摘要
探讨了采用支持向量机对黄瓜病害进行分类的方法;提取了病斑的形状、颜色、质地、发病时期等特征作为特征向量,利用支持向量机分类器,选取4种常见核函数,以Matlab7.0为平台对10类常见病害进行识别。结果表明,SVM方法在处理小样本问题中具有良好的分类效果,线性核函数和径向基核函数的SVM分类方法在黄瓜病害的识别方面优于其他类型核函数的SVM。
Classification of cucumber diseases had been discussed by support vector Machine.First,we extracted the shape,color,texture,onset period of diseased spots as feature vector,then created svm classifier by four common functions. Ten kinds of cucumber diseases had been classified in Matlab7.0 platform.The results show us that the method of svm has better performance in solving small sample problems,and the classification based on linear kernel function and rbf kernel function is better than other functions on recognition of cucumber diseases.
出处
《农机化研究》
北大核心
2009年第3期36-39,共4页
Journal of Agricultural Mechanization Research
基金
辽宁省自然科学基金项目(20052125)
辽宁省教育厅攻关计划课题(2005367)
关键词
分类识别
支持向量机
黄瓜病害
特征选取
classification and recognition
support vector machine
cucumber disease
feature extracted
