Water quality monitoring method and system for surface water

Abstract

The invention discloses a water quality monitoring method and system for surface water, and relates to the technical field of water quality monitoring. Obtaining water body state data of the target area, and performing feature extraction on the spectral data through a preset algorithm to obtain training data; inputting the training data into a water body evaluation model for evaluation to obtain water body abnormal parameters, obtaining ultrasonic detection data, and determining water body turbidity parameters of the target area according to the ultrasonic detection data; combining the water body abnormal parameter and the water body turbid parameter to obtain a water quality evaluation parameter, and evaluating the water quality of the target area according to the water quality evaluation parameter; the preset algorithm performs feature extraction on the spectral data to generate training data, so that the model training efficiency and accuracy are improved. Training data are input into the water body evaluation model, water body abnormal parameters are automatically evaluated, water quality evaluation parameters are obtained in combination with turbidity parameters detected through ultrasonic waves, comprehensive water quality evaluation is conducted, an alarm is automatically given when a threshold value is exceeded, and the monitoring efficiency and the emergency response speed are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of water quality monitoring, and in particular relates to a method and system for monitoring surface water quality. Background Art

[0002] Surface waters are often affected by the discharge of untreated wastewater, which leads to eutrophication, excessive algal blooms and reduced water clarity. Given this range of negative impacts, continuous monitoring of water quality in urban reservoirs is an essential step in inland water management systems.

[0003] Patent No.: CN118225713A, discloses a water quality monitoring method and system based on full spectrum analysis; the first machine learning model is used to predict the influence of environmental factors on the spectrum, and the second machine learning model is used to predict water quality parameters; during the monitoring process, the spectral data and environmental data of the water body are collected in real time, and the first model is used to predict the influence of the environment on the spectrum, and the real-time spectral data is corrected; the corrected spectral data is then optimized and input into the second model to predict the water quality index value; finally, the predicted water quality index value is compared with the preset threshold, and an alarm message is sent if the threshold is exceeded. By collecting reference data to train the machine learning model, the influence of the environment on the spectral data is predicted and corrected, the accuracy of the spectral data is improved, and then the accuracy of water quality monitoring is improved.

[0004] Although the above method solves some problems, there are still some problems. For example, the process of water quality data analysis is too simple. Analysis only from spectral data leads to reduced accuracy of water quality data monitoring and increased probability of false alarms and missed reports. Summary of the invention

[0005] The purpose of the present invention is to solve the problem that the process of water quality data analysis is too simple, and only analyzing from spectral data leads to reduced accuracy of water quality data monitoring and increased probability of false alarms and missed alarms, and to propose a water quality monitoring method and system for surface water.

[0006] In a first aspect of the present invention, a method for monitoring surface water quality is first proposed, the method comprising: Acquire water body status data of the target area, and extract features from the spectral data using a preset algorithm to obtain training data; the water body status data includes: spectral data and water sample data; The training data is input into a water body assessment model to evaluate and obtain water body abnormality parameters, ultrasonic detection data is obtained, and water turbidity parameters of the target area are determined according to the ultrasonic detection data; The water body abnormality parameter and the water body turbidity parameter are combined to obtain a water quality assessment parameter, and the water quality of the target area is assessed according to the water quality assessment parameter.

[0007] Optionally, extracting features from the spectral data using a preset algorithm to obtain training data includes: Determine the data type of the water sample data, and divide each data type in the spectral data according to a preset division standard to obtain a classification data set; the classification data set includes multiple classification data sets; Oversampling the classified data set in the classified data set to obtain a synthetic data set, and mathematically transforming the spectral data through a preset operation rule to obtain a transformed data set; Determining the data type of the converted data set, and fusing the converted data with the synthesized data set according to the data type to obtain a fused data set; The band features in the fused data set are globally filtered by a genetic algorithm to obtain initial band features, and the initial band features are selected by a variance expansion factor to obtain optimal band features, and the optimal band features are used as training data.

[0008] Optionally, before inputting the training data into a water body assessment model for assessment to obtain water body abnormality parameters, the method further includes: Step 1: Divide the training data into a preset number to obtain a training data set, input the training data set into a first preset model, perform training according to a preset number of training times to obtain a result set, and vertically merge the result set to obtain a column vector ; Step 2: Repeat the steps until the number of column vectors is the same as the preset number to obtain the column vector , input the training data into the second preset model to obtain a column vector ; The column vector With the column vector have the same number of column vectors; Step 3: Obtain water quality assessment parameters, and combine the water quality assessment parameters, the training data and the column vector Inputting the input into the third preset model for training to obtain model parameters, and updating the third preset model according to the model parameters to obtain a water body assessment model; The training data is input into the water body assessment model for assessment to obtain water body abnormal parameters, including: The column vector The water body abnormal parameters are input into the water body assessment model for assessment.

