Studies On The Hyperspectral Characteristics And Content Of Ferric Oxide In The Development Of Basaltic Volcanic Debris

The iron oxide in soil is derived from the redeposition of the weathering products of the parent material and can be divided into free iron oxide(Fed)and amorphous iron oxide(Feo)by chemical methods.The content of iron oxide in soil can be used as a diagnostic indicator of soil types in the classification of soil systems,and is considered as a function of soil occurrence and development.Although the traditional method of measuring iron oxide content in soil is of high precision,it is time-consuming and laborious,and can not get real-time monitoring data of soil physical and chemical properties in time.Using hyperspectral remote sensing technology,the spectral information of soil can be quickly obtained by using the high spectrometer,and the relationship between the measured chemical data and the physical model is combined to realize the quantitative calculation of the soil information.It has unparalleled advantages of the traditional soil iron oxide measurement methods,such as fast and strong,and so on.It provides broad application prospects for the prediction of spectral information and content prediction of iron oxide in soil.In this study,the soil of the basaltic volcanic debris in Northeast China was collected,and the hyperspectral data of the indoor soil and the chemical analysis of the iron oxide content were carried out.The Savitzky Golay convolution smoothness,the first derivative,the second derivative,the standard normal variable transformation,the continuum removal,the multivariate scattering correction and so on were selected.By using 3 modeling methods of principal component regression,partial least squares regression and support vector machine,the characteristics of different spectral curves of iron oxide,the extraction of sensitive bands and the establishment of the prediction model are carried out to establish the iron oxide of the developed soil of a specific area of basaltic volcaniclastic.Content prediction model.The main conclusions are as follows:1.The spectral curve of ferric oxide developed in the basaltic volcaniclastic is rising sharply in the visible light band,the shape is steep,the change of the slope in the near infrared band is not uniform,and several twists are formed.The spectral reflectance of different forms of iron oxide content is also different.The absorption and reflection characteristics are mainly concentrated in the range of 400~1100nm band,and the specific response bands are concentrated in the 420 nm,480nm,910 nm,and the absorption valley.The size of the reflection peak or absorption Valley is related to the iron oxide content..2.Based on the original spectral reflectance SG convolution smoothing,first order differential,two order differential,standard normal variable correction,continuum removal and multiple scattering correction treatment,the correlation analysis of free iron content and amorphous iron content in the soils developed by basalt Pyroclasts was analyzed,and the correlation was significantly enhanced.The best way to deal with the problem is to make multivariate scatter correction for the original spectrum.3.In the range of total spectrum,the prediction model of 2 forms of iron oxide content is established by using 3 modeling methods: principal component regression,partial least square regression and support vector machine.The better method for the inversion of soil iron oxide content is multiple scattering correction and first derivative treatment.The support vector machine model is the best estimation model of the high spectrum of soil iron oxide content,which can be used to estimate the content of soil iron oxide.