Automatic monitoring system and method for environmental pollutants
By deploying a variety of pollutant sensors and positioning modules in the target area, combining 5G with satellite communications, and conducting data preprocessing and in-depth analysis, the real-time and accuracy issues of environmental pollutant monitoring in existing technologies have been solved, and comprehensive and rapid monitoring of pollutants and emergency response have been achieved.
Patent Information
- Application Number
- CN202510874023.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-27
- Publication Date
- 2025-10-14
AI Technical Summary
Existing environmental pollutant monitoring technologies have problems such as poor real-time performance, limited monitoring range, insufficient data processing capabilities, and poor communication stability, making it difficult to achieve comprehensive and accurate monitoring and timely response to multiple pollutants.
Monitoring nodes equipped with various pollutant sensors and positioning modules are deployed in the target area, and data is transmitted by combining 5G and satellite communications. In-depth analysis is performed through data preprocessing, cluster analysis and regression models to generate pollution maps and emergency response plans.
It achieves comprehensive and real-time monitoring of environmental pollutants, improves the accuracy and availability of data, enables rapid response to pollution incidents, generates accurate pollution area divisions and concentration trend forecasts, and supports environmental protection decision-making.
Smart Images
Figure CN120779050A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of environmental pollution monitoring, in particular to an automatic monitoring system and method for environmental pollutants. BACKGROUND
[0002] With the rapid development of industrialization and urbanization, the types and emissions of environmental pollutants are increasing, which poses a serious threat to the ecological environment and human health. Accurate and timely monitoring of the distribution and trend of environmental pollutants is crucial for environmental protection and pollution control.
[0003] Traditional methods of monitoring environmental pollutants mainly rely on manual sampling and laboratory analysis, which has many limitations. On the one hand, the frequency of manual sampling is low, which cannot achieve real-time monitoring of environmental pollutants, and it is difficult to capture the instantaneous changes and dynamic processes of pollutants. On the other hand, the range of manual sampling is limited, which cannot fully cover large monitoring areas, and may lead to incomplete and inaccurate monitoring data. In addition, the laboratory analysis process is complex and time-consuming, which cannot meet the timeliness requirements of environmental pollution monitoring.
[0004] In recent years, although some automated monitoring equipment and technologies have emerged, these technologies often have problems such as single monitoring parameter, insufficient data processing capability, poor communication stability, etc. For example, some existing monitoring equipment can only monitor a few types of pollutants, which cannot meet the demand for comprehensive monitoring of multiple pollutants. In terms of data processing, there is a lack of effective data preprocessing and deep analysis methods, making it difficult to extract valuable information from a large amount of monitoring data. In the data transmission process, the communication method is single, and in complex environments, signal interruption or data loss may occur, affecting the reliability and stability of the monitoring system.
[0005] Therefore, an automatic monitoring system and method for environmental pollutants have become a problem that needs to be solved. SUMMARY
[0006] The technical problem to be solved by the present application is to provide an automatic monitoring method and system for environmental pollutants, which can realize efficient and accurate monitoring of environmental pollutants, in-depth data analysis and rapid emergency response.
[0007] To solve the above technical problems, the technical solution provided by the present application is as follows: an automatic monitoring method for environmental pollutants, comprising the following steps:
[0008] S1, in the target monitoring area, according to the regional terrain and pollution source distribution, a plurality of monitoring nodes are arranged, each monitoring node is provided with a plurality of pollutant sensors and a positioning module, the pollutant sensors collect real-time environmental pollutant data of different types, and the positioning module obtains the geographic position information of the monitoring node;
[0009] S2, the pollutant sensor of each monitoring node collects environmental pollutant data in real time according to the set sampling frequency, and transmits the collected data to the data processing unit; the data processing unit pre-processes the original data, including removing outliers and filling missing values;
[0010] S3, the pre-processed data is transmitted to the data aggregation center through the wireless communication module;
[0011] S4, after receiving the data transmitted by each monitoring node, the data aggregation center performs clustering analysis on the monitoring data by clustering algorithm, divides the monitoring area into different pollution areas; uses regression algorithm to establish the relationship model between pollutant concentration and environmental factors, and predicts the change trend of pollutant concentration; generates pollution map and monitoring report according to the analysis result;
[0012] S5, when the monitoring data exceeds the preset pollution threshold, the data aggregation center immediately sends an alarm signal, and notifies the relevant environmental protection departments and staff through SMS and email; at the same time, an emergency treatment scheme is generated according to the pollution situation.
