Atmospheric pollution trend analysis method and system

Abstract

The invention discloses an atmospheric pollution trend analysis method and system, and relates to the technical field of atmospheric pollution trend analysis, and the method comprises the steps: collecting multi-source data in a target region, calculating the concentration change rate of all nodes by using the law of conservation of mass, updating the concentration of the nodes by using an Euler method, and inputting the updated concentration of the nodes into a pre-trained GCN model, calculating the optimized concentration and production abnormity list; modulating a propagation weight by using a Hebbian learning rule, modulating the optimized concentration by using the modulated weight, calculating a joint concentration by using a conservation type path flux method, combining the joint concentration by using a weighted linear combination, and calculating the joint concentration; and constructing a long-short-term memory network model to predict the future trend of the pollutants. The space gridding and the graph neural network are introduced, the recognition and control capability of the pollution source is improved, and the prediction precision and reliability of the analysis model are enhanced by combining the diffusion flux model and the convection flux model.

Description

Technical Field

[0001] The present invention relates to the technical field of air pollution trend analysis, and particularly to an air pollution trend analysis method and system. Background Art

[0002] With the increasingly severe global environmental pollution problems, the monitoring and analysis of air pollution have become an important research field. With the improvement of sensor technology and data processing capabilities, environmental monitoring means have gradually developed from traditional manual sampling to automation and intelligence. Especially in air pollution monitoring, sensor technology has been widely used. Although modern environmental monitoring systems have made significant progress in data collection, existing air pollution trend analysis methods still have certain limitations.

[0003] There are deficiencies in existing air pollution trend analysis methods. Existing methods overly rely on a single data source and fail to fully explore the correlations between different types of data, resulting in insufficient accuracy and reliability of analysis results. Existing methods generally lack effective modeling of the propagation laws of air pollutants in space and time. Especially in dynamic and complex urban environments, the changes in pollutant concentrations are affected by a variety of factors. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides an air pollution trend analysis method and system to solve the problems that existing methods overly rely on a single data source, fail to fully explore the correlations between different types of data, resulting in insufficient accuracy and reliability of analysis results, and existing methods generally lack effective modeling of the propagation laws of air pollutants in space and time. Especially in dynamic and complex urban environments, the changes in pollutant concentrations are affected by a variety of factors.

[0006] To solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides an air pollution trend analysis method, which includes,

[0008] Collect multi-source data in a target area, divide the target area into grids using spatial gridification to generate a directed graph, calculate the concentration change rate of all nodes using the law of conservation of mass, update the concentration of nodes using the Euler method, input the updated concentration of nodes into a pre-trained GCN model, calculate the optimized concentration, and generate an anomaly list;

[0009] Modulate the propagation weights using the Hebbian learning rule, modulate the optimized concentration using the modulated weights, calculate the combined concentration using the conservative path flux method, combine the combined concentration using weighted linear combination, and calculate the combined concentration;

[0010] Construct a long short-term memory network model to predict the future trend of pollutants and classify and monitor the future trend of pollutants;

[0011] Construct a visualization interface to display the monitoring results and store the multi-source data generated by collection and analysis.

[0012] As a preferred embodiment of the air pollution trend analysis method described in the present invention, wherein: collecting multi-source data in the target area, using the law of conservation of mass, calculating the concentration change rate of all nodes, updating the concentration of nodes using the Euler method, inputting the updated concentration of nodes into a pre-trained GCN model, calculating the optimized concentration, and generating an anomaly list, including:

[0013] Deploy intelligent sensors in the target area to collect multi-source data;

[0014] The intelligent sensors include chemical, optical, temperature and humidity, ultrasonic anemometer, camera, and current sensor;

[0015] The multi-source data includes pollutant concentration, temperature and humidity, wind speed, wind direction angle, traffic flow, and current data;

[0016] The pollutant concentration includes PM2.5, PM10, sulfur dioxide, nitrogen dioxide, and ozone data;

[0017] Synchronize the time of the multi-source data using the Network Time Protocol, identify and delete abnormal data using the interquartile range, fill in the missing data using the spline interpolation method, and normalize the multi-source data;

[0018] Divide the target area into grids using spatial gridding, with each grid as a node of a directed graph, and define the center of the grid as the node coordinates (x i , y i ) using the Cartesian coordinate system position, where i is the grid number;

[0019] Set the neighbors using the grid adjacency method to generate a set of neighbor nodes, and calculate the degree of the nodes using the node degree calculation method;

[0020] Calculate the angle between nodes using the arctangent function, and calculate the propagation intensity based on the wind speed and the angle between nodes using the propagation intensity formula. If the propagation intensity is less than 0, set the propagation intensity to 0;

[0021] Set the propagation intensity threshold using the percentile threshold method, and filter out the propagation intensity greater than the threshold as an edge of the directed graph to generate a directed graph.

