Water environment management method and platform based on smart Internet of Things and big data analysis

Through intelligent IoT and big data analysis, combined with watershed mechanism and machine learning model, the limitations of the comparison of single parameters of the water quality detection system are solved, and more accurate and timely water quality monitoring is achieved to identify pollution sources and changes.

CN118568614BActive Publication Date: 2024-11-08WUHAN LEITE TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410872882.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-07-01
Publication Date
2024-11-08
Estimated Expiration
2044-07-01

AI Technical Summary

Technical Problem

Due to the limitations of the comparison of single parameters and thresholds, the existing water quality detection system cannot fully reflect the actual conditions of water quality, resulting in low accuracy and the inability to timely identify potential pollution sources or environmental changes.

Method used

Using a method based on smart IoT and big data analysis, we can obtain water quality parameters in real time through the Internet of Things platform, extract and fusion of feature, input a pre-constructed big data analysis model, integrate the basin mechanism model and machine learning model, and predict key water quality indicators. And analyze the concentration trend and change trend along the route, and conduct water quality warning.

Benefits of technology

It improves the accuracy and timeliness of water quality detection, can reflect the water quality more comprehensively, identify key indicators that have the greatest impact on water quality, and promptly discover potential pollution sources and water quality changes.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118568614B_ABST
    Figure CN118568614B_ABST
Patent Text Reader

Abstract

The present application discloses a water environment management method and platform based on smart Internet of Things and big data analysis, which relates to the field of water quality detection. The method comprises: obtaining a water quality parameter set of each target section in real time through an Internet of Things platform; extracting features of the water quality parameter set of each target section to obtain a water quality feature set of the watershed where each target section is located; inputting the water quality feature set of each target section into a pre-built big data analysis model to predict and obtain key water quality indicators of the watershed where each target section is located; analyzing the key water quality indicators to determine the concentration trend of the key water quality indicators of the watershed where each target section is located, as well as determining the along-the-way variation trend and diffusion path of the key water quality indicators between different target sections; and issuing a water quality warning based on the concentration trend and the along-the-way variation trend; wherein the big data analysis model is obtained based on the fusion of a pre-built watershed mechanism model and a pre-built machine learning model.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The embodiments of the present application relate to the field of water quality testing, and in particular to a water environment management method and platform based on smart Internet of Things and big data analysis. Background Art

[0002] Water quality testing of rivers / lakes plays a very important role in environmental protection. By testing the water quality of rivers / lakes, water environment problems can be discovered and solved in a timely manner. The water quality data required for current water quality testing can be obtained through sensors deployed in rivers / lakes. The water quality data obtained by sensors include pH value, dissolved oxygen, turbidity, temperature, etc. These data are transmitted to the monitoring system for evaluating water quality conditions.

[0003] The monitoring system monitors the data based on the acquired water quality data. Its underlying logic is to compare each water quality data with the corresponding preset threshold. For example, if a certain water quality data (such as pH value) exceeds the preset safety range, the monitoring system will issue an alarm. However, this simple comparison method has obvious limitations and may lead to inaccurate water quality detection. The water quality of rivers / lakes is a complex multivariate system. There are complex interactions between various water quality data. For example, changes in dissolved oxygen may be associated with changes in temperature and turbidity, not just changes in a single parameter. Assuming that the dissolved oxygen content of a river suddenly drops, but the temperature and turbidity are still within the normal range, a simple threshold comparison may only focus on the abnormality of dissolved oxygen and ignore other factors that may cause the decrease in dissolved oxygen, such as the decrease in dissolved oxygen caused by increased temperature. In this way, the water quality monitoring system may not be able to identify potential pollution sources or environmental changes in a timely manner, resulting in incorrect judgments. Therefore, simply relying on the comparison of each parameter with the corresponding threshold value cannot fully reflect the actual situation of water quality, resulting in low accuracy of water quality detection.

[0004] There is currently no effective solution to the above problems. Summary of the invention

[0005] The embodiments of the present application provide a water environment management method and platform based on smart Internet of Things and big data analysis, which are used to effectively improve the accuracy of water quality detection.

[0006] To achieve the above objectives, the embodiments of the present application adopt the following technical solutions:

[0007] In a first aspect, a water environment management method based on smart Internet of Things and big data analysis is provided, which is applied to an electronic device, wherein the electronic device is deployed with an Internet of Things platform, and the method comprises:

[0008] Acquire a water quality parameter set of each target section in real time through the Internet of Things platform;

[0009] Performing feature extraction on the water quality parameter set of each target section to obtain a water quality feature set of the watershed where each target section is located;

[0010] Input the water quality feature set of each target section into a pre-built big data analysis model to predict and obtain key water quality indicators of the basin where each target section is located;

[0011] Analyze the key water quality indicators, determine the concentration trend of the key water quality indicators in the watershed where each target section is located, and determine the along-the-way variation trend and diffusion path of the key water quality indicators between different target sections;

[0012] Conduct water quality warning according to the concentration trend and the change trend along the route;

[0013] Among them, the big data analysis model is obtained by integrating a pre-built watershed mechanism model and a pre-built machine learning model.

[0014] In a possible implementation of the first aspect, extracting features from the water quality parameter set of each target section to obtain water quality features of each target section includes:

[0015] Extracting and summarizing the basic features of the water quality parameter set of each target section to obtain a basic feature set;

[0016] Performing feature fusion on any two different basic features of the same target section in the basic feature set to obtain a feature set including a plurality of fused features;

[0017] Performing feature crossover on each of the fused features in the same watershed in the feature set including the plurality of fused features, to obtain a feature set including the plurality of crossover features;

[0018] Feature selection is performed on the feature set including a plurality of the cross features to obtain a water quality feature set for each of the river basins.

[0019] In a possible implementation manner of the first aspect, the performing feature fusion on any two different basic features of the same target section in the basic feature set to obtain a feature set including a plurality of fused features includes:

[0020] Nonlinearly combining any two different basic features of the same target section in the basic feature set to obtain a fused feature set;

[0021] Constructing a feature matrix according to the fused feature set;

[0022] Calculating a covariance matrix of the characteristic matrix, and performing eigenvalue decomposition on the covariance matrix to obtain a plurality of eigenvalues ​​and an eigenvector corresponding to each eigenvalue;

[0023] Select the eigenvectors corresponding to the first k largest eigenvalues ​​to form a principal component matrix;

[0024] Projecting the feature matrix onto the principal component matrix to obtain a feature matrix after dimension reduction;

[0025] The feature matrix after dimension reduction is used as a feature set including multiple fusion features.