[0013] Further, the environmental pollutant data includes particulate matter, sulfur dioxide and nitrogen oxides in the atmosphere, and chemical oxygen demand and ammonia nitrogen in water.
[0014] Further, the judgment of outliers adopts the 3σ principle based on statistics. If the data deviates from the mean value by more than 3 times the standard deviation, it is determined to be an outlier.
[0015] Further, the missing values are filled by using the average value of adjacent data, and the specific method is as follows:
[0016] For the sequence of environmental pollutant data collected in time sequence, the position of missing value is located, if the missing value is located at the beginning of the sequence, the average value of several data points adjacent to the right side is taken to fill; if the missing value is located at the end of the sequence, the average value of several data points adjacent to the left side is taken to fill; if the missing value is located in the middle of the sequence, the average value of several data points on the left and right sides is taken to fill.
[0017] Further, the wireless communication module adopts the combination of 5G communication and satellite communication, and 5G communication is preferentially used in areas with good 5G signal, and satellite communication is used in areas with weak 5G signal or remote areas.
[0018] Further, the specific method of clustering analysis of monitoring data by clustering algorithm and dividing the monitoring area into different pollution areas is as follows:
[0019] (1) For n monitoring nodes, each monitoring node has m characteristics related to pollution degree, and geographical position information additionally adds l characteristics; the monitoring data is represented as:
[0020]
[0021] where x ij denotes the jth eigenvalue of the ith monitoring node in matrix X, 1≤i≤n, 1≤j≤(m+l);
[0022] (2) Divide n data points into K clusters such that the intra-cluster SSE is minimized:
[0023]
[0024] where μ k is the centroid of the kth cluster C k , ||·|| denotes the Euclidean distance;
[0025] (3) For different K values, K = 1, 2, …, K max , calculate the corresponding SSE value; by observing the curve of SSE changing with K, find the turning point where the decreasing amplitude of SSE suddenly slows down, and the K value corresponding to this point is the appropriate number of clusters;
[0026] (4) Execute the clustering algorithm
[0027] Initialize the centroid: randomly select K data points as the initial centroid
[0028] Assign data points to clusters: for each data point x i , calculate its distance to each centroid and assign it to the nearest cluster, i.e. iff
[0029] Update the centroid: for each cluster update its centroid to where is the number of data points in cluster ;
[0030] Repeat the assignment of data points to clusters and the update of centroids until the centroids no longer change significantly or the maximum number of iterations is reached;
[0031] (5) Divide the pollution area
[0032] According to the final clustering result, the monitoring nodes belonging to the same cluster C K are divided into the same pollution area.
