摘要
为了提高汽车轮胎号自动识别的正确率,提出一种利用支持向量机的汽车胎号识别算法。首先采用图像采集设备获得胎号图像,并进行预处理,然后提取字符的特征向量,采用主成分分析对特征向量进行降维处理,最后采用支持向量机进行字符识别,同时利用遗传算法优化支持向量机的参数。仿真结果表明,采用支持向量机的汽车胎号识率达到97.52%,误识率和拒识率极低,很够很好满足汽车胎号识别要求。
In order to improve the correct recognition rate of tire number, this paper proposed a tire number recog-nition method based on support vector machine. Firstly, the images of tire number were collected by image device and preproeessed, then the feature vectors were extracted, and the feature dimensions were reduced by principal compo-nent analysis. Finally, the tire numbers were recognized by the support vector machine. The simulation results show that the tire number recognition rate is up to 97.52% based on the proposed method, and error rate and rejection rate is very low, which can meet the requirements of automobile tire number recognition very well.
出处
《计算机仿真》
CSCD
北大核心
2012年第9期383-386,共4页
Computer Simulation
基金
江苏省道路载运工具新技术应用重点实验室开发基金项目(BM2008206009)
关键词
支持向量机
特征提取
胎号识别
遗传算法
Support vector machine （ SVM ）
Feature extraction
Tire number recognition
Genetic algorithm（GA）