[0026] In a possible implementation of the first aspect, the step of constructing the big data analysis model includes:

[0027] Obtaining historical water quality parameters of each of the river basins, and constructing a river basin mechanism model based on the historical water quality parameters of each of the river basins, wherein the output of the river basin mechanism model is a function of the river basin and the corresponding historical water quality parameters, and the river basin mechanism model is used to simulate and predict changes in the output river basin and the corresponding water quality parameters;

[0028] A logistic regression model is constructed and trained according to the historical water quality parameters, wherein the output of the logistic regression model is a function of the historical water quality parameters, and the logistic regression model is used to predict the probability of each of the water quality parameters meeting the standard;

[0029] The output of the watershed mechanism model and the output of the logistic regression model are linearly combined by weight coefficients to obtain the output of the fusion model, wherein the sum of the weight coefficient corresponding to the output of the watershed mechanism model and the weight coefficient corresponding to the output of the logistic regression model is 1;

[0030] Using mean square error as an optimization target, optimizing the weight coefficient, and obtaining an output of an optimized fusion model;

[0031] The output of the optimized fusion model is used as the input of the random forest model to obtain a big data analysis model.

[0032] In a possible implementation manner of the first aspect, the historical water quality parameters include pH value, dissolved oxygen, turbidity and temperature, and constructing a watershed mechanism model based on the historical water quality parameters of each watershed includes:

[0033] Define a first support function of the pH value of the basin i, a second support function of the dissolved oxygen, a third support function of the turbidity, and a fourth support function of the temperature, wherein the support function is used to quantify the contribution of each of the historical water quality parameters to the water quality of the basin i, and i is a positive integer;

[0034] Using the parameter weight of the watershed i, weighting the first support function, the second support function, the third support function and the fourth support function, to obtain the historical support function of the watershed i, wherein each watershed has different parameter weights for the first support function, the second support function, the third support function and the fourth support function;

[0035] Averagely merge the historical support functions of all the river basins to obtain a comprehensive support function;

[0036] The historical water quality parameters are used to train a linear regression model to obtain the regression coefficient of each water quality parameter change, and a watershed mechanism model is constructed based on the regression coefficient of each water quality parameter change and the comprehensive support function.

[0037] In a possible implementation manner of the first aspect, using the mean square error as an optimization target, optimizing the weight coefficient, and obtaining an output of the optimized fusion model includes:

[0038] Execute the loop steps until the mean square error is minimized and obtain the output of the optimized fusion model;

[0039] Wherein, the circulation step comprises:

[0040] Use the mean square error formula to calculate the initial mean square error;

[0041] Calculating the partial derivative of the initial mean square error with respect to the weight coefficient of the output of the watershed mechanism model;

[0042] Using a gradient descent formula, updating the weight coefficient of the output of the watershed mechanism model according to the partial derivative, to obtain an updated weight coefficient of the output of the watershed mechanism model;

[0043] The mean square error is calculated based on the updated weight coefficient of the output of the watershed mechanism model.

[0044] In a possible implementation of the first aspect, analyzing the key water quality indicators to determine the concentration trend of the key water quality indicators in the watershed where each target section is located includes:

[0045] Performing time series analysis on key water quality indicators in the river basin where each target section is located to obtain concentration trends of key water quality indicators in the river basin where each target section is located;

[0046] After analyzing the key water quality indicators and determining the concentration trend of the key water quality indicators in the river basin where each target section is located, the method further includes:

[0047] According to the concentration trend of key water quality indicators in the watershed where each target section is located, a moving average graph and a trend line graph are drawn.

[0048] In a possible implementation of the first aspect, determining the along-the-way variation trend and diffusion path of the key water quality indicator between different target sections includes:

[0049] Obtaining the geographical location of each target section;

[0050] Using Kriging interpolation, constructing a Kriging equation according to the geographical location of each target section, wherein the Kriging equation is used to calculate the estimated value of the key water quality index of the final target section;

[0051] Generate an interpolation network for the watershed i, estimate the key water quality index of each grid point in the interpolation network according to the Kriging equation, and obtain an interpolation network including a plurality of estimation results;

[0052] According to the interpolation network including multiple estimation results, a contour map of the basin i is drawn, wherein the contour map is used to show the spatial distribution of key water quality indicators in the basin i, and the spatial distribution includes the along-the-way variation trend and diffusion path of the key water quality indicators between different target sections.

[0053] In a second aspect, the present application provides an Internet of Things platform, including:

[0054] A data acquisition module is used to obtain the water quality parameter set of each target section in real time;

[0055] A feature extraction module is used to extract features from a water quality parameter set of each target section to obtain a water quality feature set of the watershed where each target section is located;

[0056] A prediction module, used to input the water quality feature set of each target section into a pre-built big data analysis model to predict and obtain key water quality indicators of the basin where each target section is located;

[0057] An analysis module, used to analyze the key water quality indicators, determine the concentration trend of the key water quality indicators in the basin where each target section is located, and determine the along-the-way change trend and diffusion path of the key water quality indicators between different target sections;

[0058] An early warning module, used for making a water quality early warning according to the concentration trend and the change trend along the process;

[0059] Among them, the big data analysis model is obtained by integrating a pre-built watershed mechanism model and a pre-built machine learning model.

[0060] In a third aspect, the present application provides an electronic device, wherein the electronic device is deployed with the above-mentioned Internet of Things platform, and the electronic device includes:

[0061] a memory configured to store instructions; and

[0062] The processor is configured to call the instructions from the memory and to implement the above-mentioned water environment management method based on smart Internet of Things and big data analysis when executing the instructions.

[0063] Through the above technical solution, the water quality feature set of each target section is input into the pre-built big data analysis model, which can predict and obtain the key water quality indicators of the basin where each target section is located. Among them, the big data analysis model integrates the basin mechanism model and the machine learning model, which can comprehensively consider the impact of multiple factors on water quality, more accurately predict water quality changes, identify the key indicators that have the greatest impact on water quality, and effectively improve the accuracy of water quality detection; after obtaining the key water quality indicators, the key water quality indicators are analyzed to determine the concentration trend of the key water quality indicators in the basin where each target section is located, as well as the along-the-way change trend and diffusion path between different target sections, so as to facilitate the timely discovery of potential pollution sources and water quality changes.

[0064] Other features and advantages of the embodiments of the present application will be described in detail in the subsequent specific implementation section. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 A schematic diagram of a water environment management method based on smart IoT and big data analysis provided in an embodiment of the present application;

[0066] Figure 2 An architecture diagram of an Internet of Things platform provided in an embodiment of the present application;

[0067] Figure 3 One of the display pages of the Internet of Things platform provided in the embodiment of the present application;

[0068] Figure 4 Another architecture diagram of an Internet of Things platform provided in an embodiment of the present application;

[0069] Figure 5 An interface display diagram of the Internet of Things platform provided in an embodiment of the present application. DETAILED DESCRIPTION

[0070] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It should be understood that the specific implementation methods described herein are only used to illustrate and explain the embodiments of the present application, and are not used to limit the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0071] It should be noted that if the embodiments of the present application involve directional indications (such as up, down, left, right, front, back...), such directional indications are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0072] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present application, the descriptions of "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or suggesting their relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0073] Figure 1 The following schematically shows a flow chart of a water environment management method based on smart IoT and big data analysis according to an embodiment of the present application. Figure 1 As shown, an embodiment of the present application provides a water environment management method based on smart Internet of Things and big data analysis, which is applied to electronic devices, and the electronic devices are deployed with an Internet of Things platform. The method may include the following steps.

