A water quality monitoring method and system based on multi-source data fusion

Abstract

The present application provides a water quality monitoring method and system based on multi-source data fusion, which is used to solve the problem of low monitoring accuracy caused by the existing high turbidity masking the spectral reflectance of other water quality parameters. The steps are: obtaining the water parameter data of the water body to be monitored and its corresponding spectral data from historical data, wherein the water parameter data includes turbidity and water quality concentration data; constructing a turbidity inversion model and a water quality concentration inversion model at low turbidity respectively; calculating the optimal parameters of the turbidity inversion model and the water quality concentration inversion model through all turbidity data, low turbidity water quality concentration data and spectral data; calculating the model correction parameters of water quality concentrations at different turbidities at high turbidity through the inversion model; collecting the characteristic spectral data of water body parameters in real time, calculating the model prediction values ​​of water body parameters and correcting them. The present application improves the accuracy of water quality concentration monitoring by constructing correction parameters of other water quality concentrations under high turbidity.

Description

Technical Field

[0001] The present invention relates to the field of water quality monitoring, and in particular to a water quality monitoring method and system based on multi-source data fusion. Background Art

[0002] With the rapid economic growth and accelerated urbanization, water pollution has become increasingly serious. The watershed water ecological environment monitoring system is large and decentralized. Traditional water quality monitoring methods mainly rely on manual sampling and laboratory analysis. This method has obvious limitations, including limited coverage, time-consuming and labor-intensive operations, and difficulty in achieving continuous and extensive monitoring. Especially in complex terrain conditions such as mountains or hills, field investigations and sample collection are extremely challenging. In recent years, with the development of spectral detection technology, spectral detection technology has been widely used in the field of water quality detection.

[0003] Spectral detection of water quality is to collect the reflectance spectra of several water quality parameters through ground-based spectral equipment, unmanned aerial vehicle spectral equipment and hyperspectral satellites, and construct a mathematical model of the reflectance spectra and water quality parameters to detect the water quality. However, in the existing detection process, high concentration of turbidity will mask the reflectance spectra of other water quality parameters, resulting in low detection accuracy of other water quality parameters. Summary of the invention

[0004] The purpose of the present invention is to provide a water quality monitoring method and system based on multi-source data fusion, which is used to solve the technical problem that high turbidity masks the spectral reflectance of chlorophyll a concentration and phycocyanin concentration in the existing water quality monitoring method based on spectral data, resulting in low monitoring accuracy of chlorophyll a concentration and phycocyanin concentration.

[0005] A water quality monitoring method based on multi-source data fusion, the specific steps are:

[0006] S1: Acquire water body parameter data of the water body to be monitored and its corresponding spectral data from historical data, wherein the water body parameter data includes turbidity data and water quality concentration data;

[0007] S2: Select characteristic spectra of several water parameters respectively, and construct the inversion model of turbidity with respect to its characteristic spectrum, and the inversion model of water quality concentration with respect to its characteristic spectrum at low turbidity;

[0008] S3: Calculate the optimal parameters of the turbidity inversion model through the turbidity data and its characteristic spectrum data, and calculate the optimal parameters of the water quality concentration inversion model through the water quality concentration data at low turbidity and its characteristic spectrum data;

[0009] S4: Calculate the model prediction value of water quality concentration at high turbidity through the inversion model, and calculate the model correction parameters of water quality concentration at different turbidities at high turbidity through the model prediction value and the actual collection value;

[0010] S5: Collect characteristic spectral data of water body parameters in real time, calculate model prediction values ​​of water body parameters through inversion model, and select model correction parameters through turbidity data to correct and output water quality concentration data.

[0011] Optionally, the water quality concentration in step S1 includes chlorophyll a concentration and phycocyanin concentration, the spectral data includes spectral reflectance data of characteristic spectra, and the water parameter data of the water body to be monitored is high-precision data obtained through manual sampling and analysis to construct a water parameter data set N :

[0012] ,in: SS i For the i The turbidity data of the subsamples, Chal i For the i Chlorophyll a concentration data of the sampling times, PC i For the i The phycocyanin concentration data of the sampling time, n is the total number of samples in the dataset, .

