摘要
介绍一种利用图像边沿特征实现图像相关跟踪的方法。从图像相关跟踪原理出发,通过对原始图像滤波、特征提取、相关函数及模板更新方法的改进,给出了实现特征相关跟踪的具体方法以及实验结果。利用目标图像边沿特征实现相关跟踪,解决了工程应用中相关跟踪算法存在的计算量大和跟踪目标易受背景影响的问题。此方法可直接利用数字图像的灰度特征,在低信噪比的情况下目标跟踪点漂移减小,目标跟踪能够较好地适应不同灰度分布的背景。使相关跟踪的性能得到了提高。
A practical method of correlation tracking based on image edge character is presented.The method to realize the character correlation tracking is introduced in detail.Also the results of experiments to prove the validity of the correlation tracking are shown.This method realized the correlation tracking by image filtering,edge-extracting,improving the correlation function and the rule to renew a sample.And the request for image quality is not strict.It could be performed using gray-scale image with lower SNR,and can be applied to the situation of complex object and background.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2004年第z1期698-699,708,共3页
Chinese Journal of Scientific Instrument
关键词
相关跟踪
边沿提取
数字图像处理
Correlation tracking Edge-extracting Digital image processing