[0074] S110, obtaining the water quality parameter set of each target section in real time through the Internet of Things platform;

[0075] S120, extracting features from the water quality parameter set of each target section to obtain a water quality feature set for the watershed where each target section is located;

[0076] S130, inputting the water quality feature set of each target section into a pre-built big data analysis model to predict and obtain key water quality indicators of the watershed where each target section is located;

[0077] S140, analyzing key water quality indicators, determining the concentration trend of key water quality indicators in the basin where each target section is located, and determining the change trend and diffusion path of key water quality indicators along the way between different target sections;

[0078] S150, water quality warning based on concentration trend and change trend along the way;

[0079] Among them, the big data analysis model is obtained by integrating the pre-built watershed mechanism model and the pre-built machine learning model.

[0080] This embodiment obtains the water quality parameter set of each target section in real time through the Internet of Things platform. Sensors are installed at each target section, and the sensors can monitor and collect water quality parameters such as pH value, dissolved oxygen, turbidity, temperature, ammonia nitrogen, total phosphorus, etc. in real time. The sensors transmit data to the Internet of Things platform through a wireless network. The Internet of Things platform can be a system that integrates data reception, storage and processing, and can perform preliminary processing and storage immediately after receiving the data.

[0081] Figure 2 The architecture diagram of an Internet of Things platform provided by this embodiment is shown, which includes a perception layer, a basic layer and a data layer. The perception layer is used to collect water quality parameters. In addition, the Internet of Things platform provided by this embodiment also includes an application platform. The application platform is used to integrate data applications and mobile applications. Intelligent display can be realized based on the application platform, such as Figure 3 shown.

[0082] Reference Figure 3 , which shows one of the display pages of the Internet of Things platform provided in this embodiment, which is used to display the water quality conditions of the river basins and sections in a certain area, wherein each section is installed with sensors for monitoring, and the Internet of Things platform can perform data analysis based on the data transmitted by the sensors to determine the water quality conditions.

[0083] Specifically, firstly, feature extraction is performed on the water quality parameter set of each target section to obtain the water quality feature set of the watershed where each target section is located. Specifically, after obtaining the real-time water quality parameters, the data processing algorithm is used to extract features from these parameters. The feature extraction process can include data cleaning, normalization and feature selection. Data cleaning is to remove noise and outliers to ensure the accuracy and consistency of the data. Normalization is to convert parameters of different dimensions to the same dimension for subsequent analysis. Feature selection is to select key features that can represent the water quality situation based on the importance and relevance of water quality parameters. The water quality feature set can reflect the overall water quality status of the watershed where the section is located.

[0084] The water quality feature set of each target section is input into the pre-built big data analysis model to predict and obtain the key water quality indicators of the basin where each target section is located. The big data analysis model is constructed based on the fusion of the basin mechanism model and the machine learning model. The basin mechanism model can simulate the changing laws of water quality parameters in the basin, while the machine learning model has the ability to predict water quality changes through training with a large amount of historical data. The extracted water quality feature set is input into the big data analysis model, and the model will predict water quality changes based on the hydrological characteristics and historical data of the basin. Through model calculation, the key water quality indicators of each target section in the future period of time, such as temperature, total phosphorus concentration, etc., can be obtained. These key water quality indicators are key parameters that affect the quality of the water environment and can directly reflect the quality of the water.

[0085] Analyze key water quality indicators, determine the concentration trend of key water quality indicators in the basin where each target section is located, and determine the change trend and diffusion path of key water quality indicators along the way between different target sections. Specifically, after obtaining the key water quality indicators, use time series analysis methods to analyze the concentration change trend of these indicators. For example, through methods such as moving average and exponential smoothing, the change patterns of key water quality indicators in different time periods can be observed. At the same time, combined with the basin's geographic information system (GIS), the spatial change trend and diffusion path of key water quality indicators between different target sections can be analyzed. Through spatial interpolation methods, such as Kriging interpolation, a spatial distribution map of key water quality indicators can be drawn, which intuitively shows the diffusion and change trend of water quality indicators in the basin.

[0086] Water quality warning can be carried out according to the concentration trend and the change trend along the way. Specifically, a water quality warning model can be established and a warning threshold can be set. When the concentration of key water quality indicators exceeds the warning threshold or the change trend along the way shows an abnormality, the warning model will issue a warning signal. The warning signal can be sent to relevant managers in a timely manner through SMS, email or the notification function of the Internet of Things platform. Managers can take corresponding response measures based on the warning information, such as increasing the frequency of water quality monitoring, launching emergency treatment plans, etc. In this way, water quality problems can be discovered and dealt with in time before they occur, avoiding further deterioration of the water environment and ensuring ecological security in the basin.

[0087] The embodiments of the present application can effectively solve the limitations of the current water quality detection system and improve the accuracy and timeliness of water quality detection through more comprehensive and intelligent means.

[0088] First, the real-time acquisition of water quality parameter sets for each target section through the IoT platform solves the problems of real-time and comprehensive data acquisition. Sensors installed in rivers or lakes can monitor water quality parameters such as pH, dissolved oxygen, turbidity, temperature, etc. 24 hours a day, and transmit data to the IoT platform through wireless networks. The IoT platform can not only receive and store this data in real time, but also perform preliminary processing and analysis to ensure the accuracy and completeness of the data. In this way, the lag and limitations of traditional manual sampling and detection methods can be avoided, and real-time monitoring of water quality can be achieved.

[0089] Secondly, feature extraction is performed on the water quality parameter set of each target section to obtain the water quality feature set of the basin where each target section is located. Through data cleaning, normalization and feature selection, key features that can represent the water quality are extracted. This step not only takes into account the changes in a single parameter, but also reveals the complex interactions between various water quality parameters through feature extraction methods. For example, changes in dissolved oxygen may be related to changes in parameters such as temperature and turbidity. Through feature extraction, the actual water quality situation can be more comprehensively reflected, avoiding the limitations of simply relying on the comparison of each parameter with the corresponding threshold.

[0090] Then, the water quality feature set of each target section is input into the pre-built big data analysis model to predict and obtain the key water quality indicators of the basin where each target section is located. The big data analysis model integrates the basin mechanism model and the machine learning model, and can accurately predict water quality changes based on the hydrological characteristics of the basin and a large amount of historical data. Through training on a large amount of historical data, the machine learning model can identify the complex relationship and change patterns between various water quality parameters, so that when predicting water quality changes, the interaction between multiple variables can be taken into account. For example, the model can identify the trend that increased temperature may lead to a decrease in dissolved oxygen, so that when the dissolved oxygen is abnormal, the temperature change can be considered at the same time to avoid misjudgment.

[0091] Furthermore, the key water quality indicators are analyzed to determine the concentration trend of the key water quality indicators in the basin where each target section is located, as well as the change trend and diffusion path of the key water quality indicators along the different target sections. Through time series analysis and spatial interpolation methods, the change trend and spatial distribution of key water quality indicators can be comprehensively analyzed. For example, through moving average or Kriging interpolation methods, the spatial distribution map of key water quality indicators can be drawn to intuitively display the diffusion path and change trend of water quality indicators.

