摘要
通过对双目立体成像基本原理的研究,依据立体匹配算法的概况,描述了一种基于图像特征值的匹配方法 (SIFT算法)。介绍了SIFT算法的特点、主要思想和具体实现步骤,并在MATLAB7.0环境下进行匹配模拟实现SIFT算法。仿真分析表明,SIFT特征匹配对的位置信息、方向信息和尺度信息都是准确的。
Based on the principle of the bioncular stereo formation of image,the categories and the improved algorithms in process of the development are emphatically analyzed,then focusing on a value matching in image registration（SIFT,Scale Invariant Feture Transform）.It includes the algorithm characteristics,main ideas and implementation steps,and then a virtual model is constructed,and eventually the algorithm of SIFT is realizesd..The simulation results show that the information of the location,the orientation and the scale of the SIFT feature matching is accurate.
出处
《机械工程与自动化》
2011年第5期41-42,45,共3页
Mechanical Engineering & Automation
关键词
双目立体匹配
SIFT算法
尺度空间
特征值
binocilar stereo matching
SIFT algorithm
scale space
characteristic value
