River and lake water quality information Internet of Things sensing system and control method

Abstract

The invention discloses a river and lake water quality information internet-of-things sensing system and a control method. A miniature water quality automatic monitoring station comprises a flow cell and a control unit; the flow cell collects water and conveys the water to the spectrum analyzer; the spectrum analyzer is electrically connected with the control unit; the control unit is electrically connected with the communication module; the controller controls the spectrum analyzer to regularly detect water quality to obtain spectrum monitoring data and send the spectrum monitoring data to the server; a water quality correlation model and a fitting error model are preset in the server, the spectrum monitoring data are input into the water quality correlation model to obtain a water quality evaluation initial value, and the spectrum monitoring data are input into the fitting error model to obtain a water quality evaluation error; superposing the water quality evaluation initial value and the water quality evaluation error to obtain a river and lake water quality sensing result; the water quality and the ecological environment of rivers and lakes are accurately monitored.

Description

Technical Field

[0001] The present invention belongs to the field of river and lake water quality monitoring design, and specifically relates to a river and lake water quality information Internet of Things perception system and a control method. Background Art

[0002] As residents pay more and more attention to environmental issues closely related to their health, the monitoring and analysis of the ecological environment of river and lake water is becoming more and more important. At present, the regular monitoring of the water ecological environment of rivers, lakes, reservoirs, ponds, etc. generally adopts the method of manual sampling and testing, which is not only time-consuming, but also lacks continuous monitoring of the water quality and ecological environment in the monitoring area, and cannot meet the needs of water area emergency and management. In addition, the water quality and ecological environment monitoring samples of rivers and lakes are characterized by large data volume, high vector dimension, and many attributes. The existing water quality monitoring instruments and monitoring systems are difficult to meet actual needs. Summary of the invention

[0003] The present invention provides an Internet of Things sensing system and a control method for river and lake water quality information, which can accurately monitor the water quality and ecological environment of rivers and lakes.

[0004] In order to achieve the above object, the technical solution adopted by the present invention is: The first aspect of the present invention provides a river and lake water quality information Internet of Things perception system, including a micro water quality automatic monitoring station, a communication module and a server; a plurality of micro water quality automatic monitoring stations are connected to the server via the communication module; The micro water quality automatic monitoring station includes a circulation pool and a control unit; the circulation pool collects water and then transmits it to the spectrometer; the spectrometer is electrically connected to the control unit; the control unit is electrically connected to the communication module; The controller controls the spectrometer to regularly detect water quality, obtain spectral monitoring data, and send it to the server; the server is preset with a water quality correlation model and a fitting error model, the spectral monitoring data is input into the water quality correlation model to obtain an initial water quality evaluation value, the spectral monitoring data is input into the fitting error model to obtain a water quality evaluation error, and the initial water quality evaluation value and the water quality evaluation error are superimposed to obtain the river and lake water quality perception result.

[0005] Furthermore, a float type water sampling device is provided at the monitoring point of the micro water quality automatic monitoring station, the water intake pipe of the circulation pool is connected to the float type water sampling device, and the water inlet hole of the water intake pipe of the circulation pool is located 0.5m to 1m below the water surface.

[0006] Furthermore, the micro water quality automatic monitoring station is equipped with a sensor group electrically connected to the controller; the sensor group includes a pH sensor, a conductivity sensor, a temperature sensor, and a flow rate and flow sensor; the pH sensor measures the pH of river and lake water; the conductivity sensor measures the content of dissolved salts in river and lake water; the temperature sensor measures water temperature; the flow rate and flow sensor measures the speed and flow of river and lake water; The server calculates and obtains river and lake water quality assessment results based on pH, dissolved salt content, water temperature, nitrate nitrogen content, water flow speed and flow, and river and lake water quality perception results.