[0022] Assign the pollutant concentration to the nodes in the directed graph. If a node has no pollutant concentration, use the inverse distance weighted method for interpolation to generate the initial pollutant concentration.

[0023] Set the diffusion coefficient using the theoretical calculation method, calculate the diffusion flux using Fick's law, calculate the convective flux using the convective flux model, add the diffusion flux and the convective flux to calculate the total flux, and if the total flux is less than 0, set the total flux to 0.

[0024] Use the law of conservation of mass to calculate the concentration change rate of all nodes, and use the Euler method to update the concentration of the nodes.

[0025] Input the updated concentration of the nodes into the pre-trained GCN model to calculate the optimized concentration, set the classification threshold using the statistical distribution threshold method, and filter out the optimized concentration greater than the threshold to generate the anomaly list A(t).

[0026] As a preferred scheme of the air pollution trend analysis method described in the present invention, wherein: modulating the propagation weight using the Hebbian learning rule, modulating the optimized concentration using the modulated weight, calculating the combined concentration using the conservative path flux method, and combining the combined concentration using the weighted linear combination, calculating the combined concentration, including:

[0027] Based on the propagation intensity, modulate the propagation weight using the Hebbian learning rule, modulate the optimized concentration using the modulated weight, and calculate the propagation probability using the Logistic function method.

[0028] If node i ∈ A(t), introduce the control intensity to adjust the propagation probability, otherwise where P ij (t) is the propagation probability of the pollutant from the i-th node to the j-th node at time t, is the adjusted propagation probability of the pollutant from the i-th node to the j-th node at time t;

[0029] Based on the adjusted propagation probability, calculate the adjusted propagation intensity, and calculate the combined concentration using the conservative path flux method.

[0030] Calculate the product of the current and time to obtain the electricity consumption, and perform standardization processing. Use linear mapping to map the standardized electricity consumption to the economic activity index range to generate the economic activity index.

[0031] Use the weighted linear combination to combine the combined concentration to calculate the combined concentration.

[0032] As a preferred embodiment of the air pollution trend analysis method of the present invention, wherein: constructing a long short-term memory network model to predict the future trend of pollutants includes:

[0033] Collecting combined concentration data from the development data platform to generate a training set;

[0034] Constructing a long short-term memory network model, including an input layer, an LSTM layer, and a fully connected output layer;

[0035] Setting the format of the input layer as combined concentration;

[0036] Training the long short-term memory network model with the training set, and using a loss function and an Adam optimizer to iteratively optimize the model parameters;

[0037] Inputting the combined concentration into the trained long short-term memory network model to predict the future trend of pollutants.

[0038] As a preferred embodiment of the air pollution trend analysis method of the present invention, wherein: classifying and monitoring the future trend of pollutants includes:

[0039] Calculating the trend component of the future trend of pollutants using the moving average method, calculating the change rate of the trend component using linear regression, setting the classification threshold range using statistical analysis, determining that the change rate greater than the classification threshold range is an upward trend, determining that the change rate equal to the classification threshold range is a stable trend, and determining that the change rate less than the classification threshold range is a downward trend;

[0040] Setting classification weights using the threshold segmentation weighting method, calculating the risk index using the weighted integral method, setting the risk threshold using the threshold screening method, screening the risk index greater than the risk threshold, and generating high-risk areas.

[0041] As a preferred embodiment of the air pollution trend analysis method of the present invention, wherein: constructing a visual interface to display the monitoring results includes:

[0042] Constructing a visual interface using the Dash framework of the visualization tool Python, displaying the future trend of pollutants in real time using a data visualization tool, and displaying high-risk areas using a map visualization tool;

[0043] Allowing users who have passed real-name verification to view.

[0044] As a preferred embodiment of the air pollution trend analysis method of the present invention, wherein: storing the multi-source data generated by collection and analysis includes:

[0045] Store the collected multi-source data and the high-risk areas generated by the analysis in a central database, and set security access measures. The central database will perform cloud backup on the stored data, and regularly perform integrity detection on the stored data and the backup data. After the detection is completed, generate an integrity detection record and synchronously store it in the central database.

