基于压缩域信息的运动区域检测  

Motion Area Detection Based on Information of Compressed-Domain

作　　者：姚春莲[1] 高丽华[1] 陈谊[1] 李炜[2] 

机构地区：[1]北京工商大学数字内容学院,北京100037 [2]北京航空航天大学计算机学院数字媒体室,北京100083

出　　处：《系统仿真学报》2008年第S1期451-454,共4页Journal of System Simulation

基　　金：北京教委项目(KM200810011004);新世纪优秀大学人才计划资助

摘　　要：运动对象检测是智能视频监控的基础,现有的运动对象检测算法在实时性和检测精度方面仍有待提高,提出了一种利用压缩域的信息运动区域检测算法,首先利用压缩过程中产生的运动矢量信息去除摄像机的背景运动,然后利用宏块的DCT系数、宏块类型等信息去除虚假的运动块,最后使用了形态学的腐蚀与扩张处理以提高检测精度。测试结果表明,本文算法速度快、能够准确检测出运动区域。Motion object detection is the basis of video surveillance;the real-time performance and the detection accuracy of the existing algorithms need to be improved.Motion area detection algorithm based on compressed-domain information is proposed in this paper.Motion vector produced in the process of compression is used to be rid of background motion of camera,then information of DCT coefficient and macro-block type are used to wipe off illusive motion block.Morphological post-processing method is introduced to reduce noise and improve detection precision.Testing with standard sequence,this algorithm has the advantage of high-speed,and can detect motion area exactly.

关 键 词：运动对象检测 运动矢量 压缩域 DCT 宏块类型 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]