基于背景差分法的运动目标检测要点.doc

本 科 生 毕 业 论 文 题 目： 院 系： 专 业： 学生姓名： 学 号： 指导教师： （职 称） 二〇 年 月  视频监控在现代社会的安保系统中发挥了非常重要的作用。由于现代社会的复杂性，传统的那种由专门人员值守电视监控屏幕的视频监控系统已经无法满足现代社会安保的要求，于是智能视频监控技术应运而生。智能监控技术是基于计算机视觉的具有高度智能的自动化监控技术，其核心内容包括运动检测、目标分类、目标跟踪、行为识别四个方面。 本文针对运动目标检测这一方面进行探究。本文采用的检测方法为背景差分法，即从视频图像序列中将当前帧的图像与预先设定的背景图像做差分，则可得到运动目标的位置和大小等信息。之后对差分图像进行阈值化和形态学处理等操作，并判断此运动物体的大小。如果超过规定的范围，则认为出现了异常情况，将此运动目标加上红色矩形框显示出来，做出报警。 背景差分法计算复杂度适中，简单实用，易于实现。在本文的试验中，对视频中运动目标的检测取得了良好的效果，实验结果令人满意。 关键词：智能监控；运动检测；行为识别；背景差分 Abstract Video surveillance in modern society plays a very important role in the security system. As the complexity of modern society, traditional video surveillance system which need specialized workers pay much attention on the television monitor screen has been unable to meet the security requirements of modern society, so intelligent video surveillance technology came into being. Intelligent monitoring technology based on computer vision is highly intelligent automated monitoring technology, and its core content, including motion detection, object classification, object tracking, behavior recognition four aspects. In this paper, I intend to explore the area of moving target detection. Detection methods used in this paper is background subtraction, make difference from the current frame video sequence of images and the pre-set background image , then we can get the information of the moving target location and size. After thresholding and morphological processing operations on differential image, determine the size of the moving object. If it excess the prescribed range, think that the anomaly occurred, mark this moving target with a red rectangle and make alarm. Background subtraction has moderate computational complexity, and is simple, practical, easy to implement. In this experiment, the moving object in video detection achieved good results, experimental results are satisfactory. Keywords: Intelligent Monitoring；Motion Detection; Behavior Identity; Background Subtraction 目 录