序列图像的自适应背景提取及车型分类

计算机工程与应用 ›› 2007, Vol. 43 ›› Issue (12): 239-242.

序列图像的自适应背景提取及车型分类

胡炜炜李树广吴舟舟

上海交通大学上海交通大学图像处理与模式识别研究所

收稿日期:2006-05-24 修回日期:1900-01-01 出版日期:2007-04-20 发布日期:2007-04-20
通讯作者: 胡炜炜

Adaptive Background Update of The Sequence Image and Vehicle Classification

Zhouzhou Wu

(B0403291, No.800 Dongchuan Road, Shanghai, Shanghai Jiaotong University. 200240 )

Received:2006-05-24 Revised:1900-01-01 Online:2007-04-20 Published:2007-04-20

摘要/Abstract

摘要： 本文提出了一种基于计算机图像视觉模式识别理论针对车型识别的自适应背景提取得新方法。基于彩色图像的特征，利用三通道的差分图象获得对象掩模。并根据两种不同的情况分别采用了不同的背景更新方法。最后利用软传感器来触发车型提取工作。与同类方法相比，本方法速度快，准确率高，能很好的满足实时性的要求。

关键词: 智能交通系统自适应背景更新车型识别区域填充实时性
　

Abstract: his paper presents a novel method of extracting the self-adaptation background applying to the vehicle shape recognization. Based on the colorful images characters, the object mask is obtained using RGB 3-channels difference images. Two different background update methods are chosen according to 2 situations and then the vehicle shape extraction is triggered through soft sensor. Compare with the other approach, it shows a higher efficiency and accuracy, which enables the excellent application in the real-time system.

Key words: ITS, Adaptive Background Update, Vehicle classification, Region Filling, Real-time

胡炜炜李树广吴舟舟. 序列图像的自适应背景提取及车型分类[J]. 计算机工程与应用, 2007, 43(12): 239-242.

Zhouzhou Wu. Adaptive Background Update of The Sequence Image and Vehicle Classification[J]. Computer Engineering and Applications, 2007, 43(12): 239-242.