[0046] In a second aspect, the present invention provides an air pollution trend analysis system, including

[0047] A collection optimization module for collecting multi-source data in a target area, dividing the target area into grids using spatial gridding to generate a directed graph, calculating the concentration change rate of all nodes using the law of conservation of mass, updating the concentration of nodes using the Euler method, inputting the updated concentration of nodes into a pre-trained GCN model to calculate the optimized concentration, and generating an anomaly list;

[0048] A modulation joint module for modulating the propagation weight using the Hebbian learning rule, modulating the optimized concentration using the modulated weight, calculating the joint concentration using the conservative path flux method, and combining the joint concentration using weighted linear combination to calculate the joint concentration;

[0049] A construction monitoring module for constructing a long short-term memory network model to predict the future trend of pollutants and classify and monitor the future trend of pollutants;

[0050] A visualization storage module for constructing a visualization interface to display the monitoring results and storing the multi-source data collected and generated by the analysis.

[0051] In a third aspect, the present invention provides a computer device, including a memory and a processor, where the memory stores a computer program, and: when the computer program is executed by the processor, any step of the air pollution trend analysis method described in the first aspect of the present invention is implemented.

[0052] In a fourth aspect, the present invention provides a computer-readable storage medium, on which a computer program is stored, and: when the computer program is executed by the processor, any step of the air pollution trend analysis method described in the first aspect of the present invention is implemented.

[0053] The beneficial effects of the present invention are as follows: By collecting multi-source data in the target area, the present invention divides the target area into grids using spatial gridding, generates a directed graph, calculates the concentration change rate of all nodes using the law of conservation of mass, updates the concentration of nodes using the Euler method, inputs the updated concentration of nodes into a pre-trained GCN model to calculate the optimized concentration, and generates an anomaly list; modulates the propagation weights using the Hebbian learning rule, modulates the optimized concentration using the modulated weights, calculates the combined concentration using the conservative path flux method, combines the combined concentration using weighted linear combination to calculate the combined concentration; constructs a long short-term memory network model to predict the future trend of pollutants, introduces spatial gridding and graph neural networks to improve the identification and control ability of pollution sources, and combines the diffusion flux and advection flux models to enhance the prediction accuracy and reliability of the analysis model. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0055] Figure 1 It is a flowchart of the air pollution trend analysis method in Embodiment 1.

[0056] Figure 2 It is a schematic diagram of the air pollution trend analysis system in Embodiment 1.

[0057] Figure 3 It is a flowchart of calculating the combined concentration in Embodiment 1.

[0058] Figure 4 It is a schematic diagram of constructing a long short-term memory network model to predict the future trend of pollutants in Embodiment 1. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0059] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following will provide a detailed description of the specific embodiments of the present invention in conjunction with the drawings in the specification.

[0060] Many specific details are set forth in the following description to facilitate a full understanding of the present invention. However, the present invention can also be implemented in other ways different from those described herein. Those skilled in the art can make similar generalizations without departing from the spirit of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0061] Second, the "one embodiment" or "embodiment" referred to herein means a specific feature, structure, or characteristic that may be included in at least one implementation manner of the present invention. The phrase "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that excludes other embodiments.

[0062] Embodiment 1, referring to Figures 1 to 4 , is the first embodiment of the present invention. This embodiment provides a method for analyzing the trend of air pollution, including the following steps:

[0063] S1. Collect multi-source data in the target area, divide the target area into grids using spatial gridding to generate a directed graph, calculate the concentration change rate of all nodes using the law of conservation of mass, update the concentration of nodes using the Euler method, input the updated concentration of nodes into a pre-trained GCN model, calculate the optimized concentration, and generate an anomaly list;

[0064] Specifically, collecting multi-source data in the target area, calculating the concentration change rate of all nodes using the law of conservation of mass, updating the concentration of nodes using the Euler method, inputting the updated concentration of nodes into a pre-trained GCN model, calculating the optimized concentration, and generating an anomaly list, including:

[0065] Deploy intelligent sensors in the target area to collect multi-source data;

[0066] The intelligent sensors include chemical, optical, temperature and humidity, ultrasonic anemometer, camera, and current sensors;

[0067] The multi-source data includes pollutant concentration, temperature and humidity, wind speed, wind direction angle, traffic flow, and current data;

[0068] The pollutant concentration includes PM2.5, PM10, sulfur dioxide, nitrogen dioxide, and ozone data;

[0069] Synchronize the time of multi-source data using the Network Time Protocol, identify and delete abnormal data using the interquartile range, fill in missing data using spline interpolation, and normalize the multi-source data;

[0070] Divide the target area into grids using spatial gridding. Each grid is used as a node of the directed graph, and the center of the grid is defined as the node coordinate (x i , y i ) using the Cartesian coordinate system position, where i is the grid number;

