摘要
车辆自动识别分类技术是智能运输系统的重要组成部分,它对特定地点和时间的车辆进行识别和分类,并以之作为交通管理、收费、调度、统计的依据。要实现我国公路收费自动化、管理规范科学化,车型自动识别方法的研究势在必行。本文研究基于车型图像代数特征的车型识别方法。该方法首先利用背景差分法从背景图像中提取出运动车辆,并对车型图像进行预处理,然后采用特征并行融合的方法即用PCA方法,最后通过支持向量机分类器进行车型识别。
Automatic Vehicle Identification and Classification of Intelligent Transport System technology is an important part of its specific place and time the identification and classification of vehicles,and used as traffic management,fees,scheduling,statistical basis.China's road toll to achieve automation,standardized and scientific management,automatic Vehicle identification method imperative.This paper explores the characteristics of the vehicle model image algebra recognition.This method first uses background subtraction to extract from the background image moving vehicles,and vehicle image is preprocessed,and then use the parallel feature fusion method using principal component analysis,and finally through the support vector machine classifier for vehicle identification.
出处
《科技广场》
2010年第11期75-78,共4页
Science Mosaic
关键词
车型识别
特征融合
特征提取
决策支持向量机
Vehicle Recognition
Feature Fusion
Feature Extraction
Decision Support Vector Machine