The Research About SOM Mapping Application Based On A Variety Of Models

Soil is a natural resource for human survival, an element of the geographicalenvironment, and as an important component of the soil, soil organic matter plays acrucial role in agricultural production. Compared to the traditional conventionalmethods to measure soil organic matter content, using hyper spectral remote sensingtechnology to predict soil information (organic matter content) has a short cycle, lowcost, and a series of advantages.In order to analyze oil organic matter content in the study area better, this paperwill take the soil near Dexing Copper Mine tailings as the research object, using theASD device to measure68groups of soil samples from Dexing Copper Mine tailingsarea, with the measured spectra and Hyperion satellite remote sensing data matched.By studying the spectral curve of soil, soil reflectance spectral characteristics and soilcharacteristic absorption spectrum band were obtained. Through the differentsampling points at different depths, the research on different positions, differentorganic matter content of soil spectral reflectance curve was reached. In predictionmodels of organic matter content, this paper carried many mathematical transform ofsoil reflectance spectra, analyzing the correlation between the spectral reflectance ofsoil and organic matter content in different forms, selecting high correlation transformby comparing as the dependent variable of predictive model on soil organic matterwith multivariate statistical analysis, fuzzy math, principal component analysismodeling. Based on the analysis of the data in the study area, the nonlinear model hasa higher accuracy than the linear model, and comparatively speaking, the method offuzzy mathematics has a higher accuracy than the other two methods, with a highercorrelation coefficient, a smaller average relative error.Finally, according to the established prediction model of organic matter, withENVI image processing software, using three mathematical models, respectively mapthe organic content on Hyperion remote sensing image in the study area. It is provedthat based on the measured soil spectra, using soil reflectance spectral curve, organicmatter content prediction model was built. It has the potential for rapid estimation of soil organic matter content, and has been more widely used.