[0009] Optionally, determining the water turbidity parameter of the target area according to the ultrasonic detection data includes: Determine the water flow velocity, the transmitting signal and the receiving signal according to the ultrasonic detection data, calculate the water flow rate according to the water flow velocity, and calculate the signal attenuation according to the transmitting signal and the receiving signal; The signal attenuation is input into a particle assessment model for training to obtain particle concentration and water particle size; the particle assessment model is trained by historical signal attenuation data, particle concentration data and water particle size data; The water turbidity parameter is calculated according to the water flow, the particle concentration and the water particle size.

[0010] Optionally, evaluating the water quality of the target area according to the water quality evaluation parameters includes: If the water quality assessment parameter is greater than the alarm threshold, the log is uploaded and an alarm instruction is sent to cause the system to alarm; otherwise, the log is uploaded.

[0011] In a second aspect of the present invention, a water quality monitoring system for surface water is proposed, comprising: a training data acquisition module, a water body parameter acquisition module and a water quality assessment and early warning module: The training data acquisition module is used to acquire water state data of the target area, and extract features from spectral data using a preset algorithm to obtain training data; the water state data includes: spectral data and water sample data; The water body parameter acquisition module is used to input the training data into the water body assessment model to evaluate and obtain water body abnormal parameters, obtain ultrasonic detection data, and determine the water body turbidity parameters of the target area according to the ultrasonic detection data; The water quality assessment and early warning module is used to combine the water body abnormality parameter and the water body turbidity parameter to obtain a water quality assessment parameter, and to assess the water quality of the target area according to the water quality assessment parameter.

[0012] Optionally, the training data acquisition module includes: a data set division module, a data set preprocessing module, a data set fusion module and a band screening module: The data set division module is used to determine the data type of the water sample data, and divide each data type in the spectral data according to a preset division standard to obtain a classification data set; the classification data set includes multiple classification data sets; The data set preprocessing module is used to oversample the classified data set in the classified data set to obtain a synthetic data set, and to mathematically transform the spectral data through a preset operation rule to obtain a transformed data set; The data set fusion module is used to determine the data type of the converted data set, and fuse the converted data with the synthetic data set according to the data type to obtain a fused data set; The band screening module is used to perform global filtering on the band features in the fused data set through a genetic algorithm to obtain initial band features, perform band feature selection on the initial band features through a variance expansion factor to obtain optimal band features, and use the optimal band features as training data.

[0013] Optionally, the system further includes: a column vector acquisition module, a column vector pre-training module and an evaluation model training module: The column vector acquisition module is used to divide the training data according to a preset number to obtain a training data set, input the training data set into the first preset model to perform training according to a preset number of training times to obtain a result set, and vertically merge the result set to obtain a column vector ; The column vector pre-training module is used to repeatedly execute the column vector acquisition module until the number of column vectors is the same as the preset number to obtain the column vector , input the training data into the second preset model to obtain a column vector ; The column vector With the column vector have the same number of column vectors; The assessment model training module is used to obtain water quality assessment parameters, and to combine the water quality assessment parameters, the training data and the column vector Inputting the input into the third preset model for training to obtain model parameters, and updating the third preset model according to the model parameters to obtain a water body assessment model; The water body parameter acquisition module is also used to convert the column vector The water body abnormal parameters are input into the water body assessment model for assessment.

[0014] Optionally, the water body parameter acquisition module includes: a signal attenuation acquisition module, a particle evaluation model module and a water body turbidity parameter module: The signal attenuation acquisition module is used to determine the water flow velocity, the transmission signal and the reception signal according to the ultrasonic detection data, calculate the water flow rate according to the water flow velocity, and calculate the signal attenuation according to the transmission signal and the reception signal; The particle assessment model module is used to input the signal attenuation into the particle assessment model for training to obtain particle concentration and water particle size; the particle assessment model is trained by historical signal attenuation data, particle concentration data and water particle size data; The water turbidity parameter module is used to calculate the water turbidity parameter according to the water flow, the particle concentration and the water particle size.

[0015] Optionally, the water quality assessment and warning module is also used to upload the log and send an alarm instruction to cause the system to alarm if the water quality assessment parameter is greater than the alarm threshold; otherwise, upload the log.

