attention from scholars. In order to get a good mosaic image and reduce the geometric distortion of image space, the registering image is needed. Surf image registration algorithm is one of the hot topics in recent years. Based on the the-oretical analysis of the traditional surf registration algorithm, an improved algorithm is proposed. When the feature points are detected, the use of threshold to remove redundant feature points and the MATLAB software is used to simulate the improved algorithm. The experimental results show that the number of feature points is reduced and the success rate of registration is increased in the improved algorithm. The matching performance of the improved algo-rithms is the same as the before one,but less time is used.
展开▼