基于AForge.NET类库的视频移动目标检测  

Moving target detection for videos based on AForge.NET library

作　　者：楼亮亮[1] 金彦亮[1] 

机构地区：[1]上海大学通信与信息工程学院,上海200072

出　　处：《现代电子技术》2015年第17期58-60,64,共4页Modern Electronics Technique

基　　金：上海市科学技术委员会重点项目(12511503300;11DZ0512500;12DZ2250200);上海市科学技术委员会资助(12511501700)

摘　　要：运动目标的检测是数字图像处理和模式识别的基础,也是计算机视觉研究的一个重要领域。以C#为主要研究工具,对基于相邻帧差法及背景差分法的视频目标检测算法进行了研究,主要对其原理和算法进行研究。最后利用以AForge.NET架构类库,利用图像灰度的绝对值是否大于设置的阈值实现了对运动目标进行检测,实验结果表明,采用该算法可以对运动目标进行较为精确的检测。Moving target detection is the foundation of digital image processing and pattern recognition,and an important domain in computer vision research. Taking C# as the main research tool,the algorithms of target detection for videos based on adjacent frame difference method and background difference method are studied,and the principles and algorithms are studied mainly.The moving target detection on the basis of AForge.NET library was achieved when the absolute value of grayscale image was larger than the set threshold. The experimental results show that background difference method can detect moving target accurately.

关 键 词：运动检测 AForge.NET 帧差法 背景差分法 

分 类 号：TN958[电子电信—信号与信息处理]