Urban air pollution monitoring management system and method based on LSTM model

Abstract

The application discloses a city air pollution monitoring management system and method based on an LSTM model, relates to the technical field of air pollution monitoring, and discloses an air pollution monitoring module, an air pollution monitoring point analysis module, a city air pollution comprehensive module, and an air pollution type classification module. The air pollution monitoring module and the air pollution monitoring point analysis module are arranged, the collected data of various air pollution monitoring points in the city are analyzed in depth, the subsequent air pollution situation is accurately predicted by using an LSTM model, the optical reaction effect of various types of city air pollution data is predicted, the pollution fluctuation of the air pollution monitoring points is combined, the air pollution situation of the air pollution monitoring points is comprehensively judged, the air pollution monitoring points with significant pollution are marked, the city air pollution comprehensive module and the air pollution type classification module are arranged, the correlation of pollution regions between the significant pollution points is further analyzed, and then the air pollution situation of the city is comprehensively analyzed.

Description

Technical Field

[0001] This invention relates to the field of air pollution monitoring technology, and more specifically, to an urban air pollution monitoring and management system and method based on an LSTM model. Background Technology

[0002] With the acceleration of urbanization and rapid industrialization, urban air pollution has become increasingly severe, posing a serious threat to people's health and quality of life. To effectively address this issue, environmental protection departments need to establish an efficient and accurate air pollution monitoring and management system to promptly and comprehensively grasp the urban air pollution situation and formulate and implement corresponding control measures accordingly.

[0003] In recent years, air pollution monitoring methods based on big data and machine learning technologies have gradually emerged. Among them, the LSTM (Long Short-Term Memory) model, as a special type of recurrent neural network (RNN), has been widely used in the field of air pollution prediction due to its ability to process long-sequence data and capture potential patterns in time-series data.

[0004] However, despite the excellent performance of LSTM models in air pollution prediction, existing LSTM-based air pollution monitoring and management systems still have some shortcomings. For example, these systems lack in-depth analysis of air pollution monitoring data, cannot analyze the correlation between data and regions, and can only provide predictions for specific data, making it difficult to comprehensively analyze the city's air pollution situation.

[0005] Therefore, this invention proposes an urban air pollution monitoring and management system and method based on an LSTM model. Summary of the Invention

[0006] In view of the shortcomings of the existing technology, the purpose of this invention is to provide an urban air pollution monitoring and management system and method based on LSTM model.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] The urban air pollution monitoring and management system based on the LSTM model includes an air pollution monitoring module, an air pollution monitoring point analysis module, an urban air pollution comprehensive module, and an air pollution type classification module.

[0009] The air pollution monitoring module is used to identify all air pollution monitoring points in the city, and the air pollution monitoring points collect air pollution data of various types of cities in real time.

[0010] The air pollution monitoring point analysis module acquires the comprehensive air pollution value of each air pollution monitoring point after each monitoring cycle. Based on the comparison between the comprehensive air pollution value and the comprehensive air pollution threshold, it determines whether to mark the air pollution monitoring point as a significant pollution point.

[0011] The urban air pollution integrated module is used to acquire urban air pollution monitoring values;

[0012] The air pollution type classification module determines the air pollution type of a city based on the comparison results of air pollution monitoring values ​​with high and low air pollution monitoring values.

[0013] Furthermore, the method for obtaining the comprehensive air pollution measurement values ​​at air pollution monitoring stations is as follows: Obtain the predicted air pollution value Zstg and the air pollution volatility value Aekb for the monitoring period at each monitoring station, and then use the formula... The comprehensive air pollution value Ezvm of the air pollution monitoring station is obtained, where q1 is the air pollution prediction coefficient and q2 is the air pollution fluctuation coefficient.

[0014] Furthermore, the method for obtaining air pollution prediction values ​​for the monitoring period at air pollution monitoring stations is as follows: Obtain the phased actual sampling feature sets of various types of air pollution data within the air pollution monitoring stations; obtain the actual sampling feature prediction models corresponding to each type of air pollution data; input the phased actual sampling feature sets of each type of urban air pollution data into the corresponding actual sampling feature prediction models; output the actual sampling prediction features of each type of urban air pollution data; obtain the pollution feature analysis models corresponding to each type of urban air pollution data; input the actual sampling prediction features of each type of urban air pollution data into the corresponding pollution feature analysis models. The model obtains the pollution prediction values ​​of air pollution data for each type of city. The pollution prediction values ​​of all types of city air pollution data are summed and averaged to obtain the comprehensive prediction value Lcdy. A photochemical relationship diagram is obtained, and the photochemical prediction values ​​Ted of air pollution data for two types of cities with a photochemical relationship are obtained from the diagram, where e = 1, 2, ..., E, e is the number of the photochemical prediction value, and E is the total number of photochemical prediction values. The photochemical prediction coefficient is set as fd, d = 1, 2, 3, ..., d-1, d, f1 < f2 < f3 < ... < fd-1 < fd. The formula is then used... The predicted air pollution values ​​Zstg for the monitoring period are obtained from the air pollution monitoring stations, where ha is the comprehensive prediction coefficient.

