基于支持向量机的苹果叶部病害识别方法研究  

Study on Apple Leaf Disease Recognition Method Based on Support Vector Machine

作　　者：王建玺[1] 宁菲菲[1] 鲁书喜[1] 

机构地区：[1]平顶山学院软件学院,河南平顶山467000

出　　处：《山东农业科学》2015年第7期122-125,141,共5页Shandong Agricultural Sciences

基　　金：河南省科技厅科技发展计划项目(No.134300510037)

摘　　要：为提高苹果叶部病害自动识别水平并实现快速有效地识别苹果叶部病害,本研究首先采用小波滤波算法对采集的苹果叶部锈病、斑点落叶病的图像进行去噪平滑,然后利用病斑颜色差异和边界跟踪算法对病斑进行分离,最后提取病斑颜色、形状和纹理等方面的特征,采用支持向量机(SVM)技术对病害进行自动识别。试验表明,该方法对苹果叶锈病和斑点落叶病样本进行处理识别的正确率较高,能够满足实际需求。该结果对苹果叶部病害的自动快速诊断和防治具有一定的指导意义。To improve the level of automatic identification and implement quickly and effectively identifi- cation of apple leaf diseases, the images of rust and alternaria leaf spot were firstly pre - processed using wavelet filtering algorithm, then, the disease spots were correctly segmented based on color difference and boundary following algorithm, finally, the disease features of color, shape and texture were extracted for auto - recognizing via SVM technology. The tests showed that the method showed a high accuracy for disease recognition and could meet the actual needs. The results had certain significance to the rapid automatic diagnosis and control of apple leaf diseases.

关 键 词：苹果叶部病害 小波滤波算法 最大类间方差法 边界跟踪算法 支持向量机(SVM) 模式识别 

分 类 号：S436.611[农业科学—农业昆虫与害虫防治] TP391.4[农业科学—植物保护]