[0092] Finally, water quality warnings are issued based on concentration trends and change trends along the route. When the concentration of key water quality indicators exceeds the warning threshold or the change trend along the route shows an abnormality, the warning model will issue a warning signal. The warning system not only considers the abnormality of a single parameter, but also combines the interaction and change trend between multiple variables, which can more accurately identify potential sources of pollution or environmental changes. For example, if the dissolved oxygen drops abnormally and an increase in temperature is detected at the same time, the warning system can combine these two factors to promptly identify possible sources of pollution or environmental changes and avoid misjudgment.

[0093] In summary, this embodiment inputs the water quality feature set of each target section into the pre-built big data analysis model, which can predict and obtain the key water quality indicators of the river basin where each target section is located. Among them, the big data analysis model integrates the river basin mechanism model and the machine learning model, which can comprehensively consider the impact of various factors on water quality, more accurately predict water quality changes, identify the key indicators that have the greatest impact on water quality, and effectively improve the accuracy of water quality detection; after obtaining the key water quality indicators, the key water quality indicators are analyzed to determine the concentration trend of the key water quality indicators in the river basin where each target section is located, as well as the along-the-way change trend and diffusion path between different target sections, so as to facilitate the timely discovery of potential pollution sources and water quality changes.

[0094] In one implementation of this embodiment, feature extraction is performed on the water quality parameter set of each target section to obtain the water quality feature of each target section, including the following steps:

[0095] S210, extracting and summarizing the basic features of the water quality parameter set of each target section to obtain a basic feature set;

[0096] S220, performing feature fusion on any two different basic features of the same target section in the basic feature set to obtain a feature set including a plurality of fused features;

[0097] S230, performing feature crossover on each fused feature in the same watershed in a feature set including multiple fused features, to obtain a feature set including multiple crossover features;

[0098] S240. Perform feature selection on a feature set including multiple cross-features to obtain a water quality feature set for each watershed.

[0099] Extract and summarize the basic characteristics of the water quality parameter set for each target section. First, in order to ensure the accuracy and consistency of the data, the water quality parameter set of each target section needs to be cleaned and normalized to obtain the basic feature set of each target section. The basic feature set can fully reflect the basic conditions of water quality. Specifically, data cleaning includes removing noise data and outliers. For example, if the data of a sensor at a certain point in time is abnormally high or low, it may be due to sensor failure or external interference, and these data need to be eliminated. Normalization is used to convert parameters of different dimensions to the same dimension, making subsequent analysis more convenient and accurate. For example, the minimum-maximum normalization method can be used to scale all parameter values ​​in the water quality parameter set to between 0 and 1.

[0100] The feature fusion of any two different basic features of the same target section in the basic feature set can reveal the relationship between various water quality parameters. Feature fusion is used to combine different basic features to form new features. The fused features can more comprehensively reflect the complex relationship between water quality parameters.

[0101] The feature set including multiple fusion features in the same basin is cross-featured to further reveal the deep relationship between water quality parameters. Feature cross-feature refers to the generation of new cross-features by combining different fusion features. The feature set including multiple cross-features can more comprehensively and deeply reflect the complex relationship between water quality parameters.

[0102] Feature selection is performed on a feature set that includes multiple cross-features to retain the most useful information. Feature selection is to select the features that have the greatest impact on the target variable (such as water quality) through an algorithm. For example, methods such as LASSO regression and recursive feature elimination (RFE) can be used to sort the features according to their importance and select the most important features. The water quality feature set can more accurately reflect the actual status of water quality.

[0103] This implementation can reflect water quality conditions more comprehensively and accurately, and improve the performance and efficiency of the prediction model.

[0104] In one implementation of this embodiment, feature fusion is performed on any two different basic features of the same target section in the basic feature set to obtain a feature set including multiple fused features, including the following steps:

[0105] S310, performing nonlinear combination on any two different basic features of the same target section in the basic feature set to obtain a fused feature set;

[0106] S320, constructing a feature matrix according to the fused feature set;

[0107] S330, calculating the covariance matrix of the characteristic matrix, and performing eigenvalue decomposition on the covariance matrix to obtain multiple eigenvalues ​​and an eigenvector corresponding to each eigenvalue;

[0108] S340, selecting eigenvectors corresponding to the first k largest eigenvalues ​​to form a principal component matrix;

[0109] S350, projecting the feature matrix onto the principal component matrix to obtain a feature matrix after dimension reduction;

[0110] S360: Using the feature matrix after dimension reduction as a feature set including multiple fusion features.

[0111] Any two different basic features of the same target section in the basic feature set are nonlinearly combined. Nonlinear combination refers to combining two different basic features through a nonlinear function to generate a new fusion feature. Nonlinear combination can capture the complex relationship between basic features and reveal deep information that a single feature cannot reflect. In this way, a feature set containing multiple fusion features can be generated, and the fusion feature set can more comprehensively reflect the interaction between water quality parameters.

[0112] According to the fusion feature set, constructing the feature matrix is ​​to arrange all the fusion features in a certain format to form a matrix. Assuming there are n target sections and each section has m fusion features, the dimension of the feature matrix is ​​n*m. The construction of the feature matrix needs to ensure that all fusion features of each target section are accurately recorded. Among them, each row of the feature matrix corresponds to a section, and each column corresponds to a fusion feature. In this way, the fusion features of all target sections can be systematically recorded to provide a data basis for subsequent analysis.

[0113] Calculate the covariance matrix of the feature matrix and perform eigenvalue decomposition on the covariance matrix to analyze the correlation between fused features. The covariance matrix is ​​a symmetric matrix whose elements represent the covariance between features. Through eigenvalue decomposition, the covariance matrix can be decomposed into eigenvalues ​​and eigenvectors. The eigenvalue represents the importance of the feature in the data, and the eigenvector represents the direction of the feature.

[0114] The eigenvectors corresponding to the first k largest eigenvalues ​​are selected to form the principal component matrix for feature dimensionality reduction. The principal component matrix is ​​a matrix composed of the first k columns of the eigenvectors, which represents the main direction of the data. Selecting the eigenvectors corresponding to the first k largest eigenvalues ​​can retain the most important information in the data while reducing the dimension of the data. In this way, the dimension of the data can be effectively reduced and the computational efficiency can be improved.

[0115] Project the feature matrix onto the principal component matrix to obtain the reduced dimension feature matrix, which is used to simplify the data structure. Projection is achieved through matrix multiplication, and the calculation formula is Y=XP, where X is the feature matrix, P is the principal component matrix, and Y is the reduced dimension feature matrix. Through projection, the high-dimensional feature matrix can be converted into a low-dimensional feature matrix, retaining the main information of the data while reducing the redundancy of the data. For example, if the dimension of the original feature matrix is ​​n*m and the dimension of the principal component matrix is ​​m*k, the dimension of the reduced dimension feature matrix is ​​n*k. Through dimensionality reduction, the data structure can be simplified and the efficiency of data processing can be improved.

