三帧差分运动目标检测算法分析与验证  

Three-image Difference Algorithm for Moving Target Detection

作　　者：尹红娟[1] 栾帅[1] 

机构地区：[1]山东师范大学物理与电子科学学院,济南250014

出　　处：《计算机与数字工程》2017年第1期69-71,87,共4页Computer & Digital Engineering

摘　　要：运动目标检测准确与否决定了后续对运动目标行为的识别及适时跟踪的成败,在军事侦查、医学图像分析、科学研究等领域有广泛的应用。论文分析了三帧差分运动目标检测算法,利用OpenCV函数库进行仿真测试,并将其与普通帧间差分算法进行对比分析。实验结果表明,该算法准确率高,运算速度快,能满足实时要求。The accuracy of moving target detection determines the success or failure of subsequent recognition and motion tracking of the target behavior in a timely manner,moving target detection is widely used in military reconnaissance,medical image processing,scientific research,ect.This paper presents a three-frame difference algorithm for moving target detection.Some simulation tests are carried out by the algorithm and the basic frame difference method,respectively,moreover,the results are analyzed comparatively.The results illustrate that the proposed algorithm is more accurate and rapid,which can fulfill the realtime requirement.

关 键 词：三帧差分 运动目标检测 OPENCV 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]