摘要
农作物生殖生长时期的生理生态性状,可通过其冠层图像表现型特征呈现出来,为解决农作物冠层图像特征难以提取的问题,以北方寒地大豆为研究对象,将数字图像处理技术与多维特征选取机制相结合,研究了农作物冠层图像特征提取技术,并应用在大豆叶部病害诊断中。首先计算农作物冠层器官形状和颜色多维特征指标,然后利用主成分分析技术筛选有效特征参数,最后完成农作物病斑害图像特征提取及诊断方法。其仿真实验中病害种类诊断准确率为97.5%。该方法实现了农作物冠层图像特征提取及信息处理过程,为大田农作物全面系统地开展生长过程监测及病害防治提供重要理论支持。
Physiological and ecological traits during crop reproductive growth could be visualized by their canopy image phenotype.In order to solve the problem of difficult extraction of crop canopy image feature,northern winter soybean was as the research object,through the combination of digital image processing technology and multidimensional feature selection mechanism,the crop canopy image feature extraction technique was studied and applied to the diagnosis of soybean leaf disease.First,calculate the multi-dimensional characteristics of the shape and color of the canopy of crops,and then use the principal component analysis to screen the effective characteristic parameters.At last,the image feature extraction and the diagnosis method of the crop blight were completed.The diagnostic accuracy of disease types in the simulation experiment was 97.5%.The method realized crop canopy image feature extraction and information processing and provided important theoretical support for the comprehensive and systematic monitoring of the growth process and the prevention and treatment of diseases in the field crops.
作者
马晓丹
朱可心
关海鸥
冯佳睿
刘梦
郑明
Ma Xiaodan;Zhu Kexin;Guan Haiou;Feng Jiarui;Liu Meng;Zheng Ming(College of Electrical and Information,Heilongjiang Bayi Agricultural University,Daqing 163319)
出处
《黑龙江八一农垦大学学报》
2019年第2期93-99,共7页
journal of heilongjiang bayi agricultural university
基金
国家自然科学基金项目(31601220)
黑龙江省自然科学基金项目(QC2016031)
中国博士后科学基金项目(2016M601464)
关键词
农作物冠层
图像特征
提取技术
病害诊断
病害防治
crop canopy
image features
extraction technology
disease diagnosis
disease control