[0007] Furthermore, the server calculates and obtains river and lake water quality assessment results based on pH, dissolved salt content, water temperature, nitrate nitrogen content, water flow speed and flow rate, and river and lake water quality perception results, and the process includes: The pH, dissolved salt content, water temperature, nitrate nitrogen content, water velocity and flow rate are used as water quality monitoring parameters. The abnormal value detection range is set based on the water quality monitoring parameters. The expression formula is: ; ; ; In the formula, is the parameter value that 25% of the data points in the b-th water quality monitoring parameter set are less than or equal to; is the parameter value that 75% of the data points in the b-th water quality monitoring parameter set are less than or equal to; is the distribution range of the middle 50% of the data in the bth water quality monitoring parameter set; is the maximum value in the abnormal value detection range of the bth water quality monitoring parameter; is the minimum value in the abnormal value detection range of the bth water quality monitoring parameter; The water quality monitoring parameters are screened for outliers based on the outlier detection range; the water quality monitoring parameters are denoised, and the expression formula is: ;

[0008] In the formula, It is represented as the output data after denoising the b-th water quality monitoring parameter at time t; is the water quality monitoring parameter of the bth type at time k; is the moving average window size; t is the current time series; A linear interpolation algorithm is used to fill in the missing data in the water quality monitoring parameters, and then the evaluation values ​​of the water quality monitoring parameters and the evaluation values ​​of the river and lake water quality perception results are obtained based on the preset mapping relationship table; the evaluation values ​​of the water quality monitoring parameters and the evaluation values ​​of the river and lake water quality perception results are weighted and summed to obtain the river and lake water quality assessment results.

[0009] Furthermore, it also includes a river and lake virtual display platform, which constructs a three-dimensional scene model based on water system distribution data and terrain data; Establish a WebGL scene architecture based on the viewing angle lighting scene and web page architecture, convert and compress the 3D scene model, embed the model scene data with the WebGL scene architecture to build a river and lake virtual model; add the river and lake water quality assessment results to the river and lake virtual model; The received viewpoint coordinates and target point coordinates are input into the river and lake virtual model, and the river and lake water quality assessment results of the target point are displayed through the river and lake virtual model.

[0010] The second aspect of the present invention provides a method for sensing river and lake water quality information through the Internet of Things, comprising: Control the spectrometer to regularly detect water quality to obtain spectral monitoring data, input the spectral monitoring data into the water quality correlation model to obtain the initial value of water quality evaluation, input the spectral monitoring data into the fitting error model to obtain the water quality evaluation error, and superimpose the initial value of water quality evaluation and the water quality evaluation error to obtain the river and lake water quality perception results.

[0011] Furthermore, the water quality correlation model and fitting error model construction process includes: The water quality correlation model is obtained by coupling the water quality spectrum-turbidity correlation function, the water quality spectrum-COD correlation function and the water quality spectrum-nitrogen content correlation function. A fitting error model is constructed based on a convolutional neural network, and water quality training data is obtained from a water quality database; the water quality training data includes spectral characteristics, water turbidity, chemical oxygen demand, and water nitrogen content; the water turbidity, chemical oxygen demand, and water nitrogen content are weighted and summed to obtain a true label; The spectral features in the water quality training data are input into the water quality correlation model to obtain the water quality prediction value; the spectral features in the water quality training data are input into the fitting error model to obtain the prediction error training value, and the training loss value is calculated according to the water quality prediction value, the prediction error training value and the true label; the fitting error model is optimized according to the training loss value, and the training process of the neural convolution model is repeated until the training loss value converges to output the trained fitting error model.

[0012] Furthermore, the process of obtaining the water quality spectrum-turbidity correlation function, the water quality spectrum-COD correlation function and the water quality spectrum-nitrogen content correlation function includes: After obtaining several spectral characteristic curves affected only by turbidity, the least squares regression algorithm was used to fit the water quality turbidity characteristics to obtain the water quality spectrum-turbidity correlation function; after obtaining several spectral characteristic curves affected only by chemical oxygen demand, the least squares regression algorithm was used to fit the water quality spectrum-COD correlation function; after obtaining several spectral characteristic curves affected only by water nitrogen content, the least squares regression algorithm was used to fit the water quality nitrogen content to obtain the water quality spectrum-nitrogen content correlation function.

[0013] Furthermore, the fitting error model sequentially includes a first residual layer, a second residual layer, a third residual layer, a dimensionality reduction layer, a first convolutional layer, a first dropout layer, a second convolutional layer, and a second dropout layer.

[0014] Furthermore, the training loss value is calculated based on the water quality prediction value, the prediction error training value and the true label. The calculation process includes: ; ; ; ; In the formula, It is represented as the water quality fitting value of the ith monitoring point; n is represented as the number of monitoring points; is the water quality prediction value of the i-th monitoring point; is the prediction error training value of the i-th monitoring point; is the true label; It is expressed as the mean square error of the fitting error model training; Represents the cross entropy loss value for fitting error model training; is the training loss value of the fitting error model; is represented as a weight parameter.