[0116] The feature matrix after dimensionality reduction is used as a feature set including multiple fusion features for subsequent analysis and modeling. The feature matrix after dimensionality reduction retains the main information of the data, while reducing the redundancy and noise of the data, and can more accurately reflect the relationship between water quality parameters. These fusion features can be used as input data and applied to various machine learning algorithms, such as classification and regression, for water quality prediction and analysis, which can effectively improve the accuracy and efficiency of the model and provide a scientific basis for water quality management.

[0117] In one implementation of this embodiment, the step of building a big data analysis model includes the following steps:

[0118] S410, obtaining historical water quality parameters of each river basin, and constructing a river basin mechanism model based on the historical water quality parameters of each river basin, wherein the output of the river basin mechanism model is a function of the river basin and the corresponding historical water quality parameters, and the river basin mechanism model is used to simulate and predict the changes of the output river basin and the corresponding water quality parameters;

[0119] S420, constructing and training a logistic regression model based on historical water quality parameters, wherein the output of the logistic regression model is a function of the historical water quality parameters, and the logistic regression model is used to predict the probability of each water quality parameter reaching the standard;

[0120] S430, linearly combining the output of the watershed mechanism model and the output of the logistic regression model through weight coefficients to obtain the output of the fusion model, wherein the sum of the weight coefficient corresponding to the output of the watershed mechanism model and the weight coefficient corresponding to the output of the logistic regression model is 1;

[0121] S440, using the mean square error as the optimization target, optimizing the weight coefficient, and obtaining the output of the optimized fusion model;

[0122] S450. Using the output of the optimized fusion model as the input of the random forest model to obtain a big data analysis model.

[0123] Based on these historical data, the basin mechanism model can simulate the law of water quality parameters changing over time and environment through mathematical and physical models. According to the historical water quality parameters, a logistic regression model is constructed and trained to predict the probability of each water quality parameter reaching the standard. Among them, the logistic regression model is a statistical model used to predict binary classification results.

[0124] The output of the watershed mechanism model and the output of the logistic regression model are linearly combined through weight coefficients to obtain the output of the fusion model, which can comprehensively utilize the advantages of the two models. Assuming that the output of the watershed mechanism model is M1 and the output of the logistic regression model is M2, the output Mf of the fusion model can be expressed as:

[0125] Mf=ω1M1+ω2M2;

[0126] Among them, ω1 and ω2 are weight coefficients, and they satisfy ω1+ω2=1. Through linear combination, the physical and chemical knowledge of the watershed mechanism model can be combined with the statistical advantages of the logistic regression model to obtain a more accurate and stable prediction model.

[0127] The mean square error is used as the optimization target to optimize the weight coefficient to improve the accuracy of the fusion model. The mean square error (MSE) is an indicator to measure the difference between the predicted result and the true value. Its calculation formula is:

[0128] ;

[0129] In the formula, is the true value, is the predicted value, and n is the number of samples. By minimizing the mean square error, the optimal weight coefficients ω1 and ω2 can be found to minimize the difference between the prediction result of the fusion model and the true value. The optimization process can use optimization algorithms such as the gradient descent method to gradually approach the optimal solution by continuously adjusting the weight coefficients. The output of the optimized fusion model can more accurately reflect the changes in water quality parameters and improve the reliability of the prediction.

[0130] Replacing the output of the random forest model with the output of the optimized fusion model to obtain a big data analysis model can further improve the performance of the model. The random forest model is an integrated learning method that constructs multiple decision trees and votes to obtain the final prediction result. Using the output of the optimized fusion model as the input of the random forest model can further improve the accuracy and stability of the prediction by utilizing the powerful learning ability of the random forest model. In specific implementation, the output of the fusion model can be used as the feature of the random forest model for training and prediction.

[0131] This implementation method can fully utilize the advantages of the fusion model and the random forest model to build an efficient and accurate big data analysis model to provide a scientific basis for water quality management.

[0132] In one implementation of this embodiment, the historical water quality parameters include pH value, dissolved oxygen, turbidity and temperature. The watershed mechanism model is constructed based on the historical water quality parameters of each watershed, including the following steps:

[0133] S510, defining a first support function of pH value, a second support function of dissolved oxygen, a third support function of turbidity, and a fourth support function of temperature of basin i, wherein the support function is used to quantify the contribution of each historical water quality parameter to the water quality of basin i, where i is a positive integer;

[0134] S520, using the parameter weight of the basin i, weighting the first support function, the second support function, the third support function and the fourth support function to obtain the historical support function of the basin i, wherein each basin has different parameter weights for the first support function, the second support function, the third support function and the fourth support function;

[0135] S530, averaging and fusing the historical support functions of all watersheds to obtain a comprehensive support function;

[0136] S540. Use historical water quality parameters to train a linear regression model to obtain the regression coefficient of each water quality parameter change, and construct a watershed mechanism model based on the regression coefficient of each water quality parameter change and the comprehensive support function.

[0137] The first support function of pH value, the second support function of dissolved oxygen, the third support function of turbidity and the fourth support function of temperature of basin i are defined to quantify the contribution of each historical water quality parameter to the water quality of basin i. The support function is a mathematical tool used to describe the degree of influence of a variable on a system.

[0138] In this embodiment, pH values ​​between 6.5 and 8.5 are generally considered suitable for aquatic life. The support function can be defined as a Gaussian function with a center at 7.5 and a standard deviation of 1.

[0139] The first support function indicates that the closer the pH value is to 7.5, the higher the support. pH for:

[0140] ;

[0141] In the formula, X pH is the water quality parameter pH, standard deviation σ pH =1.

[0142] The second support function indicates that the higher the dissolved oxygen concentration, the higher the support. DO for:

[0143] ;

[0144] In the formula, X DO DO is the water quality parameter.

[0145] The third support function indicates that the lower the turbidity, the higher the support. Turb for:

[0146] ;

[0147] In the formula, X Turb is the water quality parameter turbidity Turb.

[0148] The fourth support function indicates that the lower the temperature, the higher the support. T for:

[0149] ;

[0150] In the formula, X T is the water quality parameter temperature T.

[0151] The parameter weights of basin i are used to weight the first support function, the second support function, the third support function and the fourth support function to obtain the historical support function of basin i. The parameter weights are used to reflect the importance of different water quality parameters in the basin. The parameter weights of each basin can be determined by expert knowledge or data-driven methods to ensure that the weighted historical support function can accurately reflect the comprehensive situation of the basin water quality. Through the weighted method, the influence of different water quality parameters can be comprehensively considered to obtain a more comprehensive basin water quality evaluation index.

[0152] Specifically, the comprehensive support function is:

[0153] ;

[0154] Where α1, α2, α3, and α4 are the weight coefficients of each support function, reflecting the relative importance of each water quality parameter to the overall water quality.

[0155] By averaging the historical support functions of all basins to obtain a comprehensive support function, a unified water quality evaluation standard can be constructed. By averaging the historical support functions of all basins, the outliers of individual basins can be smoothed out to obtain a more robust water quality evaluation index. The comprehensive support function can be used to compare the water quality of different basins, help identify basins with poor water quality, and guide water quality management and governance.

