Hyperspectral Inversion Of Soil Chromium Content In Tangshan Iron Tailings Area

With the rapid economic development,my country’s demand for mineral resources is increasing.With the continuous development of mineral resources,along with the annual increase in tailings emissions,heavy metal pollution in the soil around the tailings has become increasingly serious.The traditional detection methods of soil heavy metal content have disadvantages such as expensive and easy to cause secondary pollution.The advantages of low cost and pollution-free hyperspectral remote sensing technology provide new ideas and methods for quickly and accurately obtaining soil heavy metal content.The paper studied the characteristics of soil spectrum in Tangshan iron tailings area and the influence of soil chromium content on soil spectral reflectance.Perform spectral re-sampling on the original data after spectral smoothing at 2nm and 10nm intervals,perform 8 different forms of mathematical transformation on the original data and resampled data,and use correlation analysis methods for the 9 spectral forms including the original spectrum itself.Continuous projection method and competitive adaptive re-weighted sampling method to select characteristic bands,combined with multiple linear regression,principal component regression and partial least square regression three regression models to construct the soil chromium content inversion model.The main research conclusions are as follows:（1）The analysis shows that the spectral reflectance of different soil samples in the same waveband is different,but the overall trend is the same.There are obvious absorption peaks in the soil spectral curve at 1400nm,1900nm,and 2200nm.Analyzing the influence of soil chromium content on the reflectance of the soil spectral curve,it can be seen that the spectral reflectance difference is the largest in the range of 750 to1900nm,and the spectral reflectance and the chromium content are obviously positively correlated.（2）Analyze the same characteristic wavebands extracted by different characteristic waveband extraction methods.The same wavebands extracted in the visible part of the spectrum of different transform forms are all located near the Cr³⁺absorption peak positions of 400nm and 550nm,verifying the correctness of the experiment.（3）From the perspective of different mathematical transformations of spectra,the best performance in the original spectral data soil chromium content inversion model is the original spectrum R,and the best performance in the 2nm resampled spectral data soil chromium content inversion model is the second-order Differential spectrum SD,10nm resampled spectrum data.The best performance in the soil chromium content inversion model is the original spectrum R;from the perspective of three different characteristic band extraction methods,the inversion effect of the model established based on the continuous projection method extraction band Optimal.From the three different sampling intervals of spectral data,the best inversion model RPD established by the three different sampling intervals is greater than 2.0,and they are all excellent models.The best inversion model established by the three different sampling intervals is the best inversion model.In the model,the optimal inversion effect of the 2nm resampled spectral data is the best.The research provides a theoretical basis for the hyperspectral monitoring of soil heavy metal chromium content in Tangshan iron tailings area.Figure 42;Table 28;Reference 57...