The Study On Hyper Spectral Characteristics And Inversion Of Som On Loess Plateau

This study analyzed the spectral characteristics of Calcic kastanozem, Loessial soil and cinnamon soil. And we learn the characteristics of soil aggregate spectrum. aggregate. In the laboratory the collected soil was processed such as been air-dried and ground,and then we measure the reflectivity spectrum curve of each soil layer in the laboratory. Finally, we builded the models by two methods of the stepwise multiple linear regression analysis(SMLR) and partial least squares regression(PLS).Conclusions are as follows:(1)The soil reflectivity had strong negative correlation with soil organic matter content and Iron oxide content in visible band and near-infrared band from 475 to 975nm.The most relevant band in nine hundred seventy-five nano organic matter content.(2)The reflectivity of loess region soil for different vegetations was different,the grass land was more than pinus tabulaeformis, the mixed forests and seabuckthorn.It can be seen that seabuckthorn had the stongest enrichment for nutrientsubstance.(3) This study build the soil organic matter models for loess tableland area by multiple stepwise linear regression analysis and partial least squares regression analysis, and verified this models, we found that the number of partial least squares regression model of the first-order differential transformation under optimum stability and predictability.(4) From the visible to near-infrared wavelengths, the reflectivity of aggregates with a shorter wavelength from the visible to the near infrared range is increasing. In each wavelength of the particle,when the size is 0.5mm,the reflectivity is maxiumum,and when is 2mm, the reflectivity is minimum. And effect of particle size less than the reflectivity of Organic Matter on the reflectivity of the obvious. Under under different vegetation types, the maximum average organic matter content of soil aggregates was happen in pure forest, pine, sea buckthorn, Blacked Locust lowest organic matter content, while the highest average reflectance.(5) This study builded PLS models For different sizes of aggregates of organic matter, we can analyzed the particle size has a very important impact on the ability of the model, which is a negative correlation relationship. And also found that the 0.5mm diameter model is the best one to predict the effect of origanic matter content.