[0013] Optionally, in step S2, a characteristic spectrum of turbidity is selected, and an inversion model of turbidity with respect to its characteristic spectrum is constructed as follows:

[0014] ;

[0015] In the formula, SS is turbidity, and is the characteristic spectral band of turbidity, a and b is the turbidity inversion model parameter, for The spectral reflectance, for spectral reflectance.

[0016] Optionally, in step S2, a characteristic spectrum of chlorophyll a concentration is selected to construct an inversion model of chlorophyll a concentration with respect to the characteristic spectrum:

[0017] ;

[0018] In the formula, Chla is the chlorophyll a concentration, and is the characteristic spectral band of chlorophyll a concentration, c and d is the chlorophyll a concentration inversion model parameter, for The spectral reflectance, for spectral reflectance.

[0019] Optionally, in step S2, a characteristic spectrum of phycocyanin concentration is selected to construct an inversion model of phycocyanin concentration with respect to its characteristic spectrum:

[0020] ;

[0021] In the formula, PC is the concentration of phycocyanin, , and is the characteristic spectral band of phycocyanin concentration, e and f is the inversion model parameter of phycocyanin concentration, for The spectral reflectance, for The spectral reflectance, for spectral reflectance.

[0022] Optionally, the specific steps of calculating the optimal parameters of the model in step S3 are:

[0023] S3.1: Through water body parameter dataset N The turbidity data and its corresponding spectral data are used to calculate the parameters in the turbidity inversion model using the least squares method. a and b The optimal value of

[0024] S3.2: From water body parameter datasets N Select turbidity SS≤S 1 All data subsets of the parameter inversion set are constructed N' , through the parameter inversion set N' The chlorophyll a concentration data and its corresponding spectral data in the chlorophyll a concentration inversion model are calculated using the least squares method. c and d The optimal value of

[0025] S3.3: Inversion Sets by Parameters N' The phycocyanin concentration data and its corresponding spectral data are used to calculate the parameters in the phycocyanin concentration inversion model using the least squares method. e and f The optimal value of .

[0026] Optionally, the specific steps of calculating the model correction parameters of water quality concentrations under different turbidities in step S4 are:

[0027] S4.1: From the water body parameter dataset N Select turbidity SS>S 1 All data subsets are used to construct parameter correction sets N" , the parameter correction set N" The spectral data corresponding to the water quality concentration data is brought into the water quality concentration inversion model to calculate the predicted value of water quality concentration. y , y is one of the predicted value of chlorophyll a concentration or the predicted value of phycocyanin concentration;

[0028] S4.2: Predicted values ​​of water quality concentrations y and parameter correction set N" The collected values ​​are used to calculate the prediction accuracy ;

[0029] ;

[0030] Where: h The chlorophyll a concentration or phycocyanin concentration is collected, and the water quality concentration accuracy data set at high turbidity is obtained through the above calculation M :

[0031] ,in: The parameter correction set N" Middle j The prediction accuracy of water quality concentration data is SS j The parameter correction set N" Middle j The turbidity data corresponding to the water quality concentration data, m Water quality concentration accuracy dataset M The total number of data, ;

[0032] S4.3: Taking the prediction accuracy of water quality concentration as the goal, the K-nearest neighbor clustering algorithm is used to cluster the water quality concentration accuracy dataset M The data in is divided into several data subsets;

[0033] S4.4: Calculate the mean turbidity value and water quality concentration accuracy correction parameter for each data subset;

[0034] ;

[0035] ;

[0036] In the formula, is the mean turbidity of the t-th data subset, is the water quality concentration accuracy correction parameter for the tth data subset, is the first g The parameter concentration prediction accuracy value of the data, is the first g The turbidity value of the data, q is the total number of data in the tth data subset, 。