[0016] Beneficial effects of the present invention: The present invention proposes a water quality monitoring method for surface water, which obtains water state data of a target area, extracts features from spectral data through a preset algorithm to obtain training data; inputs the training data into a water body assessment model to obtain abnormal water body parameters, obtains ultrasonic detection data, and determines water turbidity parameters of the target area according to the ultrasonic detection data; combines abnormal water body parameters with water turbidity parameters to obtain water quality assessment parameters, and evaluates the water quality of the target area according to the water quality assessment parameters; extracts features from spectral data through a preset algorithm to generate training data, thereby improving model training efficiency and accuracy. Input the training data into a water body assessment model, automatically assess abnormal water body parameters, and obtain water quality assessment parameters in combination with turbidity parameters detected by ultrasonic waves, conduct a comprehensive water quality assessment, and automatically alarm when the threshold is exceeded, thereby improving monitoring efficiency and emergency response speed. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The present invention will be further described below in conjunction with the accompanying drawings.

[0018] Figure 1 A flow chart of a method for monitoring surface water quality is provided for an embodiment of the present invention; Figure 2 Another framework diagram of a water quality monitoring system for surface water is provided for an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. The term "and / or" herein is only a description of the association relationship of the associated objects, indicating that there can be three relationships. For example, A and B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, in the present invention, the description of "first", "second", etc. is only used for descriptive purposes, and cannot be understood as indicating or implying its relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" can 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 it must be based on the ability of ordinary technicians in the field to implement. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0020] Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative work shall fall within the scope of protection of the present invention.

[0021] The embodiment of the present invention provides a method for monitoring the water quality of surface water. Figure 1 , Figure 1 A flow chart of a method for monitoring surface water quality provided by an embodiment of the present invention. The method comprises the following steps: S101, obtaining water state data of a target area, and extracting features from spectral data using a preset algorithm to obtain training data; S102, inputting the training data into the water body assessment model to evaluate and obtain water body abnormality parameters, obtaining ultrasonic detection data, and determining water body turbidity parameters of the target area according to the ultrasonic detection data; S103, combining the water body abnormality parameter and the water body turbidity parameter to obtain a water quality assessment parameter, and assessing the water quality of the target area according to the water quality assessment parameter; Water status data include: spectral data and water sample data; Based on a water quality monitoring method for surface water provided by an embodiment of the present invention, a preset algorithm is used to extract features from spectral data to generate training data, thereby improving the efficiency and accuracy of model training. The training data is input into a water body assessment model, and abnormal water body parameters are automatically assessed. Water quality assessment parameters are obtained by combining turbidity parameters detected by ultrasonic waves, and a comprehensive water quality assessment is performed. The system monitors water quality in real time and automatically alarms when the threshold is exceeded, thereby improving monitoring efficiency and emergency response speed.

[0022] In one implementation, water state data includes: spectral data and water sample data, wherein the spectral data is obtained through a detection device installed at the water inlet, and the cross-section at the water inlet is determined by measurement by staff. The spectral data is obtained, for example, by a multi-parameter water quality monitoring spectrometer analyzer, a water quality spectral detection system, etc.; the water sample data is obtained, for example, by an ORP sensor, a pH sensor, a dissolved oxygen sensor, a water temperature sensor, etc.; wherein, the water sample data is mapped to the spectral data; and the spectral data is feature extracted by a preset algorithm (genetic algorithm) to obtain training data preset algorithm.

[0023] In one implementation, the data monitoring accuracy can be improved by using water body status data and ultrasonic detection data. Ultrasonic wave: has high measurement accuracy for the physical properties of the fluid (such as flow rate, suspended particles), but has weak detection capability for chemical components. Spectroscopy: has high detection accuracy for chemical components, but has weak measurement capability for physical parameters. The combination of the two: by measuring physical and chemical parameters at the same time, the water quality can be evaluated more comprehensively, and the accuracy and reliability of monitoring can be improved; and, ultrasonic wave: can measure the flow rate, flow rate, suspended solids concentration, etc. of the fluid. Spectroscopy: can measure chemical oxygen demand (COD), total organic carbon (TOC), dissolved oxygen (DO), turbidity, color, etc. The combination of the two: can monitor multiple parameters at the same time, provide more comprehensive water quality information, and achieve more comprehensive, accurate and reliable water quality monitoring.

[0024] In one implementation, the training data after feature extraction is input into the water body assessment model, and the model can assess the abnormal parameters of the water body based on these features. By using machine learning or deep learning algorithms, the mapping relationship between water body status and abnormal parameters is learned through existing training data, and the automatic identification and quantitative assessment of water body abnormalities are realized. Compared with traditional manual detection methods, this method can quickly and accurately identify abnormal conditions in water bodies, such as excessive pollutant concentrations and algae outbreaks.

[0025] In one implementation, ultrasonic detection data is obtained, and water turbidity parameters of the target area are determined based on these data. Ultrasonic detection technology can directly measure the concentration and particle size distribution of suspended particles in water, thereby accurately reflecting the turbidity of the water. The water turbidity parameters complement the water abnormality parameters evaluated by spectral data, providing more comprehensive water quality information. The independent measurement of turbidity parameters ensures the diversity and reliability of the data, helps to more accurately evaluate the overall state of the water, and can also reflect the flow of the water.