[0071] Set neighbors using the grid adjacency method to generate a set of neighbor nodes, and calculate the degree of the nodes using the node degree calculation method;

[0072] The included angle between nodes is calculated using the arctangent function, and the formula is:

[0073]

[0074] where θ ij is the included angle between the i-th and j-th nodes, θ wind is the wind direction angle, x i and y i are the coordinates of the i-th node, x j and y j are the coordinates of the j-th node;

[0075] Based on the wind speed and the included angle between nodes, the propagation intensity is calculated using the propagation intensity formula. If the propagation intensity is less than 0, set the propagation intensity to 0, and the formula is:

[0076] S ij = v - cos(θ ij ),

[0077] where S ij is the propagation intensity of pollutants from the i-th to the j-th node, and v is the wind speed;

[0078] The percentile threshold method is used to set the propagation intensity threshold, and the propagation intensities greater than the propagation intensity threshold are screened as an edge of the directed graph to generate a directed graph;

[0079] The pollutant concentration is assigned to the nodes in the directed graph. If a node has no pollutant concentration, the inverse distance weighted method is used for interpolation to generate the initial pollutant concentration;

[0080] The diffusion coefficient is set using the theoretical calculation method, and the diffusion flux is calculated using Fick's law. The formula is:

[0081]

[0082] where is the diffusion flux of pollutants from the i-th to the j-th node, D is the diffusion coefficient (fitted through experimental data), representing the diffusion rate of pollutants in the medium, Δx is the spatial distance between two nodes (calculated using the Euclidean distance formula), C j (t0) and C i (t0) are the initial pollutant concentrations of the i-th and j-th nodes respectively;

[0083] The convective flux is calculated using the convective flux model, and the formula is:

[0084]

[0085] where is the convective flux of pollutants from the i-th to the j-th node;

[0086] Add the diffusion flux and the convective flux to calculate the total flux. If the total flux is less than 0, set the total flux to 0;

[0087] Use the law of conservation of mass to calculate the rate of change of concentration at all nodes. The formula is:

[0088]

[0089] where is the rate of change of concentration at the i-th node, N(i) is the set of neighbor nodes of the i-th node, and J ji is the total flux between the i-th and j-th nodes;

[0090] Use the Euler method to update the concentration of the nodes. The formula is:

[0091]

[0092] where C i (t1) is the updated concentration of the i-th node, and Δt is the time step (set using a fixed time step);

[0093] Input the updated concentration of the nodes into the pre-trained GCN model. Optimize the concentration distribution through graph convolution operations to capture the non-linear propagation relationship between grids and calculate the optimized concentration. The formula is:

[0094]

[0095] where C′ i (t1) is the optimized concentration of the i-th node, σ is the ReLU activation function, σ(x) = max(0, x), x is the data to be input, d i and d j are the degrees of the i-th and j-th nodes respectively;

[0096] Use the statistical distribution threshold method to set the classification threshold, filter the optimized concentration greater than the classification threshold, and generate the anomaly list A(t).

[0097] By integrating multi-dimensional data collection methods such as chemistry, optics, meteorology, traffic, and current, it can not only reflect the concentration changes of atmospheric pollutants in real time, but also combine environmental and human activity factors to provide more comprehensive data support for pollution trend modeling. Combining the spatial relationship between wind speed, wind direction and nodes, a propagation intensity model is constructed, and effective edges are screened by the percentile threshold method to realize the dynamic screening of pollution propagation paths, thereby enhancing the real-world binding force and propagation mechanism expression ability of the model. At the same time, Fick's law and convection model are introduced to characterize pollutant fluxes, solving the limitations of single-mechanism modeling of traditional models, being able to more comprehensively reflect the diffusion behavior of pollutants under different meteorological conditions, and performing discrete updates in units of time steps, with the advantages of simple algorithm implementation and high computational efficiency, suitable for deployment on edge computing devices or real-time analysis systems. By extracting the non-linear pollution propagation mode under spatial adjacency relationship through the graph neural network model, it not only improves the concentration prediction accuracy, but also enhances the system's response ability to abnormal diffusion events. Based on the optimized concentration distribution, combined with statistical distribution methods, a list of abnormal pollution areas is dynamically generated, which helps environmental supervision departments quickly identify high-risk areas and formulate intervention strategies.