[0037] Optionally, in step S5, the specific steps of collecting characteristic spectral data of water body parameters in real time, calculating model prediction values ​​of water body parameters through inversion models, and correcting and outputting water quality concentration data by selecting model correction parameters through turbidity data are as follows:

[0038] S5.1: collecting characteristic spectral data of turbidity and water quality concentration of the water body to be monitored at the current moment in real time, and calculating the turbidity data of the entire water body to be monitored through a turbidity inversion model, wherein the characteristic spectral data of water quality concentration includes characteristic spectral data of chlorophyll a concentration and characteristic spectral data of phycocyanin concentration;

[0039] S5.2: Input the characteristic spectrum data of water quality concentration collected in step S4.1 into the water quality concentration inversion model to calculate the model predicted value of water quality concentration at the current moment;

[0040] S5.3: Matching the turbidity data of the sampling location to the data subset in step S4, and correcting the model prediction value by the water quality concentration accuracy correction parameter of the data subset;

[0041] The product of the correction parameter and the model prediction value at the current moment is used as the water quality concentration value output at the sampling location at the current moment, and S5.3 is repeated to realize the water parameter concentration monitoring of the entire water body to be monitored.

[0042] A water quality monitoring system based on multi-source data fusion, used to implement the above-mentioned water quality monitoring method based on multi-source data fusion, comprising:

[0043] Data acquisition module: used to obtain water body parameter data and corresponding spectral data of the water body to be monitored from historical data, wherein the water body parameter data includes turbidity data and water quality concentration data;

[0044] Inversion model building module: used to select characteristic spectra of several water parameters, respectively, to build the inversion model of turbidity with respect to its characteristic spectrum, and the inversion model of water quality concentration with respect to its characteristic spectrum at low turbidity;

[0045] Model parameter calculation module: used to calculate the optimal parameters of the turbidity inversion model through turbidity data and its characteristic spectrum data, and to calculate the optimal parameters of the water quality concentration inversion model through water quality concentration data at low turbidity and its characteristic spectrum data;

[0046] Correction parameter calculation module: used to calculate the model prediction value of water quality concentration when the turbidity is high through the inversion model, and calculate the model correction parameters of water quality concentration under different turbidities when the turbidity is high through the model prediction value and the actual collection value;

[0047] Water body parameter calculation module: used to collect characteristic spectral data of water body parameters in real time, calculate the model prediction value of water body parameters through the inversion model, and select model correction parameters through turbidity data to correct and output the water quality concentration data.

[0048] Due to the adoption of the above technical solution, the present invention has the following advantages:

[0049] The present application fuses the turbidity, chlorophyll a concentration and phycocyanin concentration data, calculates the parameters of the inversion model through the chlorophyll a concentration and phycocyanin concentration data under low turbidity and their spectral data, and calculates the correction parameters through the chlorophyll a concentration and phycocyanin concentration data under high turbidity and their spectral data, thereby reducing the influence of high turbidity on the spectral reflectance masking of chlorophyll a concentration and phycocyanin concentration and improving the monitoring accuracy of chlorophyll a concentration and phycocyanin concentration.

[0050] Other advantages, objectives and features of the present invention will be described in the following description to some extent, and to some extent, will be obvious to those skilled in the art based on the following examination and study, or can be taught from the practice of the present invention. The objectives and other advantages of the present invention can be realized and obtained through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The accompanying drawings of the present invention are as follows.

[0052] Figure 1 The present invention is a flow chart of a water quality monitoring method based on multi-source data fusion.

[0053] Figure 2 This is the inversion model diagram of turbidity of the present invention.

[0054] Figure 3 This is the inversion model diagram of chlorophyll a concentration of the present invention.

[0055] Figure 4 This is the inversion model diagram of the phycocyanin concentration of the present invention.

[0056] Figure 5 This is the inversion diagram of turbidity in the simulation verification of the present invention.