[0015] Compared with the prior art, the present invention has the following beneficial effects: The micro automatic water quality monitoring station of the present invention comprises a circulation pool and a control unit; the circulation pool collects water and then transmits it to a spectrometer; the spectrometer is electrically connected to the control unit; the control unit is electrically connected to a communication module; the controller controls the spectrometer to regularly detect water quality to obtain spectral monitoring data and sends it to a server; the server generates river and lake water quality perception results; the water quality and ecological environment of rivers and lakes are accurately monitored in real time; at the same time, the online full-spectrum water quality analysis technology can be directly immersed in water to realize in-situ detection of water quality, which can reduce the collected monitoring samples to a certain extent compared with traditional sensor detection.

[0016] The server described in the present invention is preset with a water quality correlation model and a fitting error model. The spectral monitoring data is input into the water quality correlation model to obtain the initial value of water quality evaluation, the spectral monitoring data is input into the fitting error model to obtain the water quality evaluation error, and the initial value of water quality evaluation and the water quality evaluation error are superimposed to obtain the river and lake water quality perception result; the fitting error model is used to compensate for the water quality prediction error caused by the spectral characteristic curve of the cross-influence of turbidity, COD and total nitrogen. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A structural diagram of the river and lake water quality information Internet of Things perception system provided in Example 1 of the present invention; Figure 2 A structural diagram of a micro automatic water quality monitoring station provided in Example 1 of the present invention; Figure 3 A flow chart of data evaluation by a server provided in Example 1 of the present invention; Figure 4 A flow chart of constructing a river and lake virtual model in the river and lake virtual display platform provided in Example 1 of the present invention; Figure 5 A flow chart showing the water quality assessment results of rivers and lakes at target points on the river and lake virtual display platform provided in Example 1 of the present invention; Figure 6 A flow chart of a method for sensing river and lake water quality information through the Internet of Things provided in Example 2 of the present invention; Figure 7 This is a structural diagram of the fitting error model provided in Example 2 of the present invention. DETAILED DESCRIPTION

[0018] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the protection scope of the present invention.

[0019] Example 1 like Figure 1 to Figure 2 As shown, this implementation provides a river and lake water quality information Internet of Things perception system, including a micro water quality automatic monitoring station, a communication module and a server; a plurality of micro water quality automatic monitoring stations are connected to the server through the communication module; The micro water quality automatic monitoring station includes a circulation pool and a control unit; the circulation pool collects water and then transmits it to the spectrometer; the spectrometer is electrically connected to the control unit; the control unit is electrically connected to the communication module; The controller controls the spectrometer to regularly detect water quality, obtain spectral monitoring data, and send it to the server; the server is preset with a water quality correlation model and a fitting error model, the spectral monitoring data is input into the water quality correlation model to obtain an initial water quality evaluation value, the spectral monitoring data is input into the fitting error model to obtain a water quality evaluation error, and the initial water quality evaluation value and the water quality evaluation error are superimposed to obtain the river and lake water quality perception result.

[0020] A float type water sampling device is provided at the monitoring point of the micro water quality automatic monitoring station, the water intake pipe of the circulation pool is connected to the float type water sampling device, and the water inlet hole of the water intake pipe of the circulation pool is located 0.5m to 1m below the water surface.

[0021] The micro water quality automatic monitoring station is equipped with a sensor group electrically connected to the controller; the sensor group includes a pH sensor, a conductivity sensor, a temperature sensor, and a flow rate and flow sensor; the pH sensor measures the acidity and alkalinity of river and lake water quality; the conductivity sensor measures the content of dissolved salts in river and lake water quality; the temperature sensor measures water temperature; the flow rate and flow sensor measures the speed and flow of river and lake water flow; like Figure 3 As shown, the server calculates and obtains the river and lake water quality assessment result according to the pH value, the content of dissolved salts, water temperature, the content of nitrate nitrogen, the speed and flow of water flow, and the river and lake water quality perception result, and the process includes: The pH, dissolved salt content, water temperature, nitrate nitrogen content, water velocity and flow rate are used as water quality monitoring parameters. The abnormal value detection range is set based on the water quality monitoring parameters. The expression formula is: ; ; ; In the formula, is the parameter value that 25% of the data points in the b-th water quality monitoring parameter set are less than or equal to; is the parameter value that 75% of the data points in the b-th water quality monitoring parameter set are less than or equal to; is the distribution range of the middle 50% of the data in the bth water quality monitoring parameter set; is the maximum value in the abnormal value detection range of the bth water quality monitoring parameter; is the minimum value in the abnormal value detection range of the bth water quality monitoring parameter; The water quality monitoring parameters are screened for outliers based on the outlier detection range; the water quality monitoring parameters are denoised, and the expression formula is: ; In the formula, It is represented as the output data after denoising the b-th water quality monitoring parameter at time t; is the water quality monitoring parameter of the bth type at time k; is the moving average window size; t is the current time series; A linear interpolation algorithm is used to fill in the missing data in the water quality monitoring parameters, and then the evaluation values ​​of the water quality monitoring parameters and the evaluation values ​​of the river and lake water quality perception results are obtained based on the preset mapping relationship table; the evaluation values ​​of the water quality monitoring parameters and the evaluation values ​​of the river and lake water quality perception results are weighted and summed to obtain the river and lake water quality assessment results.