[0098] S2. Modulate the propagation weight using the Hebbian learning rule, modulate the optimized concentration using the modulated weight, calculate the joint concentration using the conservative path flux method, and combine the joint concentration using weighted linear combination to calculate the joint concentration;

[0099] Specifically, modulating the propagation weight using the Hebbian learning rule, modulating the optimized concentration using the modulated weight, calculating the joint concentration using the conservative path flux method, and combining the joint concentration using weighted linear combination to calculate the joint concentration includes:

[0100] Based on the propagation intensity, modulate the propagation weight using the Hebbian learning rule to simulate the dynamic adjustment of synaptic strength in neuromorphology, and the formula is:

[0101]

[0102] where S′ ij (t) is the modulated propagation weight of pollutants between the i-th and j-th nodes at time t, and S ij (t) is the propagation intensity of pollutants between the i-th and j-th nodes at time t, and T i (t) and T j (t) are the temperatures of the i-th and j-th nodes at time t respectively, and H i (t) and H j (t) are the humidities of the i-th and j-th nodes at time t respectively;

[0103] Modulate the optimized concentration using the modulated weights to simulate signal transmission between neurons. The formula is:

[0104] C″ i (t) = C′ i (t) + ∑ j∈N(i) S′ ij (t)·(C j (t1) - C i (t1)),

[0105] where C″ i (t) is the concentration modulated at the i-th node at time t;

[0106] Use the Logistic function method to calculate the propagation probability. The formula is:

[0107]

[0108] where P ij (t) is the propagation probability of the pollutant from the i-th node to the j-th node at time t, θ is the propagation sensitivity parameter that controls the steepness of the Logistic function, and C″ j (t) is the concentration modulated at the j-th node at time t;

[0109] If node i ∈ A(t), introduce the control intensity to adjust the propagation probability, otherwise The formula is:

[0110]

[0111] where is the adjusted propagation probability of the pollutant from the i-th node to the j-th node at time t, and γ i (t) is the control intensity at the i-th node at time t (set using the empirical rule);

[0112] Based on the adjusted propagation probability, calculate the adjusted propagation intensity. The formula is:

[0113]

[0114] where D ij (t) is the adjusted propagation intensity of the pollutant from the i-th node to the j-th node at time t;

[0115] Use the conservative path flux method to calculate the combined concentration. The formula is:

[0116]

[0117] where is the combined concentration at the i-th node at time t, and Dji D(t) is the adjusted propagation intensity of pollutants from the j-th node to the i-th node at time t. ik D(t) is the adjusted propagation intensity of pollutants from the i-th node to the k-th node at time t.

[0118] Calculate the product of current and time to obtain the electricity consumption, and perform standardization. Use linear mapping to map the standardized electricity consumption to the economic activity index range [0, 100] to generate the economic activity index. The formula is:

[0119] E i (t) = a·U i (t) + b,

[0120] where E i (t) is the economic activity index of the i-th node at time t, a is the slope (calculated using the two-point form), U i (t) is the standardized electricity consumption of the i-th node at time t, and b is the intercept (calculated using the substitution method);

[0121] Use weighted linear combination to combine the joint concentration and calculate the joint concentration. The formula is:

[0122]

[0123] where I i (t) is the joint concentration of the i-th node at time t, M i (t) is the traffic flow of the i-th node at time t, and α and β are the modulation coefficients of traffic flow and economic activity index respectively (set using the empirical weight tuning method).

[0124] Dynamically adjust the propagation weight by means of the Hebbian learning rule, enabling the system to simulate the dynamic adaptability of pollution propagation paths in reality, improving the accuracy of depicting the migration trend of pollutants under spatio-temporal changes, modulating the propagation weight through the fused input of temperature and humidity, establishing the coupling relationship between meteorological conditions and pollution diffusion, solving the problem of the mismatch between "static propagation intensity" and "static environmental conditions" in traditional models, using the Logistic function and the control intensity mechanism to construct a flexible propagation probability function, endowing the model with the non-linear characteristics of responding to high-concentration excitation and policy regulation, contributing to the construction of a more realistic urban pollution evolution mechanism. The conservative path flux model ensures the conservation of pollutant mass, effectively avoiding the phenomenon of "pollutant disappearance" caused by information loss or path setting errors in numerical simulation, and enhancing the credibility of simulation results. By introducing the economic activity index and traffic flow index, indirect modeling of the pollution generation mechanism is realized, enabling the system to not only analyze the spatial propagation characteristics of pollution but also predict its potential generation sources, contributing to the source control of urban environmental governance. The empirical weight tuning method is used in the weighted linear combination, facilitating the rapid adaptation of the system to different cities, different seasons, or different regulatory policies, and having good scalability and deployment flexibility.

