Study On Identification Method Of Corn Rust Based On Hyperspectral Imaging Technique

Maize is one of the most important grain crops in our country,and corn rust is a kind of disease which is harmful to maize.It is a great threat to guarantee the corn yield in our country.At present,some farmers in order to maintain high security of corn,from different developmental stages of maize seedlings until maturity,regardless of whether there is corn disease,regardless of disease will be the use of pesticides order of priority processing,not only broke the natural balance of the original ecological agriculture,also let corn in most of the remnants of harmful pesticides.In order to ensure the efficient use of pesticides to dispose of diseases,and can reduce the pesticide application scale is applied to the pressure of the natural environment,we must accurately obtain accurate data of diseases(including disease types and extent of damage),then a specific implementation of prevention and treatment of diseases effectively and scientifically.The monitoring of corn diseases and insect pests has a direct impact on the agricultural economic development rate of China.How to enhance the monitoring efficiency of corn diseases and insect pests is the focus of the current agricultural analysis and attention.This paper studies the problem of detection of rust in maize leaves using hyperspectral imaging technology,and try to use hyperspectral remote sensing image classification technology to extract large area plots in the diseased region,implementation of precision agriculture,and provide a reference method.The focus of this study is to outline and feature extraction of hyperspectral images in maize leaves using RGB image processing technology,and the use of RGB image visual lesion location of hyperspectral image on the lesion of training samples,and then use the support vector machine classification algorithm(SVM)extracted from the lesion on the leaves,then the lesion and normal leaf extract the reflectivity,and establish a detection model,to lay a theoretical foundation for the monitoring of hyperspectral imaging technology on maize leaf rust disease,provide a reference for subsequent band selection.At the same time,this paper also attempts to classify the UAV hyperspectral image,mean k-classification method using unsupervised classification in(K-Means)and iterative self organization data classification method(ISODATA),maximum likelihood classification of supervision(MLC)and minimum distance classification(MDC),regional corn rust the leaf will be single drones to shoot corn field scale extracted by comparing the accuracy of different algorithms and the classification results,choose the ideal classification algorithm,for UAV remote sensing images provide a theoretical basis for the classification of some.Because this study has some hard constraints,resulting in some studies failed to achieve,but can still provide some direction for future research reference and research,analysis of disease grade multiple UAV remote sensing image mosaic,the whole field for example.