[0156] The linear regression model is trained using historical water quality parameters to obtain the regression coefficient of each water quality parameter change, and the basin mechanism model is constructed based on the regression coefficient of each water quality parameter change and the comprehensive support function. The linear regression model is a statistical method used to describe the linear relationship between one or more independent variables and the dependent variable. After obtaining the regression coefficient, the basin mechanism model can be constructed based on the comprehensive support function and the regression coefficient.

[0157] This implementation method can establish a watershed mechanism model based on historical data to simulate and predict changes in water quality parameters and provide a scientific basis for water quality management.

[0158] In one implementation of this embodiment, the mean square error is used as the optimization target to optimize the weight coefficient, and the output of the optimized fusion model is obtained, including the following steps:

[0159] S610, executing the loop step until the mean square error is minimized, and obtaining the output of the optimized fusion model;

[0160] The cycle steps include:

[0161] S620, using a mean square error formula to calculate an initial mean square error;

[0162] S630, calculating the partial derivative of the initial mean square error with respect to the weight coefficient of the output of the watershed mechanism model;

[0163] S640, using a gradient descent formula, updating the weight coefficient of the output of the watershed mechanism model according to the partial derivative, to obtain the updated weight coefficient of the output of the watershed mechanism model;

[0164] S650: Calculate the mean square error according to the weight coefficient of the output of the updated watershed mechanism model.

[0165] The loop step is executed until the mean square error is minimized, and the output of the optimized fusion model is obtained. Mean square error (MSE) is a commonly used evaluation indicator used to measure the difference between the model's predicted value and the actual value. The goal of the optimization process is to minimize the mean square error by adjusting the weight coefficient. Specifically, the loop step includes the following key steps.

[0166] Use the mean square error formula to calculate the initial mean square error;

[0167] Specifically, the formula for the initial mean square error MSE is:

[0168] ;

[0169] Where yi is the actual water quality parameter value, M 融合 ,i is the predicted value of the fusion model.

[0170] Calculate the partial derivative of the initial mean square error with respect to the weight coefficient of the output of the watershed mechanism model. The partial derivative reflects the rate of change of the mean square error relative to each weight coefficient and is used to determine the influence of each weight coefficient on the mean square error, thereby providing an update direction for the gradient descent method;

[0171] Calculate the partial derivative of the mean square error with respect to the weight coefficient ω flow and ω logistic. Since ω logistic = 1-ω flow, only one weight needs to be optimized:

[0172] ;

[0173] Using the gradient descent formula, the weight coefficient of the output of the watershed mechanism model is updated according to the partial derivative to obtain the updated weight coefficient. The gradient descent method is an iterative optimization algorithm that gradually reduces the mean square error by updating the weight coefficient along the negative gradient direction.

[0174] Specifically, the gradient descent method is used to update the weight coefficient ω domain:

[0175] ;

[0176] Where ɑ is the learning rate.

[0177] Calculate the new mean square error based on the weight coefficients of the output of the updated watershed mechanism model. Each time the weight coefficients are updated, the mean square error needs to be recalculated to evaluate the performance of the updated model. By comparing the new mean square error with the previous mean square error, it can be determined whether the model is moving in the optimization direction. If the new mean square error is smaller than the previous mean square error, it means that the model performance has improved; otherwise, the learning rate or other parameters may need to be adjusted.

[0178] This implementation is repeated until the mean square error reaches a minimum or converges to a stable value. Through continuous iteration and optimization, a fusion model with an optimal weight coefficient can be obtained to minimize the prediction error of the model. This process not only improves the prediction accuracy of the model, but also enhances the robustness and generalization ability of the model. The optimized fusion model can more accurately reflect the changes in water quality parameters in the basin, providing a more reliable basis for water quality management and decision-making.

[0179] In one implementation of this embodiment, analyzing key water quality indicators to determine the concentration trend of key water quality indicators in the watershed where each target section is located includes the following steps:

[0180] S710, performing time series analysis on key water quality indicators in the river basin where each target section is located, and obtaining concentration trends of key water quality indicators in the river basin where each target section is located;

[0181] After analyzing the key water quality indicators and determining the concentration trend of the key water quality indicators in the basin where each target section is located, it also includes:

[0182] S720. Draw a moving average graph and a trend line graph based on the concentration trend of key water quality indicators in the watershed where each target section is located.

[0183] In this embodiment, firstly, a time series analysis is performed on the key water quality indicators of the watershed where each target section is located to obtain the concentration trend of these indicators. Time series analysis is a statistical method used to analyze data that changes over time. Among them, the key water quality indicators are time series data.

[0184] Specifically, time series analysis can be performed by calculating the moving average of key water quality indicators and smoothing the data to characterize the concentration change trend of key water quality indicators, and fitting trend lines to quantify the concentration change trend.

[0185] After completing the time series analysis, draw a moving average graph and a trend line graph based on the concentration trend of the key water quality indicators in the basin where each target section is located. The moving average graph is a method of smoothing time series data. By taking the average value of the data within a certain time window, short-term fluctuations are eliminated and long-term trends are highlighted. Assuming that the time window is k, the calculation formula for the moving average is:

[0186] ;

[0187] In the formula, MA t is the moving average at time t, x i is the observed value at the i-th moment. The long-term trend of water quality indicators can be observed intuitively through the moving average graph. The trend line graph describes the overall trend of data by fitting a trend line.

[0188] This implementation method can systematically analyze key water quality indicators, determine the concentration trend of key water quality indicators in the basin where each target section is located, and intuitively display them through moving average graphs and trend line graphs. This process not only helps to understand the changing patterns of water quality indicators, but also provides a scientific basis for water quality management and governance. For example, if the ammonia nitrogen concentration in a basin shows an upward trend, timely measures can be taken to control the pollution source and improve the water quality. At the same time, through moving average graphs and trend line graphs, the effectiveness of governance measures can be monitored, the water quality improvement can be evaluated, and the achievement of water quality management goals can be ensured.

[0189] In one implementation of this embodiment, determining the change trend and diffusion path of key water quality indicators along the route between different target sections includes the following steps:

[0190] S810, obtaining the geographical location of each target section;

[0191] S820, using Kriging interpolation to construct a Kriging equation according to the geographical location of each target section, wherein the Kriging equation is used to calculate the estimated value of the key water quality index of the final target section;

[0192] S830, generating an interpolation network for the watershed i, estimating key water quality indicators of each grid point in the interpolation network according to the Kriging equation, and obtaining an interpolation network including multiple estimation results;

[0193] S840. Draw a contour map of watershed i based on an interpolation network including multiple estimation results, wherein the contour map is used to display the spatial distribution of key water quality indicators in watershed i, and the spatial distribution includes the along-the-way variation trend and diffusion path of the key water quality indicators between different target sections.

[0194] Geographical location is usually expressed as longitude and latitude coordinates, which can be obtained through GPS devices or geographic information systems (GIS). The Kriging interpolation method is used to construct the Kriging equation according to the geographical location of each target section. Kriging interpolation is a geostatistical method used to estimate the value of spatial data. The basic idea is to estimate the value of unknown points based on the numerical value and spatial position of known points. Generate an interpolation network for basin i, and estimate the key water quality indicators of each grid point in the interpolation network according to the Kriging equation. The interpolation network is a two-dimensional space composed of multiple grid points, each of which has its geographical location. Through the Kriging interpolation method, the water quality index value of each grid point can be estimated, and an interpolation network including multiple estimation results can be obtained.

