基于核K-均值聚类算法的植物叶部病害识别  

Plant Leaf Disease Recognition Based on Kernel K-means Clustering Algorithm

作　　者：王守志[1,2] 何东健[3] 李文[2] 王艳春[1] 

机构地区：[1]西北农林科技大学机械与电子工程学院,陕西杨凌712100 [2]威海职业学院机电工程系,威海264210 [3]西北农林科技大学信息工程学院,陕西杨凌712100

出　　处：《农业机械学报》2009年第3期152-155,共4页Transactions of the Chinese Society for Agricultural Machinery

基　　金：国家自然科学基金资助项目(30471138)

摘　　要：针对植物叶部病害图像的特点,首先对采集到的玉米病害彩色图像采用矢量中值滤波法去除噪声,然后提取玉米病叶彩色图像的纹理特征和颜色特征作为特征向量,利用Mercer核,把输入空间的样本映射到高维特征空间进行K-均值聚类以及植物病害识别。试验涉及的4种玉米病害识别正确率达82.5%,核K-均值聚类方法适合玉米叶部病害分类。Based on the features of plant disease image,vector median filter was firstly applied to remove noise of the acquied color images of grape leaf with disease.Then texture features and color features of color image of leaf with disease were extraccted as feature vector.And by using Mercer kernel functions,the data in the original space was maped to a high-dimensional feature space in which the data has been clustered efficiently.The precision of four kinds of experimental maize diseases recognition is 82.5%,and kernel K-means clustering algorithm suited the plant leaf disease classification recognition.

关 键 词：植物病害 病害识别 核K-均值聚类 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术] S431.9[自动化与计算机技术—计算机科学与技术]