基于机器视觉技术的黄瓜叶片白粉病识别  

Recognition of Cucumber Leaf Powdery Mildew Based on Machine Vision

作　　者：张鹏[1] 朱育强[1] 王丽莉[1] 周胜军[1] Zhang Peng Zhu Yuqiang Wang Lili Zhou Shengjun(Institute of Vegetable, Zhejiang Academy of Agriculture Sciences, Hangzhou 310021)

机构地区：[1]浙江省农业科学院蔬菜所,杭州310021

出　　处：《中国农学通报》2017年第21期134-137,共4页Chinese Agricultural Science Bulletin

基　　金：浙江省公益技术研究农业项目"黄瓜白粉病抗性基因精细定位及分子标记开发"(2016C32097);浙江省农业科学院青年人才培养项目"耐热;抗病黄瓜新品种的选育"(2016R23R08E03)

摘　　要：黄瓜叶片白粉病染病程度的判定,对于确定病灾预防措施意义重大。笔者采用机器视觉技术对叶片病斑进行有效识别,借助人工神经网络完成对叶片染病状态的模式分类。根据病斑的规模及面积分布提出了白斑区域面积比、平均白斑面积、白斑覆盖率等特征参数,借助这3个特征参数实现了叶片染病程度的定量分析,并借助人工神经网络完成了对叶片染病状态的模式分类。4类白斑叶片的正确识别率分别为88%、91%、92%、94%。To clarify the degree of cucumber leaf powdery mildew is important to determine the measures for disease prevention.In this study,leaf scab was identified effectively by the machine vision technology,as characteristic parameters,white spot area ratio,average white spot area and white spot coverage were presented based on the scale and area distribution of the scab.The level of leaf disease was quantitatively analyzed with the three characteristic parameters,and the pattern classification of leaf powdery mildew situation was accomplished by the artificial neural network.The correct recognition rate of the four classes of white spot leaf was 88%,91%,92%,and 94%,respectively.

关 键 词：黄瓜叶片白粉病 机器视觉 数字图像处理 BP网络 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]