基于边缘特征的亚像素相关匹配跟踪算法  

Subpixel Correlation Matching Tracking Algorithm Based on Edge Feature

作　　者：刘倩[1] 闫宇壮[1] 黄新生[1] 

机构地区：[1]国防科学技术大学机电工程与自动化学院,长沙410073

出　　处：《计算机工程》2011年第15期201-202,211,共3页Computer Engineering

基　　金：国家部委基金资助项目

摘　　要：基于像素灰度和基于特征的相关匹配跟踪算法跟踪精度不高,只能匹配定位到整像素点。针对该问题,采用相关函数拟合极值法对整像素匹配点进行亚像素定位。设计一种新的实验方法和基于边缘特征的相关匹配跟踪算法。实验结果表明,该算法可以达到0.1级别的亚像素定位精度。Correlation matching tracking algorithm based on pixel gray and character is commonly used. The algorithm based on edge feature is a research hotspot because of its advantages for application. However, this kind of algorithm can just locate to integer pixel for its low tracking precision. A method based on the quadratic surface fitting is selected for subpixel matching through analysis and comparison. A new experiment is designed to show that this algorithm gets as high as 0. 1 level of pixel location precision.

关 键 词：亚像素定位 边缘提取 边缘相关匹配 相关函数拟合极值法 匹配跟踪 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]