Zn Contamination Monitoring Model Of Rice Based On Fastica And Hyperspectral Index

Zn is an important metal closely related with human lives, however, a large number accumulation of Zn is harmful to the ecology system security. Zn soil concentrations exceeding 200 mg / kg may be harmful for plant to grow, excess Zn concentration can lead to plant poisoning occurred. in the prevention and treatment of heavy metal pollution ,monitoring and identification is critical. However, traditional monitoring methods is mainly chemical-based,which is high-cost, time-consuming, laborious and has limited monitoring scope; but hyperspectral remote sensing technology-based monitoring method, with a wide field of vision, informative and fast, dynamic monitoring and other features, has become an increasingly important techniques of resource extraction and soil contamination.In the ecosystem, the crop spectral reflectance will change depending on the contamination condition. zn contamination of rice with different concentrations of stress is identified by remote sensing diagnosis method from the representative spectral reflectance and potential Hyperspectral index.at spectral reflectance level,to reduce the reaction between the spectral, spectral reflectance of representative bands in visible and near infrared spectral band are decomposed using the method named independent component analysis(ICA), the independent components which reflect the change of Zn contamination concentration are found. Thus the visible–near infrared independent component space is established.at spectral index level, the responsive relationship of the Hyperspectral index and physiological parameters under the stress of zinc pollution,which is chlorophyll content, water content,is systematically analyzed. Through the experiment,Hyperspectral remote sensing indexes which reflect the change of ecological parameters and their interactive law are extracted.Conbined with the visible and near infrared independent component, the three–dimensional spectral identification model of hyperspectral remote sensing index which reflect the change of Zn contamination is established.Using multiple regression analysis, the best fit model(RED-NPCI-Depth983 model) is finded, and it is helpful to monitor Zn contamination of rice.