摘要
针对植物病害彩色纹理图像的特点 ,提出将支持向量机和色度矩分析方法相结合应用于植物病害识别中。首先利用色度矩提取植物病害叶片的特征向量 ,然后利用支持向量机分类方法进行病害的识别。黄瓜病害纹理图像识别实验分析表明 ,利用色度矩提取病害彩色纹理图像特征简便、快捷、分类效果好 ;支持向量机分类方法在病害分类时训练样本较少 ,具有良好的分类能力和泛化能力 ,适合于植物病害的分类。不同分类核函数的相互比较分析表明 ,线性核函数最适于植物病害的分类识别。
According to the features of color texture image of plant disease,recognition of plant disease using support vector machine (SVM) and chromaticity moments was introduced. At first,extracting features of chromaticity moments was done,then classification method of SVM for recognition of plant disease was discussed. Experimentation with cucumber disease was conducted and the results proved that chromaticity moments is simple,efficient,and effective for recognition of plant disease image,and the SVM method has excellent classification and generalization ability in solving learning problem with small training set of sample,and is fit for classification of plant disease. The comparison of different kernel functions for SVM shows that liner kernel function is most suitable for shape recognition of plant disease spot.
出处
《农业机械学报》
EI
CAS
CSCD
北大核心
2004年第3期95-98,共4页
Transactions of the Chinese Society for Agricultural Machinery
