摘要
为了解决复杂环境下汽车牌照识别困难的问题,建立基于改进SVM方法的图像识别模型。首先,采集牌照字符信息数据,建立字符识别数据库;然后,利用最小二乘支持向量机(LS-SVM)方法建立图像识别模型,并通过粒子群算法来实现模型中参数的优化;最后将改进的SVM分类模型应用于复杂环境车牌自动识别。实验结果表明该方法具有识别准确度高、消耗时间短的优点。
In order to solve the problem of vehicle license identification in the complex environment, constructs the image recognition model based on improved SVM. First of all, collects the character information data of licence to establish the character recognition database. Then, uses the least squares support vector machine（LS-SVM） method to create an image recognition model, and particle swarm algorithm to achieve the optimization of the parameters in the model. Applies the improve SVM classification model to the license plate automatic identification in the complex environment. The experimental result shows that it has the advantages of high recognition accu- racy and short time consumption.
基金
国家自然科学基金项目(No.61074147
No.60374062)
广东省自然科学基金项目(No.S2011010005059)
广东省自然科学基金团队项目(No.8351009001000002)
广东省教育部产学研结合项目(No.2011B090400460)
关键词
最小二乘支持向量机
图像识别
多维关联规则
混沌粒子群算法
Least Squares Support Vector Machine
Image Recognition
Association Rule
Chaos Particle SwarmAlgorithm