[0125] S3. Construct a long short-term memory network model to predict the future trend of pollutants and classify and monitor the future trend of pollutants;

[0126] Specifically, constructing a long short-term memory network model to predict the future trend of pollutants includes:

[0127] Collect combined concentration data from the development data platform to generate a training set;

[0128] Construct a long short-term memory network model, including an input layer, an LSTM layer, and a fully connected output layer;

[0129] Set the format of the input layer as combined concentration;

[0130] Use the training set to train the long short-term memory network model, and use the loss function and the Adam optimizer to iteratively optimize the model parameters;

[0131] Input the combined concentration into the trained long short-term memory network model to predict the future trend of pollutants.

[0132] The LSTM network is highly sensitive to short-term data changes, can capture the abnormal change trend of pollutants in advance, improve the prediction accuracy of pollution trends, and enhance the ability to respond to sudden pollution events.

[0133] Furthermore, classifying and monitoring the future trend of pollutants includes:

[0134] Calculate the trend component of the future trend of pollutants using the moving average method, calculate the change rate of the trend component using linear regression, set the classification threshold range using statistical analysis, and judge the change rate greater than the classification threshold range as an upward trend, the change rate equal to the classification threshold range as a stable trend, and the change rate less than the classification threshold range as a downward trend;

[0135] Set the classification weights using the threshold segmentation weighting method, calculate the risk index using the weighted integral method, set the risk threshold using the threshold screening method, and screen the risk index greater than the risk threshold to generate a high-risk area.

[0136] Take the trend change rate as the basis of the classification variable, so that the trend judgment is not only based on the concentration value itself, but also combines its change slope, avoiding misleading judgment results due to static concentration, enhancing the prediction sensitivity of the system, and being particularly suitable for capturing the rapid changes of pollutants in a local short time period. This step introduces the "classification weight" mechanism, that is, different risk weights are assigned to different trend categories (upward, stable, downward). Through the threshold segmentation method, a weight function with non-linear growth can be set for different rate intervals, more truly reflecting the impact degree of pollution changes on fields such as health, transportation and industry. It not only realizes the accurate description of pollutant trends, but also provides data support and quantitative basis for the formulation of pollution prevention and control policies.

[0137] S4. Construct a visual interface to display the monitoring results and store the multi-source data generated by collection and analysis;

[0138] Specifically, constructing a visual interface to display the monitoring results includes:

[0139] Use the Dash framework of the visualization tool Python to construct a visual interface, use the data visualization tool to display the future trend of pollutants in real time, and use the map visualization tool to display the high-risk area;

[0140] Allow users who have passed real-name verification to view.

[0141] Dash can directly integrate the output results of the analysis models in Python, enabling seamless docking of pollutant trend prediction results (such as LSTM models, risk indexes, etc.) with the front-end display, reducing the technical burden of intermediate data conversion and deployment. The user access control mechanism not only ensures the security of system data, but also enables the platform to have social service and hierarchical management capabilities.

[0142] Furthermore, store the multi-source data generated by collection and analysis, including:

[0143] The collected multi-source data and the high-risk areas generated by the analysis are stored in the central database, and security access measures are set. The central database backs up the stored data to the cloud, and regularly conducts integrity checks on the stored data and the backup data. After the checks are completed, integrity check records are generated and synchronously stored in the central database.

[0144] The multi-source data stored centrally can be quickly queried and retrieved through the database management system, greatly improving the efficiency of data access and processing. Through strict security access measures, the privacy of the stored data can be effectively protected, preventing unauthorized personnel from illegally accessing or tampering with sensitive data. Regular integrity checks provide traceability and transparency for the system. Users can view the reports and records of each check, thus increasing their trust in the data management system.

[0145] This embodiment also provides an air pollution trend analysis system, including:

[0146] A collection and optimization module, which is used to collect multi-source data in the target area, divide the target area into grids using spatial gridding to generate a directed graph, calculate the concentration change rate of all nodes using the law of conservation of mass, update the concentration of the nodes using the Euler method, input the updated concentration of the nodes into a pre-trained GCN model to calculate the optimized concentration, and generate an anomaly list;

[0147] A modulation and combination module, which is used to modulate the propagation weights using the Hebbian learning rule, modulate the optimized concentration using the modulated weights, calculate the combined concentration using the conservative path flux method, and combine the combined concentration using weighted linear combination to calculate the combined concentration;

[0148] A construction and monitoring module, which is used to construct a long short-term memory network model to predict the future trend of pollutants and classify and monitor the future trend of pollutants;

[0149] A visualization and storage module, which is used to construct a visualization interface to display the monitoring results and store the multi-source data collected and generated by the analysis.

