基于图像处理的葡萄霜霉病单叶严重度自动分级方法  

An automatic grading method of severity of single leaf infected with grape downy mildew based on image processing

作　　者：李冠林[1] 马占鸿[1] 王海光[1] 

机构地区：[1]中国农业大学农学与生物技术学院,北京100193

出　　处：《中国农业大学学报》2011年第6期88-93,共6页Journal of China Agricultural University

基　　金：国家科技支撑计划项目(007BAD57B02)

摘　　要：为了实现植物病害严重度的精确测定和自动分级,克服目前病害严重度肉眼观测存在主观随意的缺陷,以葡萄霜霉病发病叶片为研究对象,提出一种基于图像处理技术的病害单叶严重度自动分级方法。经对完整的叶部病害正投影图像进行处理,利用K_means聚类算法自动准确地将叶片区域和发病区域分别分割出来,通过像素统计的方法提取叶片和发病区域的面积特征,从而精确地计算出发病区域所占叶片总面积的百分比,并根据分级标准给出病害严重度级别。利用该方法对葡萄霜霉病样本进行测试结果表明,该方法能够精确地估计病害严重度,对葡萄霜霉病发病叶片严重度判断的准确率为93.33%。To realize accurately calculating and automatically grading of disease severity,a kind of automatic grading method of severity of single leaf infected with grape downy mildew based on image processing was proposed.In processing the completed vertical-projected images of leaf disease,leaf area and diseased area were segmented out automatically and accurately using K_means clustering（HCM） algorithm.The area features of leaf area and that of diseased area were extracted using pixels statistic.And then the assessed severity of a single leaf was obtained by calculating area ratio between diseased area and leaf area.The results show that the proposed method can assess the disease severity accurately with accuracy of 93.33%.

关 键 词：葡萄霜霉病 严重度 自动分级 图像处理 K_means聚类 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] S126[自动化与计算机技术—计算机科学与技术]