[0057] Figure 6 This is an inversion diagram of chlorophyll a concentration in the simulation verification of the present invention.

[0058] Figure 7 This is an inversion diagram of the phycocyanin concentration in the simulation verification of the present invention. DETAILED DESCRIPTION

[0059] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0060] Embodiment 1:

[0061] like Figure 1 A water quality monitoring method based on multi-source data fusion is shown, and the specific steps are as follows:

[0062] S1: Acquire water body parameter data of a water body to be monitored and its corresponding spectral data from historical data, wherein the water body parameter data includes turbidity data and water quality concentration data.

[0063] In this embodiment, the water quality concentration in step S1 includes chlorophyll a concentration and phycocyanin concentration, the spectral data includes spectral reflectance data of characteristic spectrum, and the water parameter data of the water body to be monitored is high-precision data obtained through manual sampling and analysis, and a water parameter data set is constructed. N :

[0064] ,in: SS i For the i The turbidity data of the subsamples, Chal i For the i Chlorophyll a concentration data of the sampling times, PC i For the i The phycocyanin concentration data of the sampling time, n is the total number of samples in the dataset, .

[0065] In this embodiment, the turbidity, chlorophyll a concentration and phycocyanin concentration in the water body parameter data are all high-precision data obtained by manually using an analyzer, and the chlorophyll a concentration and phycocyanin concentration are not affected by the turbidity concentration, but the scope or frequency of manual collection is limited. Therefore, it is necessary to monitor the entire water body to be monitored through spectral data. In this embodiment, the spectral data is based on multispectral data collected by hyperspectral satellites or multispectral data collected by drones, and supplementary sampling is performed by setting up ground-based spectral sampling equipment in places where hyperspectral satellites or drones are difficult to collect, and the complete multispectral data of the water area to be monitored is obtained by fusing the multispectral data collected by hyperspectral satellites or drones with the spectral data sampled by ground-based spectral sampling equipment.

[0066] S2: Select characteristic spectra of several water parameters respectively, and construct the inversion model of turbidity with respect to its characteristic spectrum, and the inversion model of water quality concentration with respect to its characteristic spectrum at low turbidity. The specific steps are as follows:

[0067] S2.1: Select the characteristic spectrum of turbidity and construct the inversion model of turbidity with respect to its characteristic spectrum:

[0068] ;

[0069] In the formula, SS is turbidity, and is the characteristic spectral band of turbidity, a and b is the turbidity inversion model parameter, for The spectral reflectance, for Spectral reflectance;

[0070] S2.2: Select the characteristic spectrum of chlorophyll a concentration and construct the inversion model of chlorophyll a concentration with respect to the characteristic spectrum:

[0071] ;

[0072] In the formula, Chla is the chlorophyll a concentration, and is the characteristic spectral band of chlorophyll a concentration, c and d is the chlorophyll a concentration inversion model parameter, for The spectral reflectance, for Spectral reflectance;

[0073] S2.3: Select the characteristic spectrum of phycocyanin concentration and construct the inversion model of phycocyanin concentration with respect to its characteristic spectrum:

[0074] ;

[0075] In the formula, PC is the concentration of phycocyanin, , and is the characteristic spectral band of phycocyanin concentration, e and f is the inversion model parameter of phycocyanin concentration, for The spectral reflectance, for The spectral reflectance, for spectral reflectance.

[0076] In this embodiment, the spectral characteristic band of turbidity and The wavelengths of 640nm and 760nm are the characteristic wavelengths of chlorophyll a concentration. and The wavelengths of phycocyanin at 645 nm and 850 nm are the characteristic wavelengths of phycocyanin concentration. , and They are 600nm, 650nm and 770nm respectively.

