基于支持向量机的车型分类的设计  

Design of Vehicle Type Classification Based on Support Vector Machine

作　　者：曹洁[1] 李浩茹[1] 陈继开[2] 

机构地区：[1]兰州理工大学电气工程与信息工程学院,兰州730050 [2]兰州理工大学机电工程学院,兰州730050

出　　处：《科学技术与工程》2007年第22期5962-5965,共4页Science Technology and Engineering

摘　　要：基于支持向量机的车型分类的设计思路是通过视频采集获得车辆图像,对车辆图像进行图像预处理和特征提取后,得到分类器所需数据特征,而后采用支持向量机和二元决策树对车型分类。采用三个支持向量机的分类器和二元决策树相结合对特征数据进行分类识别,最终实现了车型分类。通过利用Libsvm(SVM模式识别与回归的软件)进行实验,取得了较好的分类效果。The design of vehicle type classification based on Support Vector Machine is collecting characteristic parameters provided to segregator after vehicle image preprocessing and feature extraction , and classifying the type of vehicle by using technique of Support Vector Machine and binary decision tree according to vehicle image from video camera. The vehicle type classification is completed through integrating the trisegregator of Support Vector Machine and binary decision tree. In the end , the test results based on Libsvm prove software prove to be ideal.

关 键 词：车型分类 支持向量机 二元决策树 车辆图像 特征提取 

分 类 号：U491.3[交通运输工程—交通运输规划与管理]