本科毕业论文设计实时视频信号的运动画面提取算法研究黎永相.doc

本科毕业论文 黎永相 200831120210 指导教师 徐初东 讲师 学院名称 理学院 专业名称 电子信息科学与技术 论文提交日期 ????年??月 日 论文答辩日期 ????年??月 日 摘 要 现代社会对安全防范的要求越来越高，人们使用了几千年的普通门锁，随着社会的进步和需要，人们的生活水平和安防意识的提高，更先进的家用防盗产品必将成为安防消费的主流并最终取代旧的安防产品。安防市场的发展就是一种观念和相关防盗产品同时“进化”的过程，而红外防盗报警系统是现今社会上家居防盗的首选仪器。 本文主要针对主动红外CCD摄像头，作出实时视频信号的运动画面提取算法，并且介绍人体识别系统的一系列算法与应用。 人体识别系统分为两大部分，运动物体的分割还原和相关处理算法。运动物体的分割还原是人体识别系统前期的重要工作，本文算法通过对主动红外CCD摄像头采集的视频信号进行实时处理，运用背景差法对运动图像分割提取，再对运动图像容错处理，实施灰度值化，使实时运动图像实现灰度值模型化，方便计算实时灰度图于模板灰度图之间的相似度，而本文运用向量空间模型（VSM）法计算实时图于模板图之间的相似度，实现主动红外CCD摄像头对运动物体识别。 关键词 人体识别 VSM 图像相似度 实时 运动画面提取 Abstract Modern society for safety and higher to the requirement of the development of security market, is a kind of idea and relevant security products and evolution process and infrared anti-theft alarm system is now on the choice of social security. This article is mainly about using active infrare CCD camera, make real-time video signal of the movement, and introduces image extraction algorithm of human recognition system with a series of algorithm and application. Human recognition system is divided into two parts, the movement of objects segmentation algorithms and reduction. Objects in motion segmentation reduction is human recognition system is the important work, the algorithm based on CCD camera collection of ir active video signal processing, using real-time background difference method to extract image segmentation, then using Fault-tolerant method to deal with motion picture. Then gray value of the motion picture. To achieve real-time moving images of gray value model. It’s more convenient to calculate the similarity between template and the Real-time gray-scale image. And then we using the vector space model(VSM) to calculate the similarity between template and the Real-time gray-scale image. To achieve Active Infrared CCD camera on the moving object recognition. Key words：Human body recognition，VSM , Image similarity,Real-time，Movement picture extracting 1