[0077] S3: Calculate the optimal parameters of the turbidity inversion model through the turbidity data and its characteristic spectrum data, and calculate the optimal parameters of the water quality concentration inversion model through the water quality concentration data at low turbidity and its characteristic spectrum data. The specific steps are:

[0078] S3.1: Through water body parameter dataset N The turbidity data and its corresponding spectral data are used to calculate the parameters in the turbidity inversion model using the least squares method. a and b The optimal value of

[0079] S3.2: From water body parameter datasets N Select turbidity SS≤S 1 All data subsets of the parameter inversion set are constructed N' , through the parameter inversion set N' The chlorophyll a concentration data and its corresponding spectral data in the chlorophyll a concentration inversion model are calculated using the least squares method. c and d The optimal value of

[0080] S3.3: Inversion Sets by Parameters N' The phycocyanin concentration data and its corresponding spectral data are used to calculate the parameters in the phycocyanin concentration inversion model using the least squares method. e and f The optimal value of .

[0081] In this embodiment, the turbidity parameter a and b In solving, As a parameter, the least square method is used to solve the best fitting straight line between it and turbidity SS. c and d In solving, As a parameter, the least square method is used to solve its relationship with chlorophyll a concentration ChlaThe best fitting straight line. e and f In solving, As a parameter, the least square method was used to solve its relationship with phycocyanin concentration. PC The best fitting straight line.

[0082] In this embodiment, the turbidity threshold S 1 The data sets were divided into those affected by turbidity and those not affected by turbidity. The parameter inversion set whose characteristic spectra of chlorophyll a concentration and phycocyanin concentration were not affected by turbidity was selected. N' , the parameters of the chlorophyll a concentration inversion model and the phycocyanin concentration inversion model were fitted, and the fitted inversion model had high accuracy.

[0083] S4: Calculate the model prediction value of water quality concentration when the turbidity is high through the inversion model, and calculate the model correction parameters of water quality concentration under different turbidity when the turbidity is high through the model prediction value and the actual collection value. The specific steps are:

[0084] S4.1: From the water body parameter dataset N Select turbidity SS>S 1 All data subsets are used to construct parameter correction sets N" , the parameter correction set N" The spectral data corresponding to the water quality concentration data is brought into the water quality concentration inversion model to calculate the predicted value of water quality concentration. y , y is one of the predicted value of chlorophyll a concentration or the predicted value of phycocyanin concentration;

[0085] S4.2: Predicted values ​​of water quality concentrations y and parameter correction set N" The collected values ​​are used to calculate the prediction accuracy ;

[0086] ;

[0087] Where: h The chlorophyll a concentration or phycocyanin concentration is collected, and the water quality concentration accuracy data set at high turbidity is obtained through the above calculation M :

[0088] ,in: The parameter correction set N" Middle j The prediction accuracy of water quality concentration data is SS j The parameter correction set N" Middle jThe turbidity data corresponding to the water quality concentration data, m Water quality concentration accuracy dataset M The total number of data, ;

[0089] S4.3: Taking the prediction accuracy of water quality concentration as the goal, the K-nearest neighbor clustering algorithm is used to cluster the water quality concentration accuracy dataset M The data in is divided into several data subsets;

[0090] S4.4: Calculate the mean turbidity value and water quality concentration accuracy correction parameter for each data subset;

[0091] ;

[0092] ;

[0093] In the formula, is the mean turbidity of the t-th data subset, is the water quality concentration accuracy correction parameter for the tth data subset, is the first g The parameter concentration prediction accuracy value of the data, is the first g The turbidity value of the data, q is the total number of data in the tth data subset, 。

[0094] In this embodiment, the chlorophyll a concentration or the phycocyanin concentration is calculated through steps S3.1-S3.4 to obtain the turbidity mean and accuracy correction parameters of multiple data subsets of chlorophyll a concentration, and the turbidity mean and accuracy correction parameters of multiple data subsets of phycocyanin concentration, respectively.