[0033] Further, the specific method of using regression algorithm to establish the relationship model between pollutant concentration and environmental factors to predict the trend of pollutant concentration is as follows:
[0034] (1) Integrate p environmental factor data and pollutant concentration data; environmental factor data is represented as an n x p matrix Z, and pollutant concentration data is represented as an n-dimensional vector y:
[0035]
[0036] where z ij represents the jth environmental factor value of the ith sample in matrix Z, y i represents the pollutant concentration value of the ith sample in vector y; 1≤i≤n, 1≤j≤p;
[0037] (2) There is a linear relationship between pollutant concentration y and environmental factors Z = (z1, z2, …, zp), which is: p T
[0038] y = β0+ β1z1+ β2z2+ … + β p z p + ∈;
[0039] where β0, β1, …, β p are the regression coefficients to be estimated, ∈ is the error term, and ∈ ~ N(0, σ 2 );
[0040] (3) Divide the training set and the test set
[0041] Divide the data set (Z, y) into the training set (Z train , y train ) and the test set (Z test , y test ), the training set is used to estimate the regression coefficients, and the test set is used to evaluate the model performance;
[0042] (4) Train the regression model
[0043] Use the least squares method to estimate the regression coefficients to minimize the RSS:
[0044]
[0045] where n train is the number of samples in the training set, is the predicted value of the ith sample;
[0046] The estimated value of the regression coefficient is obtained by solving the normal equation :
[0047]
[0048] (5) Predict the pollutant concentration trend
[0049] For new environmental factor data Z new = (z new1 ,z new2 ,…,z newp ) T , using the trained regression model to predict the pollutant concentration:
[0050]
[0051] The application also provides an automatic monitoring system for environmental pollutants, which executes the above method, comprising:
[0052] A monitoring module comprising a plurality of monitoring nodes arranged in a target monitoring area, each monitoring node comprising a plurality of pollutant sensors and a positioning module;
[0053] A data processing unit for receiving environmental pollutant data transmitted by each monitoring node and preprocessing the raw data;
[0054] A wireless communication module for transmitting the preprocessed data to a data aggregation center;
[0055] A data aggregation center for receiving data transmitted by each monitoring node, performing deep analysis of the data using a machine learning algorithm, and generating a pollution map and a monitoring report;
[0056] An alarm and response module for issuing an alarm signal when the monitoring data exceeds a preset pollution threshold, notifying relevant environmental protection departments and personnel, and automatically generating an emergency treatment plan according to the pollution situation.
[0057] The application has the following advantages compared with the prior art: the application realizes comprehensive and real-time monitoring of environmental pollutants by reasonably arranging monitoring nodes and a plurality of sensors, the data obtained covers a plurality of types of pollutants such as air and water, and the monitoring range is wide. A scientific data preprocessing method is adopted to effectively remove outliers and fill in missing values, improving the accuracy and usability of the data and providing a reliable basis for subsequent data analysis. Deep data analysis is performed using a machine learning algorithm, the pollution area is accurately divided through a clustering algorithm, the relationship model between the pollutant concentration and the environmental factors is established through a regression algorithm, and the change trend is predicted, which can more deeply understand the distribution and change rule of environmental pollutants and provide strong support for environmental protection decisions. The application has a perfect alarm and response mechanism, which can quickly issue an alarm and generate an emergency treatment plan when a pollution event occurs, improving the response capability to environmental pollution events and helping to timely control and mitigate pollution hazards. BRIEF DESCRIPTION OF DRAWINGS
[0058] Fig. 1 is a flowchart of an automatic monitoring method for environmental pollutants.
[0059] Fig. 2 is a system block diagram of an automatic monitoring system for environmental pollutants. DETAILED DESCRIPTION
[0060] Various exemplary embodiments of the present application will be described herein below with reference to the accompanying drawings. Note that the relative arrangement, numerical expressions, and numerical values of components and steps set forth in these embodiments are not limiting to the scope of the present application unless specifically stated otherwise.
[0061] The following description of at least one exemplary embodiment is merely exemplary in nature and is in no way intended to limit the scope of the application its application or uses.
[0062] Techniques, methods, and apparatus known to those of ordinary skill in the relevant art can not be discussed in detail herein, but should be considered as if the techniques, methods, and apparatus were discussed with reference to the specification.
[0063] In all examples shown and discussed herein, any specific values should be interpreted as merely exemplary and not as a limitation. Thus, other examples of the exemplary embodiments can have different values.
[0064] An automatic monitoring system for environmental pollutants and a method thereof will be described in further detail below with reference to the accompanying drawings.
