Inversion of non-optical water quality parameters of hyperspectral remote sensing based on LBFGS-accelerated multi-layer perceptron network
Water is the source of life,the foundation of survival,a necessity for production,and the basis of ecology.However,under the dual pressures of human activities and climate change,aquatic ecosystems are facing increasingly severe challenges,particularly the serious problem of water pollution,which directly threatens the physical and mental health of residents.Water quality monitoring plays a crucial role in water pollution control,which precisely evaluates the health of water bodies and promptly adjusts control strategies,ensuring the stability and health of water environmental quality.Hyperspectral remote sensing exhibits significant potential in water quality monitoring.With the rapid development of Unmanned Aerial Vehicles(UAVs)and hyperspectral technology,UAVs equipped with hyperspectral sensors have considerably improved in terms of spectral and spatial resolutions.Accordingly,water quality parameter inversion by using hyperspectral remote sensing has gradually become a research hotspot.However,current research predominantly focuses on optical water quality parameters,with relatively less emphasis on nonoptical parameters,which also reflect the effect of human activities on water bodies.In this study,an urban river in a certain village in Guangdong Province is selected as the study area,.An experiment that involves UAV for hyperspectral remote sensing image acquisition and simultaneous water sample collection is conducted.Then,we propose a multilayer perceptron(MLP)network model accelerated by the limited-memory Broyden-Fletcher-Goldfarb-Shanno(LBFGS)method,called LBFGS-MLP,for the inversion of nonoptical water quality parameters.The parameters include Total Phosphorus(TP),Total Nitrogen(TN),and ammonia nitrogen(NH3-N),which are important indicators for measuring the nutritional status of water bodies.Through Pearson correlation analysis,spectral bands related to the three nonoptical water quality parameters(TP,TN,and NH3-N)are selected as model input.Subsequently,on the basis of exploring the effect of different network depths and optimization algorithms on model performance,the LBFGS optimization algorithm is adopted to accelerate the MLP network,and the loss function is mean squared error.Finally,the LBFGS-MLP model is applied to spatially analyze the concentrations of TP,TN,and NH3-N in the study area.Overall,the LBFGS-MLP model demonstrates significantly better accuracy on the training and testing datasets for the concentrations of TP,TN,and NH3-N compared with the random forest,CatBoost,and XGBoost models,particularly in the inversion of TN and NH3-N concentrations.The model's coefficients of determination are 0.71,0.82,and 0.72,while the mean absolute errors are 0.0118,0.0394,and 0.0601 mg/L,respectively.The concentrations of TP,TN,and NH3-N in the study area are mostly distributed at 0.1-0.3,2-5,and 0.1-0.4 mg/L,respectively,which are consistent with the survey results.Through this study,the effectiveness of the MLP algorithm in the inversion of nonoptical water quality parameters is verified,providing a theoretical basis and reference for a more comprehensive assessment of urban river water body condition.
non-optical water quality parametersmachine learninghyperspectral remote sensingconcentration inversion
万方数据
维普
