基于DM642的混合高斯模型运动目标检测系统设计与实现  

作　　者：史露[1] 朱星宇[1] 东方[1] 

机构地区：[1]武汉大学物理科学与技术学院,湖北武汉430072

出　　处：《中国科技信息》2012年第6期84-85,共2页China Science and Technology Information

摘　　要：运动目标检测是智能视频监控技术的重要环节,精确的分割出运动目标是后续监控技术得以实现的重要基础。该文以强大运算能力的DM642为核心搭建了运动目标检测系统硬件平台,实现和优化了基于混合高斯模型的检测算法。实验表明,该系统能有效地检测出运动目标,实时性较高。Moving object detection is an important part of video surveillance technology,pPecise segmentation of moving objects is an important basis for subsequent monitoring technology to achieve. This paper design moving target detection system hardware platform which is constructed with the DSP chipset TMS;520DM642 as the core,implement and optimize ＆ detection algorithm based on Gussian mixture model.The experiments show that the system can detect moving target effectively and meet the real--time requirements.

关 键 词：运动目标检测 混合高斯模型 背景差分 TMS320DM642 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]