Sea-entering estuary water quality monitoring automatic early warning network construction method

Abstract

The invention relates to the technical field of water quality early warning, in particular to a sea-entering estuary water quality monitoring automatic early warning network construction method, which comprises the following steps of data acquisition, data preprocessing, model construction, model training and verification and model application. According to the method for constructing the automatic early warning network for monitoring the water quality of the estuary, the real-time monitoring of the water quality of the estuary is realized through an automatic monitoring sensor network, the time delay of a traditional method is overcome, and early warning can be quickly given out at the initial stage of a water quality problem by adopting an advanced data processing and early warning algorithm; compared with traditional manual sampling and laboratory analysis, the technical scheme has the advantages that the monitoring cost is remarkably reduced, the working efficiency is improved, interference to river ecology is reduced, friendly monitoring on the natural environment is realized, and the water quality monitoring system can be widely applied to the field of water quality monitoring and early warning by providing detailed water quality data and analysis reports. And a scientific basis is provided for water resource management and environmental protection decision making.

Description

Technical Field

[0001] The present invention relates to the technical field of water quality early warning, and in particular to a method for constructing an automatic early warning network for water quality monitoring at an estuary entering the sea. Background Art

[0002] The water quality of estuaries, as the confluence of rivers and oceans, has an important impact on the ecological environment and human activities. However, since estuary water quality is affected by multiple factors, such as hydrometeorology, land pollution emissions, ocean tides, etc., monitoring and early warning of estuary water quality has always been a complex and challenging task.

[0003] Traditional water quality monitoring methods for estuaries usually rely on manual sampling and laboratory analysis, which is not only inefficient but also difficult to achieve real-time monitoring and rapid early warning.

[0004] With the development of science and technology, especially the advancement of Internet of Things technology and automated monitoring technology, it has become possible to build an efficient and real-time water quality monitoring and early warning network for estuaries entering the sea. Summary of the invention

[0005] The purpose of the present invention is to provide a method for constructing an automated early warning network for water quality monitoring at estuaries entering the sea, which solves the technical problem that existing methods for monitoring water quality at estuaries entering the sea usually rely on manual sampling and laboratory analysis, which is not only inefficient but also difficult to achieve real-time monitoring and rapid early warning.

[0006] To achieve the above object, the present invention provides a method for constructing an automatic early warning network for water quality monitoring at an estuary entering the sea, comprising the following steps:

[0007] Data collection: collect water quality information of rivers entering the sea on a regular or real-time basis, and transmit the collected data to the data center;

[0008] Data preprocessing, screening the received data, and after screening, performing dimensionality reduction processing on the screened data to determine the specified factors;

[0009] Model construction, first determine the key factors affecting water quality, then determine the thresholds of the key factors, and explore the impact of the associations between the key factors on the water quality, and finally, conduct real-time automatic monitoring of the data and issue warnings;

[0010] Model training and verification: using historical water quality data to train the model and optimize model parameters, analyzing the error between the model's prediction results and actual monitoring data, and evaluating the accuracy of the model;

[0011] The model is applied by automatically accessing the database interface of the real-time incoming data through the system and obtaining the data. After processing and analyzing the data, it is determined whether the system triggers the early warning mechanism according to the analysis results.

[0012] The data collection includes periodic or real-time data collection of the water quality of the estuary and transmitting the collected data to the data center. The specific steps are:

[0013] Using drones to obtain real-time remote sensing image data of key areas of estuaries, and then collecting historical data of the key areas for model construction and training;

[0014] Afterwards, a series of water quality monitoring sensors are deployed in the key areas to collect key water quality parameters of the key areas in real time;

[0015] Then, the water quality data of the key area is collected regularly or in real time using automated monitoring equipment, and the collected water quality parameters and water quality data are transmitted to the data center via wireless communication technology.

[0016] The data preprocessing includes screening the received data and performing dimensionality reduction processing on the screened data after screening. The specific steps of determining the specified factors are as follows:

[0017] Preprocessing the remote sensing image data acquired by the drone and calculating NDVI to obtain the NDVI data as a factor for monitoring the overall water quality;

[0018] Screening the collected historical water quality data to find out the peak value, mean value, range value and real-time value of the indicator stage, and forming a factor database suitable for data mining;

[0019] Afterwards, the filtered data is subjected to dimensionality reduction processing to determine the index factors.

