Preliminary Study On Inversion Of Soil Copper Content Based On Leaf Spectra Of High Vegetation Coverage Area In Mines

With developments of high spectral technology and demands of the national economy for resources,hyperspectral has become one of the state of the art technologies of mineral resource exploration,achieving good application results.It provides unique indicative information for mineral resource investigation,and thus has a wide application prospect in geological prospecting over vegetation-covered area.Generally,the common remote sensing inversion methods of heavy metal in vegetation leaves are on the basis of physiological characteristics of the estimated vegetation spectrum under heavy metal stress,and empirical models with vegetation indices are established to inverse the heavy metal content of vegetation leaves.However,the inversion research of heavy metal content in vegetation-covered soil is still rare.In this study,Dexing,Jiangxi Province is chosen as the study area.In order to find out the concentration of copper in the soil and vegetation in Dexing Copper Mine,Jiangxi,and the influence of soil copper on the measured spectral data of vegetation leaves,By comparing the inversion accuracy of various models,the sensitivity of different spectral bands and vegetation indices to soil copper elements was analyzed.Sexually,an inversion model of leaf spectrum and soil copper content in vegetation cover was finally constructed in order to achieve breakthroughs in vegetation barriers,reveal or infer potential hidden mineral deposits in the ground,and quickly and accurately obtain distribution of mineral resources for vegetation coverage areas.The national mineral resources survey provides a strong theoretical basis and technical support.In this study,we first analyzed the correlation analysis and feature selection of the measured spectral data and soil copper elements.The establishment of the correlation model between the vegetation leaf spectrum and the soil copper element content was mainly based on the optimization of the characteristic bands and the spectral index.Based on the repeated verification of the relevant processing,a new feature band is preferably selected and established.The preference of the spectral index is mainly to collect various vegetation indices mentioned in other documents,and it is preferable to select part of the vegetation index that is sensitive to soil copper.Then,an inversion model of soil copper content based on measured spectral data of vegetation leaves was established and verified by the model.The influence of leaf spectra of vegetation coverage area on the inversion accuracy of soil copper content under different spectral resolutions was explored,and the established model was applied to Hyperion spectral data and GF-5 simulation data.Finally,a quantitative inversion model of soil copper elements based on the spectrum of the high vegetation coverage area of the mine was obtained.The multivariate linear model,partial least squares model,and support vector machine model were compared to obtain the optimal results of the partial least squares model.The application of satellite imagery has provided technical support for the geological prospecting of vegetation cover areas and the further application of GF-5 satellites.