[0065] In conjunction with the accompanying Figs. 1-2 , the present application will be described in detail.
[0066] An automatic monitoring method for environmental pollutants, specifically comprising the following steps:
[0067] S1, monitoring node arrangement: In the target monitoring area, according to the regional terrain and pollution source distribution, a plurality of monitoring nodes are arranged reasonably. Each monitoring node is provided with a plurality of pollution sensors and a positioning module, wherein the pollution sensors are used to collect real-time environmental pollution data of different types, which include but are not limited to particulate matter, sulfur dioxide, nitrogen oxides in the atmosphere, and chemical oxygen demand, ammonia nitrogen in water, etc.; the positioning module is used to obtain the geographic position information of the monitoring node, so as to analyze the spatial positioning of the monitoring data subsequently.
[0068] S2, data acquisition and preprocessing: The pollution sensors of each monitoring node continuously collect environmental pollution data in real time according to the pre-set sampling frequency, and transmit the collected data to the data processing unit in time. After receiving the original data, the data processing unit performs preprocessing operation on it, mainly including removing outliers and filling missing values. The judgment of outliers adopts the 3σ principle based on statistics, that is, if the data deviates from the mean value by more than 3 times the standard deviation, it is determined that the data is an outlier and is removed.
[0069] For missing values, the method of average value of adjacent data is used to fill in, specifically: for the environmental pollutant data sequence collected in time sequence, first locate the missing value position, if the missing value is located at the beginning of the sequence, take the average value of a number of data points (a suitable positive integer set according to the actual situation, such as 3 or 5, the specific value can be determined according to the data fluctuation and monitoring demand) adjacent to the right side to fill in; if the missing value is located at the end of the sequence, take the average value of a number of data points (the same as above) adjacent to the left side to fill in; if the missing value is located in the middle of the sequence, take the average value of a number of data points on the left and right sides to fill in. Through these preprocessing operations, the accuracy and usability of the data are improved.
[0070] S3, data transmission: the preprocessed data is transmitted to the data aggregation center through the wireless communication module. The wireless communication module adopts the strategy of combining 5G communication with satellite communication, and uses 5G communication in areas with good 5G signal to realize fast data transmission; in areas with weak 5G signal or remote areas, satellite communication is adopted to ensure that data can be stably and uninterruptedly transmitted to the data aggregation center.
[0071] S4, data analysis and processing: after the data aggregation center receives the data transmitted by each monitoring node, the data is analyzed in depth by using machine learning algorithm. First, the monitoring data is analyzed by clustering algorithm, the specific process is as follows:
[0072] For n monitoring nodes, each monitoring node has m characteristics related to pollution degree, and the geographical position information additionally increases l characteristics; the monitoring data is represented as:
[0073] Wherein, x ij represents the jth feature value of the ith monitoring node in matrix X, 1≤i≤n, 1≤j≤(m+l);
[0074] Divide n data points into K clusters So that the sum of squares of within-cluster errors (SSE) is minimized:
[0075] Wherein, μ k is the centroid of the kth cluster C k , ||·|| represents the Euclidean distance;
[0076] For different K values, K = 1, 2, …, K max , calculate the corresponding SSE value; by observing the curve of SSE changing with K, find the turning point where the decreasing amplitude of SSE suddenly slows down, and the K value corresponding to this point is the appropriate clustering number;
[0077] Execute clustering algorithm:
[0078] Initialize the centroid: Randomly select K data points as the initial centroid
[0079] Assign data points to clusters: For each data point x i , calculate the distance to each centroid And assign it to the cluster with the closest distance, that is, If and only if
[0080] Update the centroid: For each cluster Update its centroid to in It is a cluster The number of data points in ;
[0081] Repeat assigning data points to clusters and updating centroids until the centroids no longer change significantly or the maximum number of iterations is reached;
[0082] According to the final clustering results, the K The monitoring nodes are divided into the same pollution area, thereby realizing the accurate division of the monitoring area.
