基于边缘轮廓特征的实时车型分类  

Real-time Vehicle Classification Based on Edge Contour Feature

作　　者：胡炜炜[1] 李树广[1] 吴舟舟[1] 

机构地区：[1]上海交通大学图像处理与模式识别研究所,上海200240

出　　处：《计算机工程》2007年第13期185-187,190,共4页Computer Engineering

摘　　要：提出了一种车辆侧面轮廓特征的实时车型分类方法。该文动态地提取背景,通过软传感器的触发,根据当前帧与背景帧的差分图获取车辆侧面轮廓特征曲线,设计了一个基于侧面轮廓特征的车型分类器。该方法速度快、准确度高,能很好地应用在实时系统中。This paper presents a method for real-time vehicle shape classification based on its side outline. It extracts the background dynamically, through the trigger of soft sensor, gets the silhouette of the vehicle from the difference between current frame and background frame and reaches the vehicle shape classifier based on side outline character. Compared with the other approach, it shows a higher efficiency and accuracy, which enables the excellent application in the real-time system.

关 键 词：智能交通 软传感器 车型分类 实时性 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]