[0095] In this embodiment, by selecting the parameter correction set N" The accuracy parameters of chlorophyll a concentration and phycocyanin concentration under different turbidities were calculated. The accuracy parameters reflected the influence of different turbidities on the characteristic spectral reflectance masking of chlorophyll a concentration and phycocyanin concentration. The accuracy correction parameters could better reflect the chlorophyll a concentration and phycocyanin concentration under high turbidity.

[0096] S5: Collect characteristic spectral data of water body parameters in real time, calculate the model prediction value of water body parameters through the inversion model, and select model correction parameters through turbidity data to correct and output the water quality concentration data. The specific steps are as follows:

[0097] S5.1: collecting characteristic spectral data of turbidity and water quality concentration of the water body to be monitored at the current moment in real time, and calculating the turbidity data of the entire water body to be monitored through a turbidity inversion model, wherein the characteristic spectral data of water quality concentration includes characteristic spectral data of chlorophyll a concentration and characteristic spectral data of phycocyanin concentration;

[0098] S5.2: Input the characteristic spectrum data of water quality concentration collected in step S4.1 into the water quality concentration inversion model to calculate the model predicted value of water quality concentration at the current moment;

[0099] S5.3: Matching the turbidity data of the sampling location to the data subset in step S4, and correcting the model prediction value by the water quality concentration accuracy correction parameter of the data subset;

[0100] The product of the correction parameter and the model prediction value at the current moment is used as the water quality concentration value output at the sampling location at the current moment, and S5.3 is repeated to realize the water parameter concentration monitoring of the entire water body to be monitored.

[0101] In this embodiment, steps S5.1-S5.3 are repeated to correct the predicted values ​​of chlorophyll a concentration or phycocyanin concentration respectively.

[0102] S6: A tributary in the southwest region was used as the target water area for simulation and verification. The historical data of the target water area was collected, and the parameters of the inversion model of turbidity, chlorophyll a concentration, and phycocyanin concentration were fitted. The optimal parameters of turbidity were obtained by the least squares method. a and b are 34.56 and -1.08 respectively. The inversion model is as follows: Figure 2 As shown; the optimal parameters of chlorophyll a concentration were obtained by the least squares method c and d are 47.83 and 22.23 respectively. The inversion model is as follows: Figure 3 As shown; the optimal parameters of phycocyanin concentration were obtained by the least squares method e and f are 2.51 and -0.28 respectively. The inversion model is as follows: Figure 4 shown.

[0103] Setting turbidity threshold S 1 The turbidity, chlorophyll a concentration and phycocyanin concentration of the target water area are obtained by using steps S3-S5, and the inversion images are drawn as follows: Figure 5 , Figure 6 and Figure 7 shown.

[0104] Embodiment 2:

[0105] A water quality monitoring system based on multi-source data fusion, used to implement a water quality monitoring method based on multi-source data fusion described in Example 1, comprising:

[0106] Data acquisition module: used to obtain water body parameter data and corresponding spectral data of the water body to be monitored from historical data, wherein the water body parameter data includes turbidity data and water quality concentration data;

[0107] Inversion model building module: used to select characteristic spectra of several water parameters, respectively, to build the inversion model of turbidity with respect to its characteristic spectrum, and the inversion model of water quality concentration with respect to its characteristic spectrum at low turbidity;

[0108] Model parameter calculation module: used to calculate the optimal parameters of the turbidity inversion model through turbidity data and its characteristic spectrum data, and to calculate the optimal parameters of the water quality concentration inversion model through water quality concentration data at low turbidity and its characteristic spectrum data;

[0109] Correction parameter calculation module: used to calculate the model prediction value of water quality concentration when the turbidity is high through the inversion model, and calculate the model correction parameters of water quality concentration under different turbidities when the turbidity is high through the model prediction value and the actual collection value;

[0110] Water body parameter calculation module: used to collect characteristic spectral data of water body parameters in real time, calculate the model prediction value of water body parameters through the inversion model, and select model correction parameters through turbidity data to correct and output the water quality concentration data.