[0022] The server is connected to the monitoring terminal through a wireless network; when the river and lake water quality assessment result is lower than the set threshold, the server retrieves the water system distribution map, adds the river and lake water quality assessment result to the water system distribution map to generate early warning information and sends it to the monitoring terminal.

[0023] like Figure 4 and Figure 5 As shown, the river and lake water quality information IoT perception system also includes a river and lake virtual display platform, which constructs a three-dimensional scene model based on water system distribution data and terrain data; A WebGL scene architecture is established based on the viewing angle lighting scene and the web page architecture. After the three-dimensional scene model is converted and compressed, the model scene data is embedded in the WebGL scene architecture to construct a river and lake virtual model. The river and lake water quality assessment results are added to the river and lake virtual model. The received viewpoint coordinates and target point coordinates are input into the river and lake virtual model, and the river and lake water quality assessment results of the target point are displayed through the river and lake virtual model.

[0024] The micro automatic water quality monitoring station is provided with a surge protector; the internal grounding bar is connected to the controller and the spectrum analyzer in the micro automatic water quality monitoring station; a support for fixing the antenna in the communication module is arranged outside the micro automatic water quality monitoring station; the outdoor grounding grid is connected to the antenna support; one end of the surge protector is connected to the power supply through the surge protector lead wire, and the other end is connected to the outdoor grounding grid through the surge protector grounding wire.

[0025] In this embodiment, a number of micro automatic water quality monitoring stations distributed at river and lake monitoring points are used for fully automatic operation to monitor the on-site water quality in real time and record changes in water quality. The monitoring data is uploaded to a remote big data platform through a communication system, thereby realizing a remote monitoring function. The micro automatic water quality monitoring station is modular in setting, has high integration, small footprint, and flexible site selection for monitoring points to meet various on-site conditions. It also has good scalability and replaceability, and can adjust or replace monitoring parameters according to the user's monitoring needs. In addition, the online full-spectrum water quality analysis technology can be directly immersed in water to realize in-situ detection of water quality.

[0026] Example 2 like Figure 6 to Figure 7 As shown, this embodiment provides a method for sensing river and lake water quality information through the Internet of Things. The method is applied to the system described in Example 1, and the method includes: Constructing water quality correlation model and fitting error model includes: Obtain several spectral characteristic curves that are only affected by turbidity; fit each spectral characteristic curve respectively to obtain multiple spectral curve fitting functions; fit the coefficient of each spectral curve fitting function with the corresponding turbidity to obtain the correlation function between each coefficient and turbidity, and use the least squares regression algorithm to fit to obtain the water quality spectrum-turbidity correlation function; Similarly, after obtaining several spectral characteristic curves that are only affected by chemical oxygen demand, the least squares regression algorithm is used to fit them with the chemical oxygen demand to obtain the water quality spectrum-COD correlation function; after obtaining several spectral characteristic curves that are only affected by the water quality nitrogen content, the least squares regression algorithm is used to fit them with the water quality nitrogen content to obtain the water quality spectrum-nitrogen content correlation function.

[0027] The water quality correlation model is obtained by coupling the water quality spectrum-turbidity correlation function, the water quality spectrum-COD correlation function and the water quality spectrum-nitrogen content correlation function. A fitting error model is constructed based on a convolutional neural network, and the fitting error model sequentially includes a first residual layer, a second residual layer, a third residual layer, a dimensionality reduction layer, a first convolutional layer, a first dropout layer, a second convolutional layer, and a second dropout layer.