[0020] The model is constructed by first determining the key factors affecting water quality, then determining the thresholds of the key factors, and exploring the impact of the associations between the key factors on the water quality. Finally, the specific steps of automatically monitoring and warning the data in real time are as follows:

[0021] Determine the key factors affecting water quality by constructing a neural network and using the neural network algorithm to analyze the pre-processed NDVI data and water quality data to obtain the key factors affecting water quality;

[0022] Determine the threshold of the key factor, discretize the historical data of the key factor using the DBSCAN clustering algorithm, and segment the data. In the segmented data, select the corresponding numerical segment as the threshold range according to the value of the time data of the water body eutrophication in the historical data, and label the numerical segment with A, B, C...;

[0023] Mining the influence of the association between factors on water quality, using FP-growth algorithm to mine the association between each of the key factors, and summarizing the association rules between all the key factors, mining the influence of the association between the key factors on the water quality;

[0024] Real-time data is automatically monitored and warned, and an automatic warning model is established using a decision tree algorithm. When real-time data is transmitted to the system, decision tree rules are set. That is, if it exceeds the threshold range or meets the association rules, the system will give feedback and issue a warning. If the above rules are not met, the system will not give feedback and issue a warning.

[0025] The key factors are determined by using a multi-input single-output feedforward neural network, dividing it into several levels, including an input layer, a hidden layer and an output layer. When training the neural network, the prediction value of the model is measured by a loss function.

[0026] Among them, the basic structure of the key factors is as follows:

[0027] Input Layer:

[0028] X=[X 1 , X 2 ,......X n ] T

[0029] Where n is the number of input water quality parameters.

[0030] Hidden Layer:

[0031] Z l =W l A l-1 +b l

[0032] A l =g l (Z l )

[0033] Among them, l means there are l hidden layers, W l is the weight matrix of the lth layer, b l is the bias vector of the lth layer, A l-1 is the activation output of the previous layer, g l() is the activation function of the lth layer, commonly used ones are ReLU and sigmoid.

[0034] Output layer:

[0035]

[0036] Among them, W l+1 and b l+1 are the weight matrix and bias vector of the output layer, A l is the activation output of the last hidden layer.

[0037] Loss function:

[0038]

[0039] Where N is the number of samples.

[0040] The model application automatically accesses the database interface of the real-time data input through the system and obtains the data. After processing and analyzing the data, the specific steps of determining whether the system triggers the early warning mechanism according to the analysis results are as follows:

[0041] The system automatically accesses the database interface of the real-time incoming data and obtains the data, and performs processing and analysis;

[0042] After processing and analyzing the data, when the monitoring data reaches a preset threshold or the analysis result indicates a potential risk, the system will automatically trigger an early warning mechanism and determine whether to send an early warning message based on the final calculation result;

[0043] If it is determined that the warning message needs to be sent, relevant managers and decision makers will be notified via SMS, email or mobile application to achieve warning.

[0044] The present invention discloses a method for constructing an automated early warning network for monitoring water quality at estuaries entering the sea. The method realizes real-time monitoring of estuary water quality through an automated monitoring sensor network, overcomes the time delay of traditional methods, and can quickly issue early warnings at the early stages of water quality problems by adopting advanced data processing and early warning algorithms, thereby winning precious time for taking countermeasures. Compared with traditional manual sampling and laboratory analysis, the technical solution of the patented invention significantly reduces monitoring costs, improves work efficiency, reduces interference with river ecology, and realizes friendly monitoring of the natural environment. By providing detailed water quality data and analysis reports, it provides scientific information for water resources management and environmental protection decision-making, and solves the technical problem that existing methods for monitoring water quality at estuaries entering the sea usually rely on manual sampling and laboratory analysis, which is not only inefficient but also difficult to achieve real-time monitoring and rapid early warning. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art are briefly introduced below.

[0046] Figure 1 It is a flow chart of establishing an automated early warning network for monitoring water quality at an estuary entering the sea according to the first embodiment of the present invention.

[0047] Figure 2 It is a flowchart of the method for constructing an automated early warning network for monitoring water quality at an estuary according to the first embodiment of the present invention.

[0048] Figure 3 This is a specific step diagram of data collection in the first embodiment of the present invention, which collects water quality conditions at estuaries at regular or real-time intervals and transmits the collected data to a data center.

