Application of multispectral remote sensing technology in water quality monitoring

Abstract

In order to understand the water quality of the Qinghe Reservoir in a timely and accurate manner,multi-spectral remote sensing technology is used to monitor the water quality. Using the OLI dataof Landsat satellite, the correlations of single-band and band combination with the chlorophyllconcentration and total suspended solids concentration of OLI data are analyzed by SPSS software. Thelargest correlation coefficient is selected to construct the ratio linear regression model and the nonlinearleast squares support vector machine (LS-SVM). The two models are used to study the multi-spectralremote sensing inversion of chlorophyll a and total suspended solids in Qinghe Reservoir. The resultsshow that compared with the ratio linear regression model, the LS-SVM model increases the decisioncoefficient R2 of the predicted and actual chlorophyll a from 0.635 to 0.966, and the root meansquare error decreases from 4.83 to 2.67; the coefficient R2 of the predicted and actual suspended solidsconcentration is increased from 0.686 to 0.88, and the average relative error is reduced from 3.52% to3.16%. The accuracy of multispectral remote sensing inversion of the concentration of chlorophyll aand total suspended solids is significantly improved by LS-SVM model.