基于小波变换和LS-SVM的事件检测算法  

Incident detection algorithm based on wavelet transform and LS-SVM

作　　者：裴瑞平[1] 梁新荣[1] 刘智勇[1] 

机构地区：[1]五邑大学信息学院,广东江门529020

出　　处：《计算机工程与应用》2007年第1期229-231,共3页Computer Engineering and Applications

基　　金：广东省自然科学基金资助项目(06300326)。

摘　　要：根据高速公路有事件发生时交通流将产生突变这一原理,提出一种基于小波变换和最小二乘支持向量机的事件自动检测算法。首先用小波变换提取特征数据,然后用最小二乘支持向量机进行分类,仿真实验表明该算法具有检测时间短、检测率高和误报率低的优点,它为事件检测提供了一种切实可行的新思路。According to the principle that an incident on a freeway produces a great impact on the traffic flow,an-automatic incident detection algorithm based on wavelet transform and least squares support vector machines （LS-SVM） is presented.The wavelet transform is first used for extracting feature data,LS-SVM is then applied to incident classification.Simulation experiments show that the proposed algorithm has such advantages as short detection time,high detection rate and low false alarm rate.This method is a novel and practical way to realize freeway incident detection.

关 键 词：高速公路 事件检测 小波变换 LS—SVM 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]