[0049] Figure 4 This is a specific step diagram of data preprocessing of the first embodiment of the present invention, which includes screening the received data, and after screening, performing dimensionality reduction processing on the screened data to determine the specified factors.

[0050] Figure 5 This is a model construction of the first embodiment of the present invention, which first determines the key factors affecting water quality, then determines the thresholds of the key factors, and explores the impact of the correlation between the key factors on the water quality. Finally, a specific step diagram is provided for real-time automatic monitoring and warning of the data.

[0051] Figure 6 This is a specific step diagram of the model application of the first embodiment of the present invention, in which the system automatically accesses the database interface of the real-time incoming data and obtains the data. After processing and analyzing the data, it is determined whether the system triggers the early warning mechanism according to the analysis results. DETAILED DESCRIPTION

[0052] Embodiments of the present invention are described in detail below. Examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, but should not be construed as limiting the present invention.

[0053] First embodiment:

[0054] See also Figures 1 to 6 The present invention provides a method for constructing an automatic early warning network for monitoring water quality at an estuary entering the sea, comprising the following steps:

[0055] S1. Data collection: collect water quality information of rivers entering the sea on a regular or real-time basis, and transmit the collected data to the data center.

[0056] Specifically, the data collection is to collect the water quality of the river estuary at regular intervals or in real time, and transmit the collected data to the data center in the following specific steps:

[0057] S11. Use drones to obtain real-time remote sensing image data of key areas of estuaries, and then collect historical data in the key areas for model construction and training.

[0058] Specifically, the historical data in the key area includes long-term NDVI data and water quality data.

[0059] S12. Afterwards, a series of water quality monitoring sensors are deployed in the key area to collect key water quality parameters of the key area in real time.

[0060] Specifically, the water quality monitoring sensor is used to collect key water quality parameters such as pH, temperature, turbidity, dissolved oxygen, conductivity, COD, and ammonia nitrogen in real time.

[0061] S13, using automated monitoring equipment to collect water quality data of the key area regularly or in real time, and transmitting the collected water quality parameters and water quality data to the data center through wireless communication technology.

[0062] Specifically, the automated monitoring equipment is a buoy or a fixed station.

[0063] S2. Data preprocessing: filtering the received data, and after filtering, performing dimensionality reduction processing on the filtered data to determine the specified factors.

[0064] Specifically, the data preprocessing includes screening the received data, and after screening, performing dimensionality reduction processing on the screened data, and determining the specific steps of the specified factors are:

[0065] S21. Preprocess the remote sensing image data acquired by the drone and calculate NDVI to obtain the NDVI data as a factor for monitoring the overall water quality.

[0066] Specifically, the preprocessing of the remote sensing image data includes radiation calibration, atmospheric correction and image cropping.

[0067] S22, screening the collected historical water quality data to screen out the peak value, mean value, range value and real-time value of the indicator stage, and forming a factor database suitable for data mining;

[0068] S23. Afterwards, the filtered data is subjected to dimensionality reduction processing to determine the index factors.

[0069] Specifically, the indicator factors include water temperature, pH value, dissolved oxygen, turbidity and nutrient salt concentration.

[0070] S3. Model construction: first determine the key factors that affect water quality, then determine the thresholds of the key factors, and explore the impact of the correlation between the key factors on the water quality. Finally, automatically monitor the data in real time and issue warnings.

[0071] Specifically, the model is constructed by first determining the key factors affecting water quality, then determining the thresholds of the key factors, and exploring the impact of the associations between the key factors on the water quality. Finally, the specific steps of automatically monitoring the data in real time and issuing warnings are as follows:

[0072] S31, determining the key factors affecting water quality, by constructing a neural network and analyzing the pre-processed NDVI data and water quality data using a neural network algorithm to obtain the key factors affecting water quality;

[0073] S32, determining the threshold of the key factor, using the DBSCAN clustering algorithm to discretize the historical data of the key factor, and segmenting the data, in the segmented data, according to the value of the time data of the water body eutrophication in the historical data, selecting the corresponding numerical segment as the threshold range, and labeling the numerical segment with A, B, C...;

[0074] S33, mining the influence of the association between factors on water quality, using FP-growth algorithm to mine the association between each of the key factors, and summarizing the association rules between all the key factors, mining the influence of the association between the key factors on the water quality;

[0075] S34, real-time data is automatically monitored and warned, and an automatic warning model is established using a decision tree algorithm. When real-time data is transmitted to the system, a decision tree rule is set. That is, when the data exceeds the threshold range or meets the association rule, the system gives feedback and issues a warning. If the above rules are not met, the system will not give feedback and issue a warning.