[0150] This embodiment also provides a computer device, which is applicable to the case of the air pollution trend analysis method, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the air pollution trend analysis method proposed in the above embodiment.

[0151] The computer device may be a terminal, which includes a processor, a memory, a communication interface, a display screen, and an input device connected via a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The communication interface of the computer device is used to communicate with external terminals in a wired or wireless manner, and the wireless manner can be achieved through WIFI, carrier networks, NFC (Near Field Communication), or other technologies. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads provided on the housing of the computer device, or an external keyboard, touchpad, or mouse, etc.

[0152] This embodiment also provides a storage medium, on which a computer program is stored. When the program is executed by a processor, it implements the method for analyzing the trend of air pollution proposed in the above embodiment; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (abbreviated as SRAM), electrically erasable programmable read-only memory (abbreviated as EEPROM), erasable programmable read-only memory (abbreviated as EPROM), programmable read-only memory (abbreviated as PROM), read-only memory (abbreviated as ROM), magnetic memory, flash memory, magnetic disks, or optical discs.

[0153] In summary, the present invention collects multi-source data in the target area, divides the target area into grids using spatial gridding to generate a directed graph, calculates the concentration change rate of all nodes using the law of conservation of mass, updates the concentration of nodes using the Euler method, inputs the updated concentration of nodes into a pre-trained GCN model to calculate the optimized concentration, and generates an anomaly list; modulates the propagation weight using the Hebbian learning rule, modulates the optimized concentration using the modulated weight, calculates the combined concentration using the conservative path flux method, combines the combined concentration using weighted linear combination to calculate the combined concentration; constructs a long short-term memory network model to predict the future trend of pollutants, introduces spatial gridding and graph neural network, improves the identification and control ability of pollution sources, and combines the diffusion flux and advection flux models to enhance the prediction accuracy and reliability of the analysis model.

[0154] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

Claims

1. A method for analyzing the trend of air pollution, characterized in that: Including Collect multi-source data in the target area, divide the target area into grids using spatial gridding to generate a directed graph, use the law of conservation of mass to calculate the concentration change rate of all nodes, use the Euler method to update the concentration of nodes, input the updated node concentration into a pre-trained GCN model to calculate the optimized concentration, and generate an anomaly list; Use the Hebbian learning rule to modulate the propagation weight, use the modulated weight to modulate the optimized concentration, use the conservative path flux method to calculate the combined concentration, and use weighted linear combination to combine the combined concentration to calculate the combined concentration; Construct a long short-term memory network model to predict the future trend of pollutants and classify and monitor the future trend of pollutants; Construct a visualization interface to display the monitoring results and store the multi-source data generated by collection and analysis.

2. The air pollution trend analysis method according to claim 1, wherein: The steps of collecting multi-source data in the target area, using the law of conservation of mass to calculate the concentration change rate of all nodes, using the Euler method to update the concentration of nodes, inputting the updated node concentration into a pre-trained GCN model to calculate the optimized concentration, and generating an anomaly list include: Deploy intelligent sensors in the target area to collect multi-source data; The multi-source data includes pollutant concentration, temperature and humidity, wind speed, wind direction angle, traffic flow, and current data; Use the Network Time Protocol to synchronize the time of multi-source data, use the interquartile range to identify and delete abnormal data, use the spline interpolation method to fill in missing data, and normalize the multi-source data; The target area is divided into grids using spatial grid division. Each grid serves as a node in a directed graph, and the center of the grid is defined as the node coordinate (x i , y i ) in the Cartesian coordinate system, where i is the grid number; Use the grid adjacency method to set neighbors and generate a set of neighbor nodes, and use the node degree calculation method to calculate the degree of nodes; Use the arctangent function to calculate the angle between nodes, and based on the wind speed and the angle between nodes, use the propagation intensity formula to calculate the propagation intensity. If the propagation intensity is less than 0, set the propagation intensity to 0; Use the percentile threshold method to set the propagation intensity threshold, screen the propagation intensity greater than the propagation intensity threshold as an edge of the directed graph, and generate a directed graph; Assign the pollutant concentration to the nodes in the directed graph. If a node has no pollutant concentration, use the inverse distance weighted method for interpolation to generate the initial pollutant concentration; Use the theoretical calculation method to set the diffusion coefficient, use Fick's law to calculate the diffusion flux, use the convective flux model to calculate the convective flux, add the diffusion flux and the convective flux to calculate the total flux. If the total flux is less than 0, set the total flux to 0; Use the law of conservation of mass to calculate the concentration change rate of all nodes and use the Euler method to update the concentration of nodes; Input the updated node concentration into a pre-trained GCN model to calculate the optimized concentration, use the statistical distribution threshold method to set the classification threshold, screen the optimized concentration greater than the classification threshold, and generate an anomaly list A(t).

