ABSTRACT
Heavy metal pollution of tailings is one of the serious problems in environmental pollution, accurate estimation of soil heavy metal content is very important for the mine soil pollution monitoring. Taking Jinduicheng mine tailings in Shaanxi as the study area, soil spectral were measure with ASD spectrometer, Plumbum element content of soil samples were obtained by laboratory analysis. The wavelet transform was applied to the soil hyperspectral data for noise reduction, and the noise reduction of soil spectrum is studied by using the first derivative spectral transform and the continuum removal method. Plumbum content in the mine tailing soil were estimated by random forests, inversion results were compared with the original high spectral data and the noise reduction spectral data. The results showed that: the estimation model on the spectral data set after noise reduced by wavelet transform achieved a correlation coefficient R 2 of 0.774, and the root mean square error of RMSE is 249.125, the prediction accuracy is better than the original hyperspectral data. The results provide a theoretical basis for exploring the characteristics of soil hyperspectral data extraction, and has important significance for the heavy metal pollution monitoring of tailing soil in the mining areas.