[0076] Specifically, the key factors are determined by using a multi-input single-output feedforward neural network, dividing it into several levels, including an input layer, a hidden layer and an output layer, and measuring the prediction value of the model through a loss function when training the neural network.

[0077] The basic structure of the key factor is as follows:

[0078] Input Layer:

[0079] X=[X 1 , X 2,......X n ] T

[0080] Where n is the number of input water quality parameters.

[0081] Hidden Layer:

[0082] Z l =W l A l-1 +b l

[0083] A l =g l (Z l )

[0084] Among them, l means there are l hidden layers, W l is the weight matrix of the lth layer, b l is the bias vector of the lth layer, is the activation output of the previous layer, and g l () is the activation function of the lth layer, commonly used ones are ReLU and sigmoid.

[0085] Output layer:

[0086]

[0087] Among them, W l+1 and b l+1 are the weight matrix and bias vector of the output layer, A l is the activation output of the last hidden layer.

[0088] Loss function:

[0089]

[0090] Where N is the number of samples.

[0091] S4. Model training and verification: Use historical water quality data to train the model and optimize model parameters, analyze the error between the model's prediction results and actual monitoring data, and evaluate the accuracy of the model.

[0092] Specifically,

[0093] S5. Model application: the system automatically accesses the database interface of the real-time incoming data and obtains the data. After processing and analyzing the data, it is determined whether the system triggers the early warning mechanism according to the analysis results.

[0094] Specifically, the model application automatically accesses the database interface of the real-time incoming data through the system and obtains the data. After processing and analyzing the data, the specific steps of determining whether the system triggers the early warning mechanism according to the analysis results are as follows:

[0095] S51, automatically accessing the database interface of the real-time incoming data through the system and obtaining the data, and performing processing and analysis;

[0096] S52. After processing and analyzing the data, when the monitoring data reaches a preset threshold or the analysis result indicates a potential risk, the system will automatically trigger an early warning mechanism and determine whether to send an early warning message based on the final calculation result;

[0097] S53. If it is determined that the warning message needs to be sent, relevant managers and decision makers are notified via SMS, email or mobile application to achieve warning.

[0098] When using the method for constructing an automated early warning network for monitoring water quality at an estuary entering the sea according to the present embodiment, real-time monitoring of the water quality at the estuary is achieved through an automated monitoring sensor network, overcoming the time delay of traditional methods. By adopting advanced data processing and early warning algorithms, early warnings can be quickly issued at the early stage of water quality problems, thereby gaining valuable time for taking countermeasures. Compared with traditional manual sampling and laboratory analysis, the technical solution of this patent significantly reduces monitoring costs, improves work efficiency, reduces interference with river ecology, and achieves friendly monitoring of the natural environment. By providing detailed water quality data and analysis reports, it provides scientific information for water resources management and environmental protection decision-making, and solves the technical problem that existing methods for monitoring water quality at estuaries entering the sea usually rely on manual sampling and laboratory analysis, which is not only inefficient but also difficult to achieve real-time monitoring and rapid early warning.

[0099] What is disclosed above is only one or more preferred embodiments of the present application, and cannot be used to limit the scope of rights of the present application. Ordinary technicians in this field can understand that all or part of the processes of implementing the above embodiments and equivalent changes made according to the claims of the present application are still within the scope covered by the present application.

Claims

1. A method for constructing an automated early warning network for water quality monitoring at an estuary, characterized in that: The following steps are involved: Data collection: collect water quality information of rivers entering the sea on a regular or real-time basis, and transmit the collected data to the data center; Data preprocessing, screening the received data, and after screening, performing dimensionality reduction processing on the screened data to determine the specified factors; Model construction, first determine the key factors affecting water quality, then determine the thresholds of the key factors, and explore the impact of the associations between the key factors on the water quality, and finally, conduct real-time automatic monitoring of the data and issue warnings; Model training and verification: using historical water quality data to train the model and optimize model parameters, analyzing the error between the model's prediction results and actual monitoring data, and evaluating the accuracy of the model; The model is applied by automatically accessing the database interface of the real-time incoming data through the system and obtaining the data. After processing and analyzing the data, it is determined whether the system triggers the early warning mechanism according to the analysis results.