[0026] In one implementation, the water body abnormality parameters and water body turbidity parameters are combined to obtain comprehensive water quality assessment parameters. The above comprehensive assessment method takes into account the chemical and physical properties of the water body. Through the comprehensive analysis of multiple parameters, the water quality of the target area can be more comprehensively and accurately assessed. The comprehensive assessment parameters not only reflect the degree of pollution of the water body, but can also be used as a judgment method for the alarm system. When the water quality assessment parameters exceed the preset alarm threshold, the log is uploaded in time and the alarm command is sent.

[0027] In one embodiment, the training data is obtained by extracting features from the spectral data using a preset algorithm, including: Determine the data type of the water sample data, and divide each data type in the spectral data according to a preset division standard to obtain a classification data set; the classification data set includes multiple classification data sets; Oversampling the classified data set in the classified data set to obtain a synthetic data set, and mathematically transforming the spectral data through a preset operation rule to obtain a transformed data set; Determine the data type of the converted data set, and fuse the converted data with the synthetic data set according to the data type to obtain a fused data set; The genetic algorithm is used to globally filter the band features in the fused data set to obtain the initial band features, and the variance expansion factor is used to select the initial band features to obtain the optimal band features, which are used as training data.

[0028] In one implementation, the spectral data is mathematically transformed by a preset operation rule, for example: mathematical transformation using the inverse, logarithm, exponent, square and square root; according to the data type, the transformed data is fused with the synthetic data set to obtain a fused data set, adding, subtracting, multiplying, distributing and standardizing, etc.; the classified data set in the classified data set is oversampled to obtain a synthetic data set. Oversampling technology can increase the diversity and quantity of data, especially in the case of data imbalance, it can effectively increase the amount of data in the minority category and make the model training more balanced. The generation of synthetic data sets not only enriches the data set, but also improves the generalization ability of the model, enabling it to better learn and identify data features of different categories.

[0029] In one implementation, the data type of the transformed data set is determined, and the transformed data is fused with the synthetic data set according to the data type to obtain a fused data set. The data fusion process integrates data from different sources and processing methods through operations such as addition, subtraction, multiplication, distribution, and standardization to generate comprehensive features. This process not only retains the rich information of the original data, but also generates new features through the interaction between data, further improving the expressiveness of the data and the predictive performance of the model.

[0030] In one implementation, the spectral data is mathematically transformed using preset operation rules, such as using reciprocals, logarithms, exponents, squares, and square roots. These mathematical transformations can enhance the characteristic performance of the data, make the distribution of the data more reasonable, and improve the sensitivity of the model to the data; determine the data type of the water sample data, and divide the data types in the spectral data according to the preset division criteria to obtain a classified data set. The above steps can classify complex spectral data according to different characteristics or attributes, making the data more organized and easy to manage. Through classification, different types of data can be processed and analyzed in a targeted manner to improve the efficiency and accuracy of data processing.

[0031] In one implementation, the genetic algorithm is used to globally filter the band features in the fused data set to obtain the initial band features. As a global optimization algorithm, the genetic algorithm can effectively search and filter out the most representative and discriminative band features, avoiding the problem of local optimal solutions. The variance inflation factor is used to select the band features of the initial band features to obtain the optimal band features. The variance inflation factor can further filter out features with low correlation and large information content, reduce redundancy between features, and improve the stability and prediction accuracy of the model. The optimal band features finally obtained are used as training data, providing high-quality input for model training, ensuring that the model can accurately learn and identify changes in water status.

[0032] In one embodiment, before inputting the training data into the water body assessment model to assess and obtain water body abnormality parameters, the method further includes: Step 1: Divide the training data into a preset number to obtain a training data set, input the training data set into the first preset model, train it according to the preset number of training times to obtain a result set, and vertically merge the result set to obtain a column vector ; Step 2: Repeat the steps until the number of column vectors is the same as the preset number to obtain the column vector , input the training data into the second preset model to obtain the column vector ; Column vector With column vector have the same number of column vectors; Step 3: Obtain water quality assessment parameters, and transform water quality assessment parameters, training data and column vector The input is input into the third preset model for training to obtain model parameters, and the third preset model is updated according to the model parameters to obtain a water body assessment model; The training data is input into the water body assessment model to evaluate and obtain water body abnormal parameters, including: The column vector The water body abnormal parameters are input into the water body assessment model for assessment.