[0083] Afterwards, a regression algorithm was used to establish a relationship model between pollutant concentrations and environmental factors to predict the changing trend of pollutant concentrations. The specific method is as follows:
[0084] Integrate p environmental factor data and pollutant concentration data; the environmental factor data is represented as an n×p matrix Z, and the pollutant concentration data is represented as an n-dimensional vector y:
[0085]
[0086] Among them, z ij represents the jth environmental factor value of the i-th sample in the matrix Z, y i Represents the pollutant concentration value of the i-th sample in vector y; 1≤i≤n, 1≤j≤p;
[0087] Pollutant concentration y and environmental factors Z = (z1, z2, ..., z p ) T There is a linear relationship between them, which is:
[0088] y=β0+γ1z1+β2z2+…+β p z p +∈;
[0089] Among them, β0, β1,…, β p is the regression coefficient to be estimated, ∈ is the error term, ∈~N(0,σ 2 );
[0090] The dataset (Z, y) is divided into a training set (Z train ,y train ) and a test set (Z test ,y test ), the training set is used to estimate the regression coefficients, and the test set is used to evaluate the model performance;
[0091] The regression coefficients are estimated using the least squares method So that the residual sum of squares (RSS) is minimized:
[0092]
[0093] Where n train is the number of samples in the training set, is the predicted value of the i-th sample;
[0094] The estimated value of the regression coefficient is obtained by solving the normal equation :
[0095]
[0096] For new environmental factor data Z new =(z new1 ,z new2 ,…,z newp ) T , the trained regression model is used to predict the pollutant concentration:
[0097]
[0098] Finally, according to the above analysis results, a pollution map and a monitoring report are generated.
[0099] S5, alarm and response: when the monitoring data exceeds the preset pollution threshold, the data aggregation center immediately sends an alarm signal, and notifies the relevant environmental protection departments and staff through SMS, email and other means. At the same time, the system automatically generates an emergency treatment plan according to the pollution situation, for example, for air pollution, it can provide pollution source investigation suggestions, emission reduction measures, etc., so that relevant personnel can take timely measures to respond to pollution incidents.
[0100] Based on the above automatic monitoring method, the present application also provides an automatic monitoring system for environmental pollutants, which executes the above-mentioned automatic monitoring method for environmental pollutants, specifically comprising:
[0101] Monitoring module: including a plurality of monitoring nodes arranged in the target monitoring area, each monitoring node includes a plurality of pollutant sensors and a positioning module, for realizing environmental pollutant data acquisition and geographic position information acquisition.
[0102] Data processing unit: used for receiving environmental pollutant data transmitted by each monitoring node, and preprocessing the raw data, including removing outliers and filling missing values, to improve data quality.
[0103] Wireless communication module: used for transmitting preprocessed data to data aggregation center, using a combination of 5G communication and satellite communication to ensure stable and efficient data transmission.
[0104] Data aggregation center: used for receiving data transmitted by each monitoring node, using machine learning algorithms for deep analysis of data, including clustering analysis and regression analysis, to generate pollution maps and monitoring reports.
[0105] Alarm and response module: when monitoring data exceeds the preset pollution threshold, an alarm signal is sent to notify relevant environmental protection departments and personnel, and an emergency treatment plan is automatically generated according to the pollution situation.
[0106] The specific implementation process of the automatic monitoring system and method of environmental pollutants is as follows:
[0107] In an industrial park and its surrounding area (target monitoring area), environmental pollutant automatic monitoring is carried out, and the specific implementation process is as follows:
[0108] I. Monitoring node arrangement
[0109] Ten monitoring nodes are arranged in the industrial park and its surrounding area. Each monitoring node is equipped with the following sensors:
[0110] Atmospheric pollutant sensor: can collect real-time concentration data of PM2.5, PM10, sulfur dioxide (SO2), and nitrogen oxides (calculated as NO2);
[0111] Water pollutant sensor: used to collect chemical oxygen demand (COD) and ammonia nitrogen (NH3-N) concentration data in nearby rivers and ditches;
[0112] Positioning module: uses GPS positioning to obtain the latitude and longitude coordinates of each monitoring node.
