Research On Analysis Of Hyperspectral Character And Experiment Of Pesticide Control Effect For Cotton Disease

This article deals with the measurement and analysis the hyperspectral character of cotton diseases(anthracnose and ring rot disease)under the level of single leaf and canopy using ASD Field Spec instrument at agricultural laboratory of Jiangsu University in year 2005, 2006. The experiment is also carried out in order to gain the quantity of pesticide of different state of two diseases. The results are as follows:(1) Cotton cultivation is done in the experiment. First, cultivate seedling both in pit plate and in nutrition plate plug. After some ten days, cotton seedlings are transplanted in soil. In order to get diseased cotton samples, infection testing by artificial inducement has been done.(2) After measuring the spectrum of normal and diseased leaves and canopies, character is analyzed between 350nm and 2500nm spectral region and sensitive wave bands are found out. In addition, the disease forecasting models are established.(3) It is indicated through the research that spectral data is positively correlative to the disease degree of graveness in visible light and near-infrared wave band while negatively correlative in mid-infrared wave band under the level of single leaf. In sensitive wave band, selecting spectral data at 560nm and RVI made up of data at 980nm and 560nm as variable, separately to make quadratic and exponential models to simulate the degree of anthracnose disease, can take out preferable effect, with which the correlative coefficient of calibration equation is 0.9862 and 0.9900, while selecting spectral data at 980nm and RVI made up of data at 980nm and 560nm as variable, separately to make quadratic and power models to simulate the degree of ring rot disease, can get the result that the correlative coefficient of calibration equation is 0.9919 and 0.9956. It is also shown that spectral data is positively correlative to the disease degree of graveness in visible light wave band while negatively correlative in near-infrared wave band under the level of canopy. The method of linear forward-step regression is adopted to develop the equation to simulate the degree of graveness of two diseases and the result is that the correlative coefficient of calibration equation is 0.9976 and 0.9981 with spectral data at former five wavelengths as variable.(4) In the experiment of controlling cotton diseases with procymidone in the field, the quantity of pesticide of five-grade state of two diseases is obtained, which is greatly decreased compared to traditional spraying means, using single-factor statistical method of random sample.The aim of the research is to provide essential theoretical evidence and technological basis for further development of forecasting cotton diseases and precise spraying technology.