[0033] In one implementation, for example, the training data is divided into five training data sets (four for training and one for validation), the preset number of training times has the same mathematical meaning as the preset number, which is five times here, and the result set is obtained by training according to the preset number of training times ( and ), the result set is merged vertically into one column to get a column vector Repeat the steps until the number of column vectors is the same as the preset number to obtain column vector , where the number of column vectors is five as well as the preset number, and As a column vector ( , i value, i = 1, 2, 3...n), input the training data into the second preset model to obtain the column vector , column vector With column vector The number of columns is the same as that of water quality assessment parameters, training data (excellent band features) and column vector Input into the third model for training to obtain the water body assessment model, training data (excellent band characteristics) and column vector Combined, the water quality assessment parameters are the scores of various chemical components in the water body at various concentrations (scoring vectors, used to characterize the pollution state of the water body). That is, the water body contains a variety of pollutants, each pollutant is set with a pollutant concentration range, and each concentration range corresponds to a score, such as: ammonia nitrogen (NH3-N), total phosphorus (TP), chemical oxygen demand (COD) and dissolved oxygen (DO). The "concentration range (mg / L)-score" of ammonia nitrogen: 0≤concentration<0.5 corresponds to 10 points, 0.5≤concentration<1.0 corresponds to 7 points, 1.0≤concentration<1.5 corresponds to 4 points, and 1.5≤concentration corresponds to 1 point; the "concentration range (mg / L)-score" of total phosphorus: 0≤concentration<0.1 corresponds to 10 points, 0.1≤concentration<0.2 corresponds to 7 points, 0.2≤concentration<0.3 corresponds to 4 points, and 0.3≤concentration corresponds to 1 point; the "concentration range (mg / L)-score" of chemical oxygen demand: 0≤concentration<0.1 corresponds to 10 points, 0.1≤concentration<0.2 corresponds to 7 points, 0.2≤concentration<0.3 corresponds to 4 points, and 0.3≤concentration corresponds to 1 point. L)-score": 0≤concentration<10 corresponds to 10 points, 10≤concentration<20 corresponds to 7 points, 20≤concentration<30 corresponds to 4 points, 30≤concentration corresponds to 1 point; "Concentration range (mg / L)-score" of dissolved oxygen: 6.0≤concentration corresponds to 10 points, 5.0≤concentration<6.0 corresponds to 7 points, 4.0≤concentration<5.0 corresponds to 4 points, 4.0<concentration corresponds to 1 point; When the test results are as follows, the ammonia nitrogen concentration is 1.2, the score is 4, the total phosphorus concentration is 0.5, the score is 1, the chemical oxygen demand concentration is 31, the score is 1, and the dissolved oxygen concentration is 5.3. The score vector of the pollutant score of the water body is: [4,1,1,7]. The lower the value of the score vector, the more serious the water pollution. It can be seen from the score vector that the scores of various pollutants in the water body are 4, indicating that the water body as a whole is at a medium pollution level and needs further treatment to improve the water quality. The above data is for reference only; the column vector Input into the water body assessment model to obtain water body anomaly parameters, column vector Containing a variety of chemical components, the water assessment model scores the chemical components according to their concentrations and assigns weights to each score, and the weighted sum is used to obtain the water anomaly parameters.

[0034] In one implementation, the first preset model is a basic model, such as: decision tree, support vector machine (SVM), k-nearest neighbor (k-NN), logistic regression; the second preset model is an integrated model, such as: random forest, gradient boosting tree (Gradient Boosting Trees, GBT), XGBoost, LightGBM; the third preset model is a meta-model, such as: linear regression, ridge regression, LASSO regression (Least Absolute Shrinkage and Selection Operator), neural network.

[0035] In one embodiment, determining the water turbidity parameter of the target area according to the ultrasonic detection data includes: Determine the water flow velocity, the transmitting signal and the receiving signal according to the ultrasonic detection data, calculate the water flow rate according to the water flow velocity, and calculate the signal attenuation according to the transmitting signal and the receiving signal; The signal attenuation is input into the particle assessment model for training to obtain the particle concentration and water particle size; the particle assessment model is trained by the historical signal attenuation data, particle concentration data and water particle size data; The water turbidity parameters are calculated based on water flow, particle concentration and water particle size.

[0036] In one implementation, the water flow velocity is determined based on ultrasonic detection data, for example, the time difference method and the Doppler effect are used to measure the flow velocity and flow rate of the fluid. The time difference method: the flow velocity of the fluid is calculated by measuring the time difference between the propagation of ultrasonic waves in the fluid downstream and upstream. Assume that the sound speed in the stationary fluid is c, the fluid flow speed is v, the distance between the transducers is L, and the installation angle is θ; the downstream propagation time for: , countercurrent propagation time for: , the time difference Δt is: ,because , so it can be simplified to , the flow velocity v of the fluid can be calculated by measuring the time difference Δt; Doppler effect: the Doppler frequency shift of sound waves is used to measure the movement speed of particles or bubbles in the fluid. The ultrasonic generator emits sound waves of a fixed frequency. When the particles or bubbles in the fluid move at a speed v, the frequency of the reflected sound waves will change. Assume that the transmission frequency is f0, the reflection frequency is f1, and the Doppler frequency shift Δf is: By measuring the Doppler frequency shift Δf, the flow velocity v of the fluid can be calculated; water flow rate, that is, water flow rate = cross-sectional area × water flow velocity.

