Research on soil Ni elements inversion based on VIS-NIR spectroscopy

Traditional soil Ni detection methods are time-consuming, labor-intensive and expensive, and can’t achieve large-scale, real-time and dynamic monitoring. To solve the above problems, this study proposed a hyperspectral remote sensing inversion method based on Principal component analysis (PCA) and Multi-scale geographical weighted regression (MGWR), in order to achieve rapid and accurate detection of soil Ni in large areas. In this study, the soil spectrum was transformed by First Derivative (FD), and the sensitive bands of soil Ni were extracted by PCA method. Two principal components were extracted as independent variables and input into Ordinary Least Squares (OLS), Geographically Weighted Regression (GWR) and MGWR models for modeling and inversion. To this end, a total of 78 sampling points were selected in Longkou City. The results showed that (1) FD transform can effectively enhance the sensitivity of soil characteristic bands. (2) MGWR model was better than GWR and OLS, the R², AIC and RMSE of MGWR were 0.536, 187.643 and 0.172, respectively.