摘要
车型分类是智能交通系统的重要组成部分,为实现车型自动分类,本文主要针对轿车、货车和客车的分类,设计了以车辆外形尺寸为特征的基于机器视觉的自动分类算法。该算法首先对车辆图像进行灰度化、背景差分、平滑、分割等预处理;然后提取顶长比、前后比等特征参量进行自动分类。该方法比较简单、易于实现,经实验验证,分级正确率可达86.25%,具有一定的应用价值。
Vehicle classification is an important part of intelligent transportation system.In order to realize automatic classification of vehicle types,this paper mainly classifies cars,trucks and buses,and designs an automatic classification algorithm based on machine vision,which is characterized by vehicle shape and size.Firstly,the algorithm preprocesses the vehicle image with gray level,background difference,smoothing and segmentation,and then extracts the top-length ratio,front-to-back ratio and other characteristic parameters for automatic classification.The method is simple and easy to implement.The experimental results show that the classification accuracy can reach 90%.It has certain application value.
作者
杜青
苗艳华
沈花玉
Du Qing;Miao Yanhua;Shen Huayu(Institute of Information and Automation,Tianjin Tianshi College,Tianjin,301700)
出处
《电子测试》
2019年第1期56-58,共3页
Electronic Test
基金
天津天狮学院自然科学研究项目"基于机器视觉的汽车车型自动分类算法设计"(K15003)
