基于PCA-SVM的麦冬叶部病害识别系统

Identification system for leaf diseases of ophiopogon japonicus based on PCA-SVM

以川麦冬叶部黑斑病、炭疽病、叶枯病3种病害图像为研究对象,采用K-Means聚类分割技术分离出病斑,通过对颜色、形状和纹理特征构成的46维特征向量进行主成分分析,再运用支持向量机设计的多级分类器进行病害识别,开发出的麦冬叶部病害识别系统识别率达到了94.4%,表明了系统对麦冬病害防治,促进麦冬产业现代化发展有重要意义。

Taking the leaf black spot,anthracnose and leaf blight of Ophiopogon japonicus as the research object,K-Means clustering segmentation technology was used to separate the lesion spots.PCA was carried out on the 46-dimensional feature vectors composed of color,shape and texture features,and then the multi-level classifier designed by SVM was used to identify the lesions.The recognition rate of the developed leaf disease recognition system of Ophiopogon japonicus was achieved 94.4%.Indicates that the system is of great significance to the prevention and control of Ophiopogon japonicus diseases and the modernization of Ophiopogon japonicus industry.