基于视频图像的安全监测算法研究  

Research on Security Monitoring Algorithms Based on Video Images

作　　者：王成 朱飒爽 王智琪 WANG Cheng;ZHU Sashuang;WANG Zhiqi(School of Telecommunications and Intelligent Manufacturing,Sias University,Zhengzhou Henan 451150)

机构地区：[1]郑州西亚斯学院电信与智能制造学院,河南郑州451150

出　　处：《软件》2024年第3期54-56,共3页Software

基　　金：郑州西亚斯学院校级科研项目(2023-D054);郑州西亚斯学院应用型本科高校建设专项教改项目(2022YYXZX01);河南省本科高校2023年度产教融合重点项目“产教融合视域下的人工智能专业集群教学育人模式创新研究与实践”。

摘　　要：针对智能视频监控中运动目标检测与提取相关算法易受噪声干扰和易出现空洞现象的问题,利用Matlab环境,将使用背景差分法与二帧差分法获得的二值图像与使用三帧差分法获得的二值图像进行“或”运算得出改进算法,并进行对比评估。实验结果表明,改进后的算法能够避免侦差法检测运动目标过程中出现的空洞现象,能够克服易出现噪声干扰的缺点,得到的目标轮廓也比之前单一算法提取的运动目标更加明确。In order to address the issues of noise interference and cavity phenomenon in motion target detection and extraction algorithms in intelligent video surveillance,an improved algorithm was developed by using Matlab environment to perform"OR"operations on binary images obtained using background difference method and two frame difference method,as well as binary images obtained using three frame difference method,and a comparative evaluation was conducted.The experimental results show that the improved algorithm can avoid the phenomenon of that occur detection of moving targets using the detection method,overcome the disadvantage of easy noise interference,and obtain clearer target contours than the previously extracted moving targets using a single algorithm.

关 键 词：视频监控 动目标检测 背景差分法 三帧差分法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]