3. The air pollution trend analysis method according to claim 2, wherein: The steps of using the Hebbian learning rule to modulate the propagation weight, using the modulated weight to modulate the optimized concentration, using the conservative path flux method to calculate the combined concentration, and using weighted linear combination to combine the combined concentration to calculate the combined concentration include: Based on the propagation intensity, use the Hebbian learning rule to modulate the propagation weight, use the modulated weight to modulate the optimized concentration, and use the Logistic function method to calculate the propagation probability; If node \(i\in A(t)\), the control intensity is introduced to adjust the propagation probability, otherwise where \(P\) ij (t) is the propagation probability of pollutants from the \(i\)-th node to the \(j\)-th node at time \(t\), is the adjusted propagation probability of pollutants from the \(i\)-th node to the \(j\)-th node at time \(t\); Based on the adjusted propagation probability, calculate the adjusted propagation intensity, and use the conservative path flux method to calculate the combined concentration; Calculate the product of current and time to obtain the power consumption, and perform normalization processing. Use linear mapping to map the normalized power consumption to the economic activity index range to generate the economic activity index; Use weighted linear combination to combine the combined concentration and calculate the combined concentration.

4. The air pollution trend analysis method according to claim 3, characterized in that: The construction of the long short-term memory network model to predict the future trend of pollutants includes: Collect combined concentration data from the development data platform to generate a training set; Construct a long short-term memory network model, including an input layer, an LSTM layer, and a fully connected output layer; Set the format of the input layer as the combined concentration; Use the training set to train the long short-term memory network model, and use the loss function and the Adam optimizer to iteratively optimize the model parameters; Input the combined concentration into the trained long short-term memory network model to predict the future trend of pollutants.

5. The air pollution trend analysis method according to claim 4, wherein: The classification and monitoring of the future trend of pollutants include: Use the moving average method to calculate the trend component of the future trend of pollutants, use linear regression to calculate the change rate of the trend component, use statistical analysis to set the classification threshold range, and judge the change rate greater than the classification threshold range as an upward trend, the change rate equal to the classification threshold range as a stable trend, and the change rate less than the classification threshold range as a downward trend; Use the threshold segmentation weighting method to set the classification weight, use the weighted integration method to calculate the risk index, use the threshold screening method to set the risk threshold, and screen the risk index greater than the risk threshold to generate a high-risk area.

6. The air pollution trend analysis method according to claim 5, characterized in that: The construction of the visualization interface to display the monitoring results includes: Use the Dash framework of the visualization tool Python to construct the visualization interface, use the data visualization tool to display the future trend of pollutants in real time, and use the map visualization tool to display the high-risk area; Allow users who have passed real-name verification to view.

7. The air pollution trend analysis method according to claim 6, characterized in that: The storage of the multi-source data generated by collection and analysis includes: Store the collected multi-source data and the generated high-risk areas in the central database, and set security access measures. The central database backs up the stored data to the cloud, and regularly conducts integrity detection on the stored data and the backup data. After the detection is completed, an integrity detection record is generated and synchronously stored in the central database.

8. An air pollution trend analysis system, based on the air pollution trend analysis method according to any one of claims 1 to 7, characterized in that: Including, The collection and optimization module is used to collect multi-source data in the target area, divide the target area into grids using spatial gridding to generate a directed graph, use the law of conservation of mass to calculate the concentration change rate of all nodes, update the concentration of nodes using the Euler method, input the updated concentration of nodes into a pre-trained GCN model to calculate the optimized concentration, and generate an exception list; The modulation joint module is used to modulate the propagation weights using the Hebbian learning rule, modulate the optimized concentration using the modulated weights, calculate the joint concentration using the conservative path flux method, combine the joint concentration using weighted linear combination, and calculate the joint concentration; The construction monitoring module is used to construct a long short-term memory network model to predict the future trend of pollutants and classify and monitor the future trend of pollutants; The visualization storage module is used to construct a visualization interface to display the monitoring results and store the multi-source data generated by collection and analysis.

9. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that: When the processor executes the computer program, the steps of the air pollution trend analysis method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, the steps of the air pollution trend analysis method according to any one of claims 1 to 7 are implemented.

Images