Research On Inversion Method Of Nitrogen And Phosphorus Content Based On UAV Hyperspectral Remote Sensing

The study of total nitrogen and total phosphorus content in water has always been a hot issue in society.The use of large amounts of nitrogen and phosphorus fertilizers in industry and agriculture,the waste water containing nitrogen and phosphorus,and the random discharge of domestic sewage have led to the eutrophication of inland rivers and offshore estuaries,breaking the ecological balance.Without timely management,water quality will deteriorate sharply and evolve into red tides and blooms,causing huge economic losses.Therefore,it is extremely necessary and urgent to control the pollution of nitrogen and phosphorus in water bodies.Conventional monitoring methods for nitrogen and phosphorus require the manual collection of water samples for analysis to obtain nitrogen and phosphorus concentration data,which is not only time-consuming and laborious,but also cannot achieve large-scale water quality monitoring.The new type of water quality remote sensing monitoring technology not only reduces labor costs but also monitors large-scale water quality in real time.It is an important development direction of water quality monitoring.Therefore,it is of great significance to study the remote sensing monitoring method of nitrogen and phosphorus in water.In this paper,based on the characteristics of high spectral resolution of UAV hyperspectral remote sensing data and complex band information,this paper studies the improvement method of the inversion model of total nitrogen and total phosphorus content to achieve the improvement of inversion accuracy,as follows:(1)A saliency band selection method based on signal matching is proposed.This method regards the spectral reflectance information of different wave bands as background information,and the concentration vector of total nitrogen and total phosphorus as the target vector.First,the water quality is selected from the background information with high information redundancy by constrained energy minimization.The bands with the most sensitive parameter changes are combined with the orthogonal principle in the orthogonal subspace projection algorithm to suppress the influence of the adjacent bands.Iterative selection is performed to select the most sensitive band set that best reflects the changes in nitrogen and phosphorus concentrations.Finally,through the regression analysis of multiple band equations through the selected band set,a remote sensing inversion model is established.(2)A model inversion method based on solving underdetermined equations is proposed.The spectral data of each sampling point and the corresponding water quality parameters are regarded as an equation.Multiple data sampling points form a set of underdetermined equations.Using QR Decompose the matrix method to obtain the minimum number of non-zero solutions of the underdetermined equations,get the sensitive band by the solution value,combine the nonlinear function to expand the sensitive band,and continue to form an expanded data with the corresponding water quality parameters.A new underdetermined equation system,through the solution of the underdetermined equation system to establish a remote sensing inversion model.In order to verify the performance of the above two remote sensing model inversion methods,this paper takes the Liaohe River estuary as the research object,designs experiments to obtain UAV hyperspectral remote sensing data,inverts the total nitrogen and total phosphorus content,and compares with traditional remote sensing methods The accuracy of the inversion is compared to prove the effectiveness of the two inversion methods proposed in this paper.