自适应权值的静态背景提取与更新算法  

Static background extraction and updating algorithm based on adaptive weights

作　　者：桑海峰[1] 李伟[1] 陈仲[1] 

机构地区：[1]沈阳工业大学信息科学与工程学院,沈阳110870

出　　处：《电子测量与仪器学报》2015年第3期358-367,共10页Journal of Electronic Measurement and Instrumentation

基　　金：国家自然科学基金(61374147);教育部高等学校博士学科点专项科研基金(20122102120004)资助项目

摘　　要：对比了多种视频中的静态背景提取方法,针对这些方法各自的局限性,提出一种新型的背景提取更新方法。该方法通过引入新参数背景权值及其变化量,背景权值作为评判背景可靠程度的标准,以其数值大小控制背景的维持或者更新。对每一帧图像进行二次判定,将读入的图像元素进行分类,分别产生不同的背景权值变化量,通过背景权值变化量控制背景权值。实验验证表明该方法具有原理简单,运算快捷,更新速度较快且抗前景干扰能力强的优点,而且通过调节定值参数可以适应多种不同情况下的背景提取。Compared with several methods of extraction from static background of video , and aimed at the limita-tions of these methods, a new type of method is presented for updating background extraction.In this method, the background weight is thought as a standard of reliability to evaluation background by introducing a new parameter value and its variation , and it is anticipated to maintain or upgrade the numerical size control background .For each frame image of two judges , the image elements will be classified , and the variation of background weights will be produced separately, control weight by weight change background.The experimental results show that this method has the advantages of simple principle, fast operation, updates faster and anti-interference ability of the foreground, and can be adapted to different circumstances background extraction through adjusting the setting parameter .

关 键 词：静态背景提取 背景权值 背景权值变化量 背景更新 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]