2. The method for constructing an automated early warning network for monitoring water quality at an estuary according to claim 1, characterized in that: The data collection is to collect the water quality of the river estuary regularly or in real time, and transmit the collected data to the data center in the following specific steps: Using drones to obtain real-time remote sensing image data of key areas of estuaries, and then collecting historical data of the key areas for model construction and training; Afterwards, a series of water quality monitoring sensors are deployed in the key areas to collect key water quality parameters of the key areas in real time; Then, the water quality data of the key area is collected regularly or in real time using automated monitoring equipment, and the collected water quality parameters and water quality data are transmitted to the data center via wireless communication technology.

3. The method for constructing an automated early warning network for monitoring water quality at an estuary according to claim 2, characterized in that: The data preprocessing includes screening the received data and, after screening, performing dimensionality reduction processing on the screened data. The specific steps of determining the specified factors are as follows: Preprocessing the remote sensing image data acquired by the drone and calculating NDVI to obtain the NDVI data as a factor for monitoring the overall water quality; Screening the collected historical water quality data to find out the peak value, mean value, range value and real-time value of the indicator stage, and forming a factor database suitable for data mining; Afterwards, the filtered data is subjected to dimensionality reduction processing to determine the index factors.

4. The method for constructing an automated early warning network for monitoring water quality at an estuary according to claim 1, characterized in that: The model is constructed by first determining the key factors that affect water quality, then determining the thresholds of the key factors, and exploring the impact of the associations between the key factors on the water quality. Finally, the specific steps of automatically monitoring and warning the data in real time are as follows: Determine the key factors affecting water quality by constructing a neural network and using the neural network algorithm to analyze the pre-processed NDVI data and water quality data to obtain the key factors affecting water quality; Determine the threshold of the key factor, discretize the historical data of the key factor using the DBSCAN clustering algorithm, and segment the data. In the segmented data, select the corresponding numerical segment as the threshold range according to the value of the time data of the water body eutrophication in the historical data, and label the numerical segment with A, B, C...; Mining the influence of the association between factors on water quality, using FP-growth algorithm to mine the association between each of the key factors, and summarizing the association rules between all the key factors, mining the influence of the association between the key factors on the water quality; Real-time data is automatically monitored and warned, and an automatic warning model is established using a decision tree algorithm. When real-time data is transmitted to the system, decision tree rules are set. That is, if it exceeds the threshold range or meets the association rules, the system will give feedback and issue a warning. If the above rules are not met, the system will not give feedback and issue a warning.

5. The method for constructing an automated early warning network for monitoring water quality at an estuary according to claim 4, characterized in that: The key factors are determined by using a multi-input single-output feedforward neural network, dividing it into several levels, including an input layer, a hidden layer and an output layer, and measuring the prediction value of the model through a loss function when training the neural network.

6. The method for constructing an automated early warning network for monitoring water quality at an estuary according to claim 5, characterized in that: The basic structure of the key factor is as follows: Input Layer: X=[X1,X2,......X n ] T Where n is the number of input water quality parameters. Hidden Layer: Z l =W l A l-1 +b l A l =g l (Z l ) Among them, l means there are l hidden layers, W l is the weight matrix of the lth layer, b l is the bias vector of the lth layer, A l-1 is the activation output of the previous layer, g l () is the activation function of the lth layer, commonly used ones are ReLU and sigmoid. Output layer: Among them, W l+1 and b l+1 are the weight matrix and bias vector of the output layer, A l is the activation output of the last hidden layer. Loss function: Where N is the number of samples.

7. The method for constructing an automated early warning network for monitoring water quality at an estuary according to claim 1, characterized in that: The model application automatically accesses the database interface of the real-time incoming data through the system and obtains the data. After processing and analyzing the data, the specific steps of determining whether the system triggers the early warning mechanism according to the analysis results are as follows: The system automatically accesses the database interface of the real-time incoming data and obtains the data, and performs processing and analysis; After processing and analyzing the data, when the monitoring data reaches a preset threshold or the analysis result indicates a potential risk, the system will automatically trigger an early warning mechanism and determine whether to send an early warning message based on the final calculation result; If it is determined that the warning message needs to be sent, relevant managers and decision makers will be notified via SMS, email or mobile application to achieve warning.

Images