[0111] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A water quality monitoring method based on multi-source data fusion, characterized in that: The specific steps are: S1: Acquire water body parameter data of the water body to be monitored and its corresponding spectral data from historical data, wherein the water body parameter data includes turbidity data and water quality concentration data; S2: Select characteristic spectra of several water parameters respectively, and construct the inversion model of turbidity with respect to its characteristic spectrum, and the inversion model of water quality concentration with respect to its characteristic spectrum at low turbidity; S3: Calculate the optimal parameters of the turbidity inversion model through the turbidity data and its characteristic spectrum data, and calculate the optimal parameters of the water quality concentration inversion model through the water quality concentration data at low turbidity and its characteristic spectrum data; S4: Calculate the model prediction value of water quality concentration at high turbidity through the inversion model, and calculate the model correction parameters of water quality concentration at different turbidities at high turbidity through the model prediction value and the actual collection value; S5: collect characteristic spectral data of water body parameters in real time, calculate the model prediction value of water body parameters through the inversion model, and select model correction parameters through turbidity data to correct and output the water quality concentration data; In step S2, the characteristic spectrum of turbidity is selected, and the inversion model of turbidity with respect to its characteristic spectrum is constructed as follows: ； In the formula, SS is turbidity, and is the characteristic spectral band of turbidity, a and b is the turbidity inversion model parameter, for The spectral reflectance, for spectral reflectance.

2. The water quality monitoring method based on multi-source data fusion according to claim 1 is characterized in that: The water quality concentration in step S1 includes chlorophyll a concentration and phycocyanin concentration, the spectral data includes spectral reflectance data of characteristic spectrum, and the water parameter data of the water body to be monitored is high-precision data obtained through manual sampling and analysis, and a water parameter data set is constructed. N : ,in: SS i For the i The turbidity data of the subsamples, Chal i For the i Chlorophyll a concentration data of the sampling times, PC i For the i The phycocyanin concentration data of the sampling time, n is the total number of samples in the dataset, .

3. The water quality monitoring method based on multi-source data fusion according to claim 2 is characterized in that: In step S2, the characteristic spectrum of chlorophyll a concentration is selected, and the inversion model of chlorophyll a concentration with respect to the characteristic spectrum is constructed as follows: ； In the formula, Chla is the chlorophyll a concentration, and is the characteristic spectral band of chlorophyll a concentration, c and d is the chlorophyll a concentration inversion model parameter, for The spectral reflectance, for spectral reflectance.

4. The water quality monitoring method based on multi-source data fusion according to claim 3 is characterized in that: In step S2, a characteristic spectrum of phycocyanin concentration is selected, and an inversion model of phycocyanin concentration with respect to its characteristic spectrum is constructed as follows: ； In the formula, PC is the concentration of phycocyanin, , and is the characteristic spectral band of phycocyanin concentration, e and f is the inversion model parameter of phycocyanin concentration, for The spectral reflectance, for The spectral reflectance, for spectral reflectance.

5. The water quality monitoring method based on multi-source data fusion according to claim 4 is characterized in that: The specific steps for calculating the optimal parameters of the model in step S3 are: S3.1: Through water body parameter dataset N The turbidity data and its corresponding spectral data are used to calculate the parameters in the turbidity inversion model using the least squares method. a and b The optimal value of S3.2: From water body parameter datasets N Select turbidity SS≤S 1 Construct parameter inversion sets from all data subsets N' , through the parameter inversion set N' The chlorophyll a concentration data and its corresponding spectral data in the chlorophyll a concentration inversion model are calculated using the least squares method. c and d The optimal value of S3.3: Inversion Sets by Parameters N' The phycocyanin concentration data and its corresponding spectral data are used to calculate the parameters in the phycocyanin concentration inversion model using the least squares method. e and f The optimal value of .

