基于EmguCV的改进背景减除法运动目标检测方法研究  

Research on Moving Object Detection Based on Improved Background Subtraction Method of EmguCV

作　　者：郭永锋[1] 曹竟如 郭清宇[3] 

机构地区：[1]中原工学院信息商务学院,郑州450007 [2]中国电子科技集团公司第二十七研究所,郑州450000 [3]中原工学院,郑州450007

出　　处：《中原工学院学报》2017年第4期86-89,共4页Journal of Zhongyuan University of Technology

基　　金：河南省科技攻关项目(112102210499)

摘　　要：针对复杂场景中运动目标检测时出现的问题,设计了一种基于EmguCV的改进背景减除法运动目标检测方法。首先对每一帧图像进行混合高斯模型处理,再通过帧间差分算法进行背景更新,继而分析了传统混合高斯模型和传统的帧差分目标检测方法,并对算法进行了改进。利用Kmeans聚类算法采用不同的背景区域对复杂场景实现不同更新速率,在EmguCV框架的基础上利用C#语言实现对运动目标的检测。实验结果表明,该方法易于使用,且性能比较好。For moving object detection in complex scene problem,a detection method is of moving target improved background subtraction method is designed based on EmguCV.First,mixed Gauss model of each frame image is processed,and then divided by frame difference algorithm to update the background,and then the traditional mixed Gauss model and the traditional frame difference method for target detection are analyzed,and the algorithm is improved,using Kmeans clustering algorithm for complex scenes with different background regions to achieve different update rate.Finally using C#language on the basis of the EmguCV framework to achieve the target.Experimental results show that this method is easy to be used and has better performance.

关 键 词：Kmeans EmguCV 改进混合高斯模型 帧差分 目标检测 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]