A Model Study On Monitoring Soil Nitrogen Of Cotton Field Based On Hyperspectral Remote Sensing

[Research object] Research on cotton canopy hyperspectral, soil hyperspectral and plant nitrogen content, soil nitrogen content of the relevant filtered characteristics of appropriate parameters and vegetation index, established the best cotton soil nitrogen content monitoring model. To improve soil nitrogen content to estimate the level of remote sensing, in order to apply satellite imagery to better spectral data to provide a basis for crop monitoring.[Research methods] Used data that were collected cotton canopy hyperspectral, soil hyperspectral in different treatments at Shihezi University Experiment Station in the year 2007-2008, meanwhile from canopy were investigated. Through qualitatively and quantitatively analyzed the correlation between reflectance spectrum,derivative spectra of cotton canopy and agricultural parameters, established Corresponding regression model, and the same time model validation and accuracy had tested.[Results] The study of cotton plants has a very good correlation between canopy spectral reflectance and plant,soil nitrogen content at different stages. Different stages of cotton nitrogen content of soil and the plant canopy is the relevance of the 1% significant level. It can be seen, the selected the plant canopy and soil total nitrogen content between the relevance of good, you can try to set up total nitrogen content of soil based on the plant canopy hyperspectral the monitoring model.Using a coverage of to established of use of plant canopy measurement of soil nitrogen content is the best model, followed by the establishment of plant nitrogen content of the model, followed by the direct monitoring of the plant canopy nitrogen content in soil, and finally the use of soil reflectance spectra the establishment of parameters of cotton canopy monitoring parameter model of soil nitrogen contentThe establishment of the coverage period of building a Square stage, Flower stage, Boll stage, Opening boll stage nitrogen content of soil and plant canopy parameters of the composite spectrum inversion model had tested. Inversion model built on the determination coefficient, RMSE, RE Analysis of the results show that different periods of time, on soil nitrogen content of the best models predict is different ND1190'in the Square stage to the accuracy of the model the highest, Flower stage, Boll stage, Opening boll stage is ND672'model to monitor the highest accuracy. Trough the establishment of the soil nitrogen content by coverage compare with not by, the accuracy of the precision of different stage there has been a marked increase, in accuracy is clear that Opening boll stage ND672 increased 12.4percent; the worst is Boll stage ND672 also increased 5.06 percent. The establishment of the soil nitrogen content by coverage in addition to the stage, but also with the parameters selected.