[0037] In one implementation, the method for determining the particle concentration and the particle size of the water body also includes: Acoustic backscatter signal analysis: By analyzing the intensity and frequency change of the reflected signal, the concentration and particle size of the suspended particles can be determined. The intensity of the reflected signal is proportional to the concentration of the suspended particles, and the frequency change of the reflected signal is proportional to the movement speed of the particles. Doppler shift measurement: By measuring the Doppler frequency shift of the reflected signal, the movement speed of the suspended particles can be determined, thereby calculating the concentration of the suspended particles. The Doppler frequency shift is proportional to the movement speed of the particles, and the frequency shift can be converted into concentration through a calibration curve.

[0038] In one implementation, an ultrasonic transmitter and a receiver are set up to allow ultrasonic waves to pass through a body of water containing suspended particles. The ultrasonic wave intensity emitted by the transmitter and the reflected signal intensity received by the receiver are measured. By comparing these two intensity values, the energy loss of the ultrasonic wave during propagation can be calculated, and then the signal attenuation can be obtained. A relationship model between signal attenuation and particle concentration is established. This model is obtained through experimental measurements, that is, the signal attenuation of ultrasonic waves is measured in a body of water with a known particle concentration, and it is used as training data for a particle assessment model (which can be a machine learning model, a neural network model, etc.), thereby establishing a mapping relationship between signal attenuation and particle concentration and particle size; the water turbidity parameter is calculated based on the water flow, particle concentration and water particle size, that is, ,in, is the water turbidity parameter, gc is the particle concentration, wp is the water particle size, wf is the water flow rate, α, β and γ are constant proportional coefficients, all of which are not zero; the values ​​of gc, wp and wf are dimensionless, and the larger the value, the more turbid the water.

[0039] In one embodiment, the water quality of the target area is evaluated according to the water quality evaluation parameters, including: If the water quality assessment parameter is greater than the alarm threshold, the log is uploaded and an alarm instruction is sent to cause the system to alarm; otherwise, the log is uploaded.

[0040] In one implementation, before evaluating the water quality of the target area according to the water quality evaluation parameter, the water quality evaluation parameter is obtained by combining the water body abnormality parameter and the water body turbidity parameter, that is, ,in is a water quality assessment parameter, is the water body abnormal parameter, is the water turbidity parameter, and are all constant proportional coefficients and are not zero.

[0041] In one implementation, upload logs: the collected data is transmitted in real time to the database and cloud server of the monitoring center via a wireless network (or wired transmission) to ensure the timeliness and accuracy of the data. The system automatically records detailed information of the alarm event, including time, location, alarm parameters, alarm level, etc., and generates an alarm log. Alarm threshold: Set the alarm threshold. When the water quality assessment parameter exceeds (greater than and not equal to) the threshold, the system automatically triggers the alarm. For example: turbidity exceeds 100 NTU, COD exceeds 20 mg / L, etc. Alarm method: Alarm methods include SMS, APP push, email and other forms to ensure that relevant personnel can receive alarm information in a timely manner.

[0042] Based on the same inventive concept, the present invention also provides a water quality monitoring system for surface water. Figure 2 , Figure 2 A schematic diagram of the structure of a water quality monitoring system for surface water provided by an embodiment of the present invention includes: a training data acquisition module, a water body parameter acquisition module and a water quality assessment and early warning module: The training data acquisition module is used to obtain the water state data of the target area, and extract the features of the spectral data through a preset algorithm to obtain the training data; the water state data includes: spectral data and water sample data; The water body parameter acquisition module is used to input the training data into the water body assessment model to evaluate and obtain water body abnormal parameters, obtain ultrasonic detection data, and determine the water turbidity parameters of the target area based on the ultrasonic detection data; The water quality assessment and early warning module is used to combine water body abnormal parameters and water body turbidity parameters to obtain water quality assessment parameters, and evaluate the water quality of the target area based on the water quality assessment parameters.

[0043] Based on a water quality monitoring system for surface water provided by an embodiment of the present invention, a preset algorithm is used to extract features from spectral data to generate training data, thereby improving the efficiency and accuracy of model training. The training data is input into a water body assessment model, and abnormal water body parameters are automatically assessed. Water quality assessment parameters are obtained by combining turbidity parameters detected by ultrasonic waves, and a comprehensive water quality assessment is performed. The system monitors water quality in real time and automatically issues an alarm when the threshold is exceeded, thereby improving monitoring efficiency and emergency response speed.