[0113] II. Data collection and preprocessing
[0114] 1. Data collection
[0115] The pollutant sensor of each monitoring node is set to collect data once every 1 hour.
[0116] 2. Data preprocessing
[0117] (1) Outlier judgment: taking the PM2.5 data of monitoring node 1 as an example, after 24 hours of data collection, the mean value of PM2.5 data in this time period is calculated Standard deviation σ = 2. According to the 3σ principle, the normal data range is If the PM2.5 data collected at a certain time is 35, it is determined that the data is an outlier and is removed.
[0118] (2) Missing value filling: If the COD data of monitoring node 2 at 3:00 is missing, the data before and after the time is: COD at 2:00 is 22, and COD at 4:00 is 24. Since the missing value is located in the middle of the sequence, the average value of the adjacent data is used for filling, that is, the filling value is (22+24) ÷ 2 = 23.
[0119] III. Data transmission
[0120] The 5G signal in most areas of this area is good, and the monitoring nodes in these areas prefer to use the 5G communication module to transmit the preprocessed data to the data aggregation center; for the two monitoring nodes at the edge of the industrial park with weak signal, satellite communication mode is adopted for data transmission, to ensure that the data is stable and fast to the data aggregation center.
[0121] IV. Data analysis and processing
[0122] 1. Cluster analysis to divide pollution areas
[0123] Data preparation: integrate all pollutant concentration data (6 pollutants, 6 features) and geographic location information (2 features) of 10 monitoring nodes to form a 10x8 monitoring data matrix X.
[0124] Determine the number of clusters: calculate the within-cluster sum of squares (SSE) for different values (from 1 to 6) by the elbow rule. The calculation results are shown in Table 1.
[0125] Table 1: Within-cluster sum of squares for different K values
[0126] K SSE 1 1200 2 600 3 300 4 200 5 180 6 170
[0127] Observation found that when K = 3, the SSE decreases slowly, so K = 3 is selected as the number of clusters.
[0128] Execute clustering algorithm: use K-Means algorithm, randomly initialize 3 centroids, and after multiple iterations (such as 10 times), finally divide 10 monitoring nodes into 3 clusters, i.e. 3 different pollution areas. For example, monitoring nodes 1, 2 and 3 are divided into cluster 1, monitoring nodes 4, 5 and 6 are divided into cluster 2, and monitoring nodes 7, 8, 9 and 10 are divided into cluster 3.
[0129] Divide pollution areas: according to the clustering results, use geographic information system (GIS) to visualize the 3 pollution areas.
[0130] 2. Regression analysis to predict pollutant concentration trends
[0131] Data Preparation: Environmental data for the area was collected, including wind speed (m / s), wind direction (degrees), temperature (°C), and humidity (%)—a total of four environmental factors. The six pollutant concentration data from 10 monitoring nodes was integrated with the four environmental factor data to form a dataset. The environmental factor data was represented as a 10×4 matrix Z, and the pollutant concentration data (using PM2.5 as an example) was represented as a 10-dimensional vector y.
[0132] Establish a regression model: PM2.5 concentration y and environmental factors Z = (z1, z2, z3, z4) T There is a linear relationship y=β0+β1z1+β2z2+β3z3+β4z4+∈ (corresponding to wind speed, wind direction, temperature, and humidity, respectively).
[0133] Divide the dataset into training set and test set: Divide the dataset into training set (8 samples) and test set (2 samples) in a ratio of 8:2.