[0195] Based on the interpolation network including multiple estimation results, a contour map of basin i is drawn. A contour map is a graph used to display the distribution of spatial data, and contour lines connect points with the same values. The contour map can intuitively display the spatial distribution of key water quality indicators in the basin, including the along-the-way variation trend and diffusion path between different target sections. The steps for drawing a contour map include: first, determining the interval of the contour lines, and selecting an appropriate contour line interval based on the range of estimated values ​​in the interpolation network. Then, using the estimated values ​​in the interpolation network, draw contour lines to connect points with the same values. Finally, mark the values ​​of the contour lines to form a complete contour map.

[0196] Specifically, this embodiment adopts Kriging interpolation, and the steps of constructing the Kriging equation according to the geographical location of each target section include:

[0197] S910. Calculate the semivariogram between every two target sections to describe the spatial correlation of the data.

[0198] The semivariogram function is:

[0199] ;

[0200] Where γ(h) is the semivariogram value at distance h, N(h) is the number of point pairs at distance h, and Z(x i ) is the position x i Key water quality indicators of the department.

[0201] S920, fitting a semivariogram function according to a Gaussian model to obtain a fitted semivariogram model;

[0202] The fitted semivariogram model is:

[0203] ;

[0204] Where C0 is the base, C is the base rise, and a is the range.

[0205] S930. Based on the fitted semi-variogram model, a Kriging equation is constructed to calculate the estimated values ​​of key water quality indicators for non-target sections.

[0206] The Kriging equation is:

[0207] ;

[0208] In the formula, Z * (x0) is the estimated value at the unknown x0, λ i are the kriging weights, satisfying the following system of equations:

[0209] ;

[0210] Where μ is the Lagrange multiplier.

[0211] For example, assuming that the key water quality indicator is ammonia nitrogen concentration, first, calculate the semivariogram: calculate the semivariogram between each pair of monitoring points. Secondly, fit the semivariogram model: select the Gaussian model to fit the semivariogram. In addition, generate an interpolation grid in the monitoring area, use Kriging interpolation to estimate the ammonia nitrogen concentration at each grid point, and draw a contour map based on the interpolation results. At the same time, construct a diffusion model to simulate the diffusion path of ammonia nitrogen in the basin, and draw a diffusion path map in combination with flow analysis. Among them, the diffusion model can be a two-dimensional plane diffusion model or a one-dimensional river diffusion model. Set initial conditions and boundary conditions, simulate the diffusion process of ammonia nitrogen in the basin, and use numerical methods (such as finite difference method or finite element method) to solve the diffusion equation to obtain the ammonia nitrogen concentration distribution at each time step. Finally, analyze the diffusion path of ammonia nitrogen in the basin, and draw an ammonia nitrogen diffusion path map based on the simulation results and flow analysis.

[0212] This implementation method can systematically analyze the changing trends and diffusion paths of key water quality indicators along the different target sections. The Kriging interpolation method is used to construct the Kriging equation, generate the interpolation network, and draw contour maps, which can intuitively display the spatial distribution of key water quality indicators. It not only helps to understand the spatial variation laws of water quality indicators, but also provides a scientific basis for water quality management and governance. For example, through contour maps, high-concentration areas and diffusion paths of water quality indicators can be identified, and timely measures can be taken to control pollution sources and improve water quality. At the same time, by monitoring the changing trends of key water quality indicators, the effectiveness of governance measures can be evaluated to ensure the achievement of water quality management goals.

[0213] Reference Figure 4 Another architecture diagram of an Internet of Things platform is shown in FIG. 1 . The embodiment of the present application further provides an Internet of Things platform, including:

[0214] A data acquisition module is used to obtain the water quality parameter set of each target section in real time;

[0215] A feature extraction module is used to extract features from the water quality parameter set of each target section to obtain a water quality feature set for the watershed where each target section is located;

[0216] The prediction module is used to input the water quality feature set of each target section into the pre-built big data analysis model to predict and obtain the key water quality indicators of the basin where each target section is located;

[0217] The analysis module is used to analyze key water quality indicators, determine the concentration trend of key water quality indicators in the basin where each target section is located, and determine the change trend and diffusion path of key water quality indicators along the way between different target sections;

[0218] Early warning module, used to issue water quality early warning based on concentration trend and change trend along the way;

[0219] Among them, the big data analysis model is obtained by integrating the pre-built watershed mechanism model and the pre-built machine learning model.

[0220] Reference Figure 5 It provides another interface display diagram of an Internet of Things platform (i.e., XX regional water environment quality monitoring data management platform), which includes multiple monitoring areas, monitoring start time, monitoring end time, water quality situation diagram of the national controlled section of surface water in one of the monitoring areas in a certain time period, and early warning icons.

[0221] Among them, this embodiment divides water quality warnings into Class I, Class II, Class III, Class IV, Class V and Class V-minus, which are used to characterize the water quality of a certain river basin.

[0222] The embodiment of the present application further provides an electronic device, on which the above-mentioned Internet of Things platform is deployed, and the electronic device includes:

[0223] a memory configured to store instructions; and

[0224] The processor is configured to call instructions from the memory and implement the above-mentioned water environment management method based on smart Internet of Things and big data analysis when executing the instructions.

[0225] An embodiment of the present application also provides a machine-readable storage medium, on which instructions are stored, and the instructions are used to enable a machine to execute the above-mentioned water environment management method based on smart Internet of Things and big data analysis.

[0226] 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 take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may take 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.

[0227] 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 block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks 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.

[0228] 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.

[0229] 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.

[0230] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0231] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0232] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0233] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0234] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A water environment management method based on smart Internet of Things and big data analysis, characterized in that: Applied to an electronic device, the electronic device is deployed with an Internet of Things platform, and the method includes: Acquire a water quality parameter set of each target section in real time through the Internet of Things platform; Performing feature extraction on the water quality parameter set of each target section to obtain a water quality feature set of the watershed where each target section is located; Input the water quality feature set of each target section into a pre-built big data analysis model to predict and obtain key water quality indicators of the basin where each target section is located; Analyze the key water quality indicators, determine the concentration trend of the key water quality indicators in the watershed where each target section is located, and determine the along-the-way variation trend and diffusion path of the key water quality indicators between different target sections; Conduct water quality warning according to the concentration trend and the change trend along the route; Wherein, the big data analysis model is obtained by integrating a pre-built watershed mechanism model and a pre-built machine learning model; The steps of constructing the big data analysis model include: Obtaining historical water quality parameters of each of the river basins, and constructing a river basin mechanism model based on the historical water quality parameters of each of the river basins, wherein the output of the river basin mechanism model is a function of the river basin and the corresponding historical water quality parameters, and the river basin mechanism model is used to simulate and predict changes in the output river basin and the corresponding water quality parameters; A logistic regression model is constructed and trained according to the historical water quality parameters, wherein the output of the logistic regression model is a function of the historical water quality parameters, and the logistic regression model is used to predict the probability of each of the water quality parameters meeting the standard; The output of the watershed mechanism model and the output of the logistic regression model are linearly combined by weight coefficients to obtain the output of the fusion model, wherein the sum of the weight coefficient corresponding to the output of the watershed mechanism model and the weight coefficient corresponding to the output of the logistic regression model is 1; Using mean square error as an optimization target, optimizing the weight coefficient, and obtaining an output of an optimized fusion model; The output of the optimized fusion model is used as the input of the random forest model to obtain a big data analysis model.