[0044] In one embodiment, the training data acquisition module includes: a data set division module, a data set preprocessing module, a data set fusion module and a band screening module: A data set division module is used to determine the data type of the water sample data, and divide each data type in the spectral data according to a preset division standard to obtain a classification data set; the classification data set includes multiple classification data sets; A data set preprocessing module is used to oversample the classified data set in the classified data set to obtain a synthetic data set, and to mathematically transform the spectral data through a preset operation rule to obtain a transformed data set; A data set fusion module is used to determine the data type of the converted data set, and fuse the converted data with the synthetic data set according to the data type to obtain a fused data set; The band screening module is used to perform global filtering on the band features in the fusion data set through a genetic algorithm to obtain the initial band features, perform band feature selection on the initial band features through a variance expansion factor to obtain the optimal band features, and use the optimal band features as training data.

[0045] In one embodiment, the system further includes: a column vector acquisition module, a column vector pre-training module and an evaluation model training module: The column vector acquisition module is used to divide the training data into a preset number to obtain a training data set, input the training data set into the first preset model to perform training according to a preset number of training times to obtain a result set, and vertically merge the result set to obtain a column vector ; The column vector pre-training module is used to repeatedly execute the column vector acquisition module until the number of column vectors is the same as the preset number to obtain the column vector , input the training data into the second preset model to obtain the column vector ; Column vector With column vector have the same number of column vectors; The evaluation model training module is used to obtain water quality evaluation parameters, and to convert water quality evaluation parameters, training data and column vectors The input is input into the third preset model for training to obtain model parameters, and the third preset model is updated according to the model parameters to obtain a water body assessment model; The water body parameter acquisition module is also used to convert the column vector The water body abnormal parameters are input into the water body assessment model for assessment.

[0046] In one embodiment, the water body parameter acquisition module includes: a signal attenuation acquisition module, a particle evaluation model module and a water body turbidity parameter module: A signal attenuation acquisition module is used to determine the water flow velocity, the transmitting signal and the receiving signal according to the ultrasonic detection data, calculate the water flow rate according to the water flow velocity, and calculate the signal attenuation according to the transmitting signal and the receiving signal; The particle assessment model module is used to input the signal attenuation into the particle assessment model for training to obtain the particle concentration and water particle size; the particle assessment model is trained by the historical signal attenuation data, particle concentration data and water particle size data; The water turbidity parameter module is used to calculate the water turbidity parameters based on water flow, particle concentration and water particle size.

[0047] In one embodiment, the water quality assessment warning module is also used to upload the log and send an alarm instruction to cause the system to alarm if the water quality assessment parameter is greater than the alarm threshold; otherwise, upload the log.

[0048] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A method for monitoring surface water quality, characterized in that: The method comprises: Acquire water body status data of the target area, and extract features from the spectral data using a preset algorithm to obtain training data; the water body status data includes: spectral data and water sample data; The training data is input into a water body assessment model to evaluate and obtain water body abnormality parameters, ultrasonic detection data is obtained, and water turbidity parameters of the target area are determined according to the ultrasonic detection data; The water body abnormality parameter and the water body turbidity parameter are combined to obtain a water quality assessment parameter, and the water quality of the target area is assessed according to the water quality assessment parameter.

2. A method for monitoring surface water quality according to claim 1, characterized in that: Extracting features from the spectral data using a preset algorithm to obtain training data includes: Determine the data type of the water sample data, and divide each data type in the spectral data according to a preset division standard to obtain a classification data set; the classification data set includes multiple classification data sets; Oversampling the classified data set in the classified data set to obtain a synthetic data set, and mathematically transforming the spectral data through a preset operation rule to obtain a transformed data set; Determining the data type of the converted data set, and fusing the converted data with the synthesized data set according to the data type to obtain a fused data set; The band features in the fused data set are globally filtered by a genetic algorithm to obtain initial band features, and the initial band features are selected by a variance expansion factor to obtain optimal band features, and the optimal band features are used as training data.

3. A method for monitoring surface water quality according to claim 1, characterized in that: Before inputting the training data into the water body assessment model to assess and obtain water body abnormality parameters, the method further includes: Step 1: Divide the training data into a preset number to obtain a training data set, input the training data set into a first preset model, perform training according to a preset number of training times to obtain a result set, and vertically merge the result set to obtain a column vector ; Step 2: Repeat the steps until the number of column vectors is the same as the preset number to obtain the column vector , input the training data into the second preset model to obtain a column vector ; The column vector With the column vector have the same number of column vectors; Step 3: Obtain water quality assessment parameters, and combine the water quality assessment parameters, the training data and the column vector Inputting the input into the third preset model for training to obtain model parameters, and updating the third preset model according to the model parameters to obtain a water body assessment model; The training data is input into the water body assessment model for assessment to obtain water body abnormal parameters, including: The column vector The water body abnormal parameters are input into the water body assessment model for assessment.