[0134] Training regression model: Using the training set data, the regression coefficient is solved by the least square method. The regression coefficient is obtained by calculation
[0135] Predict pollutant concentration change trend: New environmental factor data is z new1 =3(wind speed 3m / s), z new2 =90 (wind direction 90 degrees), z new3 =25 (temperature 25°C), z new4 =60 (humidity 60%), then the predicted PM2.5 concentration is:
[0136]
[0137] 3. Generate pollution maps and monitoring reports
[0138] Based on the pollution area division results obtained by cluster analysis and the pollutant concentration change trend predicted by regression analysis, a pollution map is generated to intuitively display the distribution of different pollution areas; at the same time, a monitoring report is generated to record in detail the data of each monitoring node, the pollution area division, the pollutant concentration prediction trend and other information.
[0139] 5. Alarm and Response
[0140] The preset PM2.5 pollution threshold is 35 (unit: conventional units). If the PM2.5 concentration data at a monitoring node exceeds this threshold, the data aggregation center will immediately issue an alarm signal and notify the local environmental protection department and park management via text message and email. At the same time, the system automatically generates an emergency response plan based on the pollution situation, such as recommending the investigation of possible nearby pollution sources (such as factory emissions and road dust), taking measures such as increasing the frequency of road watering and requiring factories to temporarily reduce emissions.
[0141] The present invention and its embodiments are described above. This description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if a person skilled in the art is inspired by this and, without departing from the purpose of the present invention, designs structures and embodiments similar to this technical solution without inventiveness, they shall fall within the scope of protection of the present invention.
Claims
1. A method for automatically monitoring environmental pollutants, characterized in that: The following steps are involved: S1. Within the target monitoring area, several monitoring nodes are arranged based on the regional terrain and pollution source distribution. Each monitoring node is equipped with multiple pollutant sensors and a positioning module. The pollutant sensors collect data on different types of environmental pollutants in real time, and the positioning module obtains the geographic location information of the monitoring node. S2. The pollutant sensors at each monitoring node collect environmental pollutant data in real time according to the set sampling frequency and transmit the collected data to the data processing unit; the data processing unit preprocesses the raw data, including removing outliers and filling missing values; S3, the pre-processed data is transmitted to the data aggregation center through the wireless communication module; S4. After receiving the data transmitted by each monitoring node, the data aggregation center performs cluster analysis on the monitoring data using a clustering algorithm, dividing the monitoring area into different pollution areas; using a regression algorithm to establish a relationship model between pollutant concentrations and environmental factors, and predict the changing trend of pollutant concentrations; and generating a pollution map and monitoring report based on the analysis results; S5. When the monitoring data exceeds the preset pollution threshold, the data aggregation center will immediately issue an alarm signal and notify the relevant environmental protection departments and staff via SMS and email; at the same time, an emergency treatment plan will be generated based on the pollution situation.
2. The automatic monitoring method for environmental pollutants according to claim 1, characterized in that: The environmental pollutant data include particulate matter, sulfur dioxide, nitrogen oxides in the atmosphere, and chemical oxygen demand and ammonia nitrogen in water bodies.
3. The automatic monitoring method for environmental pollutants according to claim 2, characterized in that: The outliers were determined using the 3σ principle based on statistics; if the data deviated from the mean by more than 3 times the standard deviation, it was determined to be an outlier.
4. The automatic monitoring method for environmental pollutants according to claim 3, characterized in that: The missing values are filled using the average value of adjacent data. The specific method is as follows: For the environmental pollutant data series collected in chronological order, locate the missing value position. If the missing value is at the beginning of the sequence, it is filled by taking the average value of several data points immediately to its right; if the missing value is at the end of the sequence, it is filled by taking the average value of several data points immediately to its left; if the missing value is in the middle of the sequence, it is filled by taking the average value of several data points on its left and right.
5. The automatic monitoring method for environmental pollutants according to claim 4, characterized in that: The wireless communication module combines 5G communication with satellite communication. 5G communication is used first in areas with good 5G signals, and satellite communication is used in areas with weak 5G signals or in remote areas.