6. The water quality monitoring method based on multi-source data fusion according to claim 5 is characterized in that: The specific steps of calculating the model correction parameters of water quality concentration under different turbidity in step S4 are: S4.1: From the water body parameter dataset N Select turbidity SS>S 1 All data subsets are used to construct parameter correction sets N" , the parameter correction set N" The spectral data corresponding to the water quality concentration data is brought into the water quality concentration inversion model to calculate the predicted value of water quality concentration. y , y is one of the predicted value of chlorophyll a concentration or the predicted value of phycocyanin concentration; S4.2: Predicted values ​​of water quality concentrations y and parameter correction set N" The collected values ​​are used to calculate the prediction accuracy ; ； Where: h The chlorophyll a concentration or phycocyanin concentration is collected, and the water quality concentration accuracy data set at high turbidity is obtained through the above calculation M : ,in: The parameter correction set N" Middle j The prediction accuracy of water quality concentration data is SS j The parameter correction set N" Middle j The turbidity data corresponding to the water quality concentration data, m Water quality concentration accuracy dataset M The total number of data, ; S4.3: Taking the prediction accuracy of water quality concentration as the goal, the K-nearest neighbor clustering algorithm is used to cluster the water quality concentration accuracy dataset M The data in is divided into several data subsets; S4.4: Calculate the mean turbidity value and water quality concentration accuracy correction parameter for each data subset; ； ； In the formula, is the mean turbidity of the t-th data subset, is the water quality concentration accuracy correction parameter for the tth data subset, is the first g The parameter concentration prediction accuracy value of the data, is the first g The turbidity value of the data, q is the total number of data in the tth data subset, 。 7. The water quality monitoring method based on multi-source data fusion according to claim 6 is characterized in that: In step S5, the characteristic spectrum data of water parameters are collected in real time, the model prediction values ​​of water parameters are calculated by the inversion model, and the water quality concentration data are corrected and output by selecting the model correction parameters through the turbidity data. The specific steps are: S5.1: collecting characteristic spectral data of turbidity and water quality concentration of the water body to be monitored at the current moment in real time, and calculating the turbidity data of the entire water body to be monitored through a turbidity inversion model, wherein the characteristic spectral data of water quality concentration includes characteristic spectral data of chlorophyll a concentration and characteristic spectral data of phycocyanin concentration; S5.2: Input the characteristic spectrum data of water quality concentration collected in step S4.1 into the water quality concentration inversion model to calculate the model predicted value of water quality concentration at the current moment; S5.3: Matching the turbidity data of the sampling location to the data subset in step S4, and correcting the model prediction value by the water quality concentration accuracy correction parameter of the data subset; The product of the correction parameter and the model prediction value at the current moment is used as the water quality concentration value output at the sampling location at the current moment, and S5.3 is repeated to realize the water parameter concentration monitoring of the entire water body to be monitored.

8. A water quality monitoring system based on multi-source data fusion, characterized in that: A method for monitoring water quality based on multi-source data fusion for implementing any one of claims 1 to 7, comprising: Data acquisition module: used to obtain water body parameter data and corresponding spectral data of the water body to be monitored from historical data, wherein the water body parameter data includes turbidity data and water quality concentration data; Inversion model building module: used to select characteristic spectra of several water parameters, respectively, to build the inversion model of turbidity with respect to its characteristic spectrum, and the inversion model of water quality concentration with respect to its characteristic spectrum at low turbidity; Model parameter calculation module: used to calculate the optimal parameters of the turbidity inversion model through turbidity data and its characteristic spectrum data, and to calculate the optimal parameters of the water quality concentration inversion model through water quality concentration data at low turbidity and its characteristic spectrum data; Correction parameter calculation module: used to calculate the model prediction value of water quality concentration when the turbidity is high through the inversion model, and calculate the model correction parameters of water quality concentration under different turbidities when the turbidity is high through the model prediction value and the actual collection value; Water body parameter calculation module: used to collect characteristic spectral data of water body parameters in real time, calculate the model prediction value of water body parameters through the inversion model, and select model correction parameters through turbidity data to correct and output the water quality concentration data.

Images