2. The method according to claim 1, characterized in that The feature extraction of the water quality parameter set of each target section to obtain the water quality feature set of the watershed where each target section is located includes: Extracting and summarizing the basic features of the water quality parameter set of each target section to obtain a basic feature set; Performing feature fusion on any two different basic features of the same target section in the basic feature set to obtain a feature set including a plurality of fused features; Performing feature crossover on each of the fused features in the same watershed in the feature set including the plurality of fused features, to obtain a feature set including the plurality of crossover features; Feature selection is performed on the feature set including a plurality of the cross features to obtain a water quality feature set for each of the river basins.

3. The method according to claim 2, characterized in that The step of fusing any two different basic features of the same target section in the basic feature set to obtain a feature set including a plurality of fused features includes: Nonlinearly combining any two different basic features of the same target section in the basic feature set to obtain a fused feature set; Constructing a feature matrix according to the fused feature set; Calculating a covariance matrix of the characteristic matrix, and performing eigenvalue decomposition on the covariance matrix to obtain a plurality of eigenvalues ​​and an eigenvector corresponding to each eigenvalue; Select the eigenvectors corresponding to the first k largest eigenvalues ​​to form a principal component matrix; Projecting the feature matrix onto the principal component matrix to obtain a feature matrix after dimension reduction; The feature matrix after dimension reduction is used as a feature set including multiple fusion features.

4. The method according to claim 1, characterized in that: The historical water quality parameters include pH value, dissolved oxygen, turbidity and temperature. The watershed mechanism model is constructed based on the historical water quality parameters of each watershed, including: Define a first support function of the pH value of the basin i, a second support function of the dissolved oxygen, a third support function of the turbidity, and a fourth support function of the temperature, wherein the support function is used to quantify the contribution of each of the historical water quality parameters to the water quality of the basin i, and i is a positive integer; Using the parameter weight of the watershed i, weighting the first support function, the second support function, the third support function and the fourth support function, to obtain the historical support function of the watershed i, wherein each watershed has different parameter weights for the first support function, the second support function, the third support function and the fourth support function; Averagely merge the historical support functions of all the river basins to obtain a comprehensive support function; The historical water quality parameters are used to train a linear regression model to obtain the regression coefficient of each water quality parameter change, and a watershed mechanism model is constructed based on the regression coefficient of each water quality parameter change and the comprehensive support function.

5. The method according to claim 1, characterized in that The method of using the mean square error as an optimization target, optimizing the weight coefficient, and obtaining the output of the optimized fusion model includes: Execute the loop steps until the mean square error is minimized and obtain the output of the optimized fusion model; Wherein, the circulation step comprises: Use the mean square error formula to calculate the initial mean square error; Calculating the partial derivative of the initial mean square error with respect to the weight coefficient of the output of the watershed mechanism model; Using a gradient descent formula, updating the weight coefficient of the output of the watershed mechanism model according to the partial derivative, to obtain an updated weight coefficient of the output of the watershed mechanism model; The mean square error is calculated based on the updated weight coefficient of the output of the watershed mechanism model.

6. The method according to claim 1, characterized in that The analyzing of the key water quality indicators to determine the concentration trend of the key water quality indicators in the watershed where each target section is located includes: Performing time series analysis on key water quality indicators in the river basin where each target section is located to obtain concentration trends of key water quality indicators in the river basin where each target section is located; After analyzing the key water quality indicators and determining the concentration trend of the key water quality indicators in the river basin where each target section is located, the method further includes: According to the concentration trend of key water quality indicators in the watershed where each target section is located, a moving average graph and a trend line graph are drawn.

7. The method according to claim 1, characterized in that Determining the change trend and diffusion path of the key water quality indicators along the different target sections includes: Obtaining the geographical location of each target section; Using Kriging interpolation, constructing a Kriging equation according to the geographical location of each target section, wherein the Kriging equation is used to calculate the estimated value of the key water quality index of the final target section; Generate an interpolation network for the watershed i, estimate the key water quality index of each grid point in the interpolation network according to the Kriging equation, and obtain an interpolation network including a plurality of estimation results; According to the interpolation network including multiple estimation results, a contour map of the basin i is drawn, wherein the contour map is used to show the spatial distribution of key water quality indicators in the basin i, and the spatial distribution includes the along-the-way variation trend and diffusion path of the key water quality indicators between different target sections.

8. An Internet of Things platform, characterized in that: include: A data acquisition module is used to obtain the water quality parameter set of each target section in real time; A feature extraction module is used to extract features from a water quality parameter set of each target section to obtain a water quality feature set of the watershed where each target section is located; A prediction module, used to input the water quality feature set of each target section into a pre-built big data analysis model to predict and obtain key water quality indicators of the basin where each target section is located; An analysis module, used to analyze the key water quality indicators, determine the concentration trend of the key water quality indicators in the basin where each target section is located, and determine the along-the-way change trend and diffusion path of the key water quality indicators between different target sections; An early warning module, used for making a water quality early warning according to the concentration trend and the change trend along the process; Wherein, the big data analysis model is obtained by integrating a pre-built watershed mechanism model and a pre-built machine learning model; The steps of constructing the big data analysis model include: Obtaining historical water quality parameters of each of the river basins, and constructing a river basin mechanism model based on the historical water quality parameters of each of the river basins, wherein the output of the river basin mechanism model is a function of the river basin and the corresponding historical water quality parameters, and the river basin mechanism model is used to simulate and predict changes in the output river basin and the corresponding water quality parameters; A logistic regression model is constructed and trained according to the historical water quality parameters, wherein the output of the logistic regression model is a function of the historical water quality parameters, and the logistic regression model is used to predict the probability of each of the water quality parameters meeting the standard; The output of the watershed mechanism model and the output of the logistic regression model are linearly combined by weight coefficients to obtain the output of the fusion model, wherein the sum of the weight coefficient corresponding to the output of the watershed mechanism model and the weight coefficient corresponding to the output of the logistic regression model is 1; Using mean square error as an optimization target, optimizing the weight coefficient, and obtaining an output of an optimized fusion model; The output of the optimized fusion model is used as the input of the random forest model to obtain a big data analysis model.

9. An electronic device, characterized in that: The electronic device is deployed with the Internet of Things platform as claimed in claim 8, and the electronic device includes: a memory configured to store instructions; and A processor is configured to call the instructions from the memory and to implement the water environment management method based on smart Internet of Things and big data analysis according to any one of claims 1 to 7 when executing the instructions.