[0015] Furthermore, the method for obtaining the phased actual collection feature set of air pollution data is as follows: obtain air pollution data of the same type of city for i consecutive monitoring periods before the current time from the air pollution monitoring point, perform data processing and feature extraction on the i air pollution data of the same type of city to obtain the data features of the i air pollution data of the same type of city, and combine the data features of the i air pollution data of the same type of city to form the phased actual collection feature set of air pollution data of this type.

[0016] Furthermore, the photochemical prediction values ​​for air pollution data from two types of cities exhibiting a photochemical relationship in the photochemical relationship diagram are obtained as follows: The pollution prediction values ​​for the two types of cities exhibiting a photochemical relationship in the photochemical relationship diagram are summed and averaged to obtain the pollution synergy value Kth. The difference between the pollution prediction values ​​of the two types of cities is calculated and the absolute value is taken to obtain the prediction difference value Bds. The prediction difference is then calculated using the formula... The photochemical prediction values ​​Ted of air pollution data for two types of cities were obtained, where g1 is the pollution synergy coefficient and g2 is the prediction gap coefficient.

[0017] Furthermore, the method for obtaining the air pollution fluctuation value of an air pollution monitoring station for a monitoring period is as follows: Obtain the air pollution prediction values ​​for j consecutive monitoring periods prior to the current time at the air pollution monitoring station; perform pairwise matching of all monitoring periods; calculate the difference between the air pollution prediction values ​​of the two matched monitoring periods and take the absolute value to obtain the matched fluctuation value; sum all matched fluctuation values ​​and take the average value to obtain the matched fluctuation mean BMk; sort all air pollution prediction values ​​in chronological order according to the monitoring period; calculate the difference between two adjacent air pollution prediction values ​​after sorting and take the absolute value to obtain the adjacent fluctuation value; sum all adjacent fluctuation values ​​and take the average value to obtain the adjacent fluctuation mean Rsw; use the formula Aekb=BMk*w1+Rsw*w2 to obtain the air pollution fluctuation value Aekb of the air pollution monitoring station for a monitoring period, where w1 is the matched fluctuation coefficient and w2 is the adjacent fluctuation coefficient.

[0018] Furthermore, the method for obtaining urban air pollution monitoring values ​​is as follows: The total number of significant pollution points is labeled Ew. All significant pollution points are paired into groups, and the pollution monitoring synergy value for each group is obtained. The pollution monitoring synergy values ​​of all groups are summed and averaged to obtain the pollution monitoring synergy mean value Hqa. The formula Citd = (Hqa + 1.15) is then used. Ew+1.48 The city's air pollution monitoring value, Citd, was obtained.

[0019] Furthermore, the method for obtaining the pollution monitoring synergy value of the significant point group is as follows: The distance difference between two significant pollution points in the significant point group is calculated to obtain the pollution monitoring distance Ls; the difference between the comprehensive air pollution monitoring values ​​of two significant pollution points in the significant point group is calculated to obtain the significant distance value Xt; and the value is obtained using the formula... The pollution monitoring synergy value Fru for this significant point group was obtained.

[0020] Furthermore, based on the comparison results of air pollution monitoring values ​​with high and low air pollution monitoring values, the air pollution type of a city is determined. Specifically, high and low air pollution monitoring values ​​are set. When the air pollution monitoring value of a city is greater than or equal to the high air pollution monitoring value, the city's air pollution type is marked as severe air pollution. When the air pollution monitoring value of a city is less than or equal to the low air pollution monitoring value, the city's air pollution type is marked as no air pollution. When the air pollution monitoring value of a city is between the high and low air pollution monitoring values, the city's air pollution type is marked as normal air pollution.

[0021] Furthermore, the urban air pollution monitoring and management method based on the LSTM model includes the following steps:

[0022] Step 1: Identify all air pollution monitoring stations in the city, and have these stations collect real-time air pollution data for various types of urban areas;

[0023] Step 2: After each monitoring cycle, obtain the comprehensive air pollution measurement values ​​for each air pollution monitoring station;