基于扩展MeanShift算法的移动目标跟踪研究及实现  

The Research and Implementation of Moving Target Tracking Based on Extended MeanShift Algorithm

作　　者：张彬[1] 王永雄[1] 邵翔[1] 

机构地区：[1]上海理工大学光电信息与计算机工程学院,上海200093

出　　处：《软件导刊》2016年第10期106-110,共5页Software Guide

基　　金：国家自然科学基金项目(61403254);沪江基金项目(C14002;B1402/D1402;D15009)

摘　　要：为克服时间差分法、模板匹配法、背景去除法以及光流法等传统移动目标跟踪方法的缺点,研究了MeanShift算法及扩展MeanShift算法,然后以LabVIEW为软件平台,实现了基于扩展MeanShift算法的移动目标跟踪程序,并用实验验证了扩展MeanShift算法在移动目标跟踪中的应用具有实时性、实用性、准确性和稳定性。实验结果表明,扩展MeanShift算法能很好地跟踪移动目标,甚至在移动目标被部分遮挡时也能对目标进行准确跟踪。To overcome the shortcomings of traditional moving target tracking method, such as the time difference method, template matching method, background to division method, optical flow method and so on , this paper studies the MeanShift algorithm and extended MeanShift algorithm, and then realizes the moving target tracking procedure based on the extended MeanShift algorithm using Labview. The extended MeanShift algorithm in the application of moving target tracking is verified by the experiments with real-time, practicability, accuracy and stability. The experimental results show that the extended MeanShift algorithm can track the moving target primely, it can track target accurately even when moving targets is partly covered.

关 键 词：MEANSHIFT算法 扩展MeanShift算法 移动目标跟踪 LABVIEW 核函数 

分 类 号：TP319[自动化与计算机技术—计算机软件与理论]