4. A method for monitoring surface water quality according to claim 1, characterized in that: Determining water turbidity parameters of a target area according to the ultrasonic detection data includes: Determine the water flow velocity, the transmitting signal and the receiving signal according to the ultrasonic detection data, calculate the water flow rate according to the water flow velocity, and calculate the signal attenuation according to the transmitting signal and the receiving signal; The signal attenuation is input into a particle assessment model for training to obtain particle concentration and water particle size; the particle assessment model is trained by historical signal attenuation data, particle concentration data and water particle size data; The water turbidity parameter is calculated according to the water flow, the particle concentration and the water particle size.

5. A method for monitoring surface water quality according to claim 1, characterized in that: The water quality of the target area is assessed according to the water quality assessment parameters, including: If the water quality assessment parameter is greater than the alarm threshold, the log is uploaded and an alarm instruction is sent to cause the system to alarm; otherwise, the log is uploaded.

6. A water quality monitoring system for surface water, characterized in that: The system includes: a training data acquisition module, a water body parameter acquisition module and a water quality assessment and early warning module: The training data acquisition module is used to acquire water state data of the target area, and obtain training data by extracting features from spectral data using a preset algorithm; the water state data includes: spectral data and water sample data; The water body parameter acquisition module is used to input the training data into the water body assessment model to evaluate and obtain water body abnormal parameters, obtain ultrasonic detection data, and determine the water body turbidity parameters of the target area according to the ultrasonic detection data; The water quality assessment and early warning module is used to combine the water body abnormality parameter and the water body turbidity parameter to obtain a water quality assessment parameter, and to assess the water quality of the target area according to the water quality assessment parameter.

7. A water quality monitoring system for surface water according to claim 6, characterized in that: The training data acquisition module includes: a data set division module, a data set preprocessing module, a data set fusion module and a band screening module: The data set division module is used to determine the data type of the water sample data, and divide each data type in the spectral data according to a preset division standard to obtain a classification data set; the classification data set includes multiple classification data sets; The data set preprocessing module is used to oversample the classified data set in the classified data set to obtain a synthetic data set, and to mathematically transform the spectral data through a preset operation rule to obtain a transformed data set; The data set fusion module is used to determine the data type of the converted data set, and fuse the converted data with the synthetic data set according to the data type to obtain a fused data set; The band screening module is used to perform global filtering on the band features in the fused data set through a genetic algorithm to obtain initial band features, perform band feature selection on the initial band features through a variance expansion factor to obtain optimal band features, and use the optimal band features as training data.

8. A water quality monitoring system for surface water according to claim 6, characterized in that: The system further comprises: a column vector acquisition module, a column vector pre-training module and an evaluation model training module: The column vector acquisition module is used to divide the training data according to a preset number to obtain a training data set, input the training data set into the first preset model to perform training according to a preset number of training times to obtain a result set, and vertically merge the result set to obtain a column vector ; The column vector pre-training module is used to repeatedly execute the column vector acquisition module until the number of column vectors is the same as the preset number to obtain the column vector , input the training data into the second preset model to obtain a column vector ; The column vector With the column vector have the same number of column vectors; The assessment model training module is used to obtain water quality assessment parameters, and to combine the water quality assessment parameters, the training data and the column vector Inputting the input into the third preset model for training to obtain model parameters, and updating the third preset model according to the model parameters to obtain a water body assessment model; The water body parameter acquisition module is also used to convert the column vector The water body abnormal parameters are input into the water body assessment model for assessment.

9. A surface water quality monitoring system according to claim 6, characterized in that: The water body parameter acquisition module includes: a signal attenuation acquisition module, a particle evaluation model module and a water body turbidity parameter module: The signal attenuation acquisition module is used to determine the water flow velocity, the transmission signal and the reception signal according to the ultrasonic detection data, calculate the water flow rate according to the water flow velocity, and calculate the signal attenuation according to the transmission signal and the reception signal; The particle assessment model module is used to input the signal attenuation into the particle assessment model for training to obtain particle concentration and water particle size; the particle assessment model is trained by historical signal attenuation data, particle concentration data and water particle size data; The water turbidity parameter module is used to calculate the water turbidity parameter according to the water flow, the particle concentration and the water particle size.

10. A surface water quality monitoring system according to claim 6, characterized in that: The water quality assessment and early warning module is also used to upload the log and send an alarm instruction to cause the system to alarm if the water quality assessment parameter is greater than the alarm threshold; otherwise, upload the log.

Images