[0028] Water quality training data is obtained from a water quality database; the water quality training data includes spectral characteristics, water quality turbidity, chemical oxygen demand and water quality nitrogen content; the water quality turbidity, chemical oxygen demand and water quality nitrogen content are weighted and summed to obtain a true label; The spectral features in the water quality training data are input into the water quality correlation model to obtain the water quality prediction value; the spectral features in the water quality training data are input into the fitting error model to obtain the prediction error training value. The calculation process includes: ; ; ; ; In the formula, It is represented as the water quality fitting value of the ith monitoring point; n is represented as the number of monitoring points; is the water quality prediction value of the ith monitoring point; is the prediction error training value of the i-th monitoring point; is the true label; It is expressed as the mean square error of the fitting error model training; Represents the cross entropy loss value for fitting error model training; is the training loss value of the fitting error model; is represented as a weight parameter.

[0029] The training loss value is calculated according to the water quality prediction value, the prediction error training value and the true label; the fitting error model is optimized according to the training loss value, and the training process of the neural convolution model is repeated until the training loss value converges to output the trained fitting error model.

[0030] Control the spectrometer to regularly detect water quality to obtain spectral monitoring data, input the spectral monitoring data into the water quality correlation model to obtain the initial value of water quality evaluation, input the spectral monitoring data into the fitting error model to obtain the water quality evaluation error, superimpose the initial water quality evaluation value and the water quality evaluation error to obtain the river and lake water quality perception result; use the fitting error model to compensate for the water quality prediction error caused by the spectral characteristic curve of the cross-influence of turbidity, COD and total nitrogen.

[0031] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0032] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0033] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0034] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0035] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A river and lake water quality information Internet of Things perception system, characterized in that: It includes a micro automatic water quality monitoring station, a communication module and a server; a plurality of micro automatic water quality monitoring stations are connected to the server through the communication module; The micro water quality automatic monitoring station includes a circulation pool and a control unit; the circulation pool collects water and then transmits it to the spectrometer; the spectrometer is electrically connected to the control unit; the control unit is electrically connected to the communication module; The controller controls the spectrometer to regularly detect water quality, obtain spectral monitoring data, and send it to the server; the server is preset with a water quality correlation model and a fitting error model, the spectral monitoring data is input into the water quality correlation model to obtain an initial water quality evaluation value, the spectral monitoring data is input into the fitting error model to obtain a water quality evaluation error, and the initial water quality evaluation value and the water quality evaluation error are superimposed to obtain the river and lake water quality perception result.

2. The river and lake water quality information Internet of Things perception system according to claim 1 is characterized in that: A float type water sampling device is provided at the monitoring point of the micro water quality automatic monitoring station, the water intake pipe of the circulation pool is connected to the float type water sampling device, and the water inlet hole of the water intake pipe of the circulation pool is located 0.5m to 1m below the water surface.

3. The river and lake water quality information Internet of Things perception system according to claim 1 is characterized in that: The micro water quality automatic monitoring station is equipped with a sensor group electrically connected to the controller; the sensor group includes a pH sensor, a conductivity sensor, a temperature sensor, and a flow rate and flow sensor; the pH sensor measures the acidity and alkalinity of river and lake water quality; the conductivity sensor measures the content of dissolved salts in river and lake water quality; the temperature sensor measures water temperature; the flow rate and flow sensor measures the speed and flow of river and lake water flow; The server calculates and obtains river and lake water quality assessment results based on pH, dissolved salt content, water temperature, nitrate nitrogen content, water flow speed and flow, and river and lake water quality perception results.

4. The river and lake water quality information Internet of Things perception system according to claim 3 is characterized in that: The server calculates and obtains river and lake water quality assessment results based on pH, dissolved salt content, water temperature, nitrate nitrogen content, water flow speed and flow rate, and river and lake water quality perception results, and the process includes: The pH, dissolved salt content, water temperature, nitrate nitrogen content, water velocity and flow rate are used as water quality monitoring parameters. The abnormal value detection range is set based on the water quality monitoring parameters. The expression formula is: ； ； ； In the formula, is the parameter value that 25% of the data points in the b-th water quality monitoring parameter set are less than or equal to; is the parameter value that 75% of the data points in the b-th water quality monitoring parameter set are less than or equal to; is the distribution range of the middle 50% of the data in the bth water quality monitoring parameter set; is the maximum value in the abnormal value detection range of the bth water quality monitoring parameter; is the minimum value in the abnormal value detection range of the bth water quality monitoring parameter; The water quality monitoring parameters are screened for outliers based on the outlier detection range; the water quality monitoring parameters are denoised, and the expression formula is: ； In the formula, It is represented as the output data after denoising the b-th water quality monitoring parameter at time t; is the water quality monitoring parameter of the bth type at time k; is the moving average window size; t is the current time series; A linear interpolation algorithm is used to fill in the missing data in the water quality monitoring parameters, and then the evaluation values ​​of the water quality monitoring parameters and the evaluation values ​​of the river and lake water quality perception results are obtained based on the preset mapping relationship table; the evaluation values ​​of the water quality monitoring parameters and the evaluation values ​​of the river and lake water quality perception results are weighted and summed to obtain the river and lake water quality assessment results.