6. The automatic monitoring method for environmental pollutants according to claim 5, characterized in that: The specific method of clustering the monitoring data and dividing the monitoring area into different pollution areas by clustering algorithm is as follows: (1) For n monitoring nodes, each monitoring node has m features related to the pollution level, and the geographical location information has an additional feature of l; the monitoring data is expressed as: Among them, x ij represents the j-th eigenvalue of the i-th monitoring node in the matrix X, 1≤i≤n, 1≤j≤(m+l); (2) Divide n data points into K clusters Minimize the SSE within the cluster: Among them, μ k is the kth cluster C k The centroid of , ||·|| represents the Euclidean distance; (3) For different K values, K=1,2,…,K max , calculate the corresponding SSE value; by observing the curve of SSE changing with K, find the turning point where the SSE decreases suddenly and slowly, and the K value corresponding to this point is the appropriate number of clusters; (4) Execute clustering algorithm Initialize the centroid: Randomly select K data points as the initial centroid Assign data points to clusters: For each data point x i , calculate the distance to each centroid And assign it to the cluster with the closest distance, that is, If and only if Update the centroid: For each cluster Update its centroid to in It is a cluster The number of data points in ; Repeat assigning data points to clusters and updating centroids until the centroids no longer change significantly or the maximum number of iterations is reached; (5) Division of pollution areas According to the final clustering results, the K The monitoring nodes are divided into the same pollution area.
7. The automatic monitoring method for environmental pollutants according to claim 6, characterized in that: The specific method of using regression algorithm to establish the relationship model between pollutant concentration and environmental factors and predict the changing trend of pollutant concentration is as follows: (1) Integrate p environmental factor data and pollutant concentration data; the environmental factor data is represented as an n×p matrix Z, and the pollutant concentration data is represented as an n-dimensional vector y: Among them, z ij represents the jth environmental factor value of the i-th sample in the matrix Z, y i Represents the pollutant concentration value of the i-th sample in vector y; 1≤i≤n, 1≤j≤p; (2) Pollutant concentration y and environmental factors Z = (z1, z2,…, z p ) T There is a linear relationship between them, which is: y=β0+β1z1+β2z2+…+β p z p +∈; Among them, γ0,γ1,…,γ p is the regression coefficient to be estimated, ∈ is the error term, ∈~N(0,σ 2 ); (3) Divide the training set and test set Divide the dataset (Z, y) into training set (Z train ,y train ) and the test set (Z test ,y test ), the training set is used to estimate the regression coefficients, and the test set is used to evaluate the model performance; (4) Training regression model Estimate regression coefficients using the least squares method Minimize RSS: Among them, n train is the number of samples in the training set, is the predicted value of the i-th sample; The estimated values of the regression coefficients are obtained by solving the normal equation get: (5) Predict pollutant concentration trends For the new environmental factor data Z new =(z new1 ,z new2 ,…,z newp ) T , use the trained regression model to predict pollutant concentrations:
8. An automatic monitoring system for environmental pollutants, which executes the automatic monitoring method for environmental pollutants according to any one of claims 1 to 7, characterized in that: include: The monitoring module includes several monitoring nodes, which are arranged in the target monitoring area. Each monitoring node includes multiple pollutant sensors and a positioning module. The data processing unit is used to receive the environmental pollutant data transmitted by each monitoring node and pre-process the raw data; A wireless communication module for transmitting pre-processed data to a data aggregation center; The data aggregation center receives data transmitted by each monitoring node, uses machine learning algorithms to conduct in-depth analysis of the data, and generates pollution maps and monitoring reports; The alarm and response module sends out an alarm signal when the monitoring data exceeds the preset pollution threshold, notifies the relevant environmental protection departments and staff, and automatically generates an emergency treatment plan based on the pollution situation.