Study On Modeling Of Soil Salinization Based On The Measured Ground Thermal Infrared Spectra And Electromagnetic Induction Technology

The reason of land desertification, land desertification and land degradation is numerous, soil salinization is a most important reason. Xinjiang as the most important strategic reserve grain base and agricultural resources base in China, and has great strategic significance for the stability of oasis agriculture and sustainable development to ensure national food security. Now the rapid development of irrigated agriculture and unreasonable development of land and the way of land use has changed the balance of water and salt in Xinjiang Oasis. The agricultural production and ecological environment of Xinjiang Oasis is under the threat of soil salinization and secondary salinization.At present, the remote sensing technology is commonly method, which used in resources investigation, environment monitoring for the worldwide. It has many obvious advantages, such as high speed, high precision, less cost and monitoring several times s in the area at different time. It has been used in large area to monitored soil salinization based on the above advantages. But the ground of the corresponding measurement experiments also has rapid development in recent years, it also has achieved fruitful results, such as electromagnetic technology and the measured thermal infrared emissivity is the typical representative. Electromagnetic induction technology(EM38 electromagnetic inductor) can directly obtain soil apparent conductivity data. It has a real-time, fast, high precision, but the details for the expression of relatively rough and can’t fully reflect the local small area of the soil salinization information. Thermal infrared emissivity spectral features of the continuous spectrum information, has a more elaborate, nanoscale spectral resolution,can reflect the features of subtle spectral features. This study applied to combined the two technologies on the realistic problem of soil salinization, to realize regional large scale salinization monitoring provides the basis as to thermal infrared remote sensing technology.This research build the f soil salinization interpretation model of light, medium and heavy, which based on electromagnetic induction technology and the measured data. Use ISSTES methods to separated soil thermal infrared emissivity and temperature. The purpose is to minimize the interference factors of the outside world(instrument, environment, human), according to the Adjacent average method(Adjacent- Averaging), Filter, quantile Savitzky- Golay filtering(the Percentile Filter)and FFT Filter methods to smoothing noising on the sample spectra data. Savitzky-the effect of Golay filtering method is the best through the final analysis.Transformation spectra after smoothing noise and salinity model using the first order differential, logarithmic transformation, inverse transformation, the reciprocal of logarithmic transformation, second order differential. And analysis the specific relationship between them, and extract the sensitive bands.Finally has carried on the fitting model, using the linear(simple linear model),nonlinear(logarithm model, power exponent model, parabolic model and exponential model), the stepwise multiple regression model and partial least squares method.Finally, find the best model through the analysis of the different situation according to precision evaluation on 7 models.