5. The river and lake water quality information Internet of Things perception system according to claim 3 is characterized in that: It also includes a river and lake virtual display platform, which constructs a three-dimensional scene model based on water system distribution data and terrain data; Establish a WebGL scene architecture based on the viewing angle lighting scene and web page architecture, convert and compress the 3D scene model, embed the model scene data with the WebGL scene architecture to build a river and lake virtual model; add the river and lake water quality assessment results to the river and lake virtual model; The received viewpoint coordinates and target point coordinates are input into the river and lake virtual model, and the river and lake water quality assessment results of the target point are displayed through the river and lake virtual model.

6. The control method of the river and lake water quality information Internet of Things perception system according to any one of claims 1 to 5, comprising: Control the spectrometer to regularly detect water quality to obtain spectral monitoring data, input the spectral monitoring data into a preset water quality correlation model to obtain an initial water quality evaluation value, input the spectral monitoring data into a preset fitting error model to obtain a water quality evaluation error, and superimpose the initial water quality evaluation value and the water quality evaluation error to obtain the river and lake water quality perception results.

7. The control method according to claim 6, characterized in that: The water quality correlation model and fitting error model construction process include: The water quality correlation model is obtained by coupling the water quality spectrum-turbidity correlation function, the water quality spectrum-COD correlation function and the water quality spectrum-nitrogen content correlation function. A fitting error model is constructed based on a convolutional neural network, and water quality training data is obtained from a water quality database; the water quality training data includes spectral characteristics, water turbidity, chemical oxygen demand, and water nitrogen content; the water turbidity, chemical oxygen demand, and water nitrogen content are weighted and summed to obtain a true label; The spectral features in the water quality training data are input into the water quality correlation model to obtain the water quality prediction value; the spectral features in the water quality training data are input into the fitting error model to obtain the prediction error training value, and the training loss value is calculated according to the water quality prediction value, the prediction error training value and the true label; the fitting error model is optimized according to the training loss value, and the training process of the neural convolution model is repeated until the training loss value converges to output the trained fitting error model.

8. The control method according to claim 7, characterized in that: The process of obtaining the water quality spectrum-turbidity correlation function, the water quality spectrum-COD correlation function and the water quality spectrum-nitrogen content correlation function includes: After obtaining several spectral characteristic curves affected only by turbidity, the least squares regression algorithm was used to fit the water quality turbidity characteristics to obtain the water quality spectrum-turbidity correlation function; after obtaining several spectral characteristic curves affected only by chemical oxygen demand, the least squares regression algorithm was used to fit the water quality spectrum-COD correlation function; after obtaining several spectral characteristic curves affected only by water nitrogen content, the least squares regression algorithm was used to fit the water quality nitrogen content to obtain the water quality spectrum-nitrogen content correlation function.

9. The control method according to claim 7, characterized in that: The fitting error model includes a first residual layer, a second residual layer, a third residual layer, a dimensionality reduction layer, a first convolutional layer, a first dropout layer, a second convolutional layer, and a second dropout layer in sequence.

10. The control method according to claim 7, characterized in that: The training loss value is calculated based on the water quality prediction value, the prediction error training value and the true label. The calculation process includes: ； ； ； ； In the formula, It is represented as the water quality fitting value of the ith monitoring point; n is represented as the number of monitoring points; is the water quality prediction value of the i-th monitoring point; is the prediction error training value of the i-th monitoring point; is the true label; It is expressed as the mean square error of the fitting error model training; Represents the cross entropy loss value for fitting error model training; is the training loss value of the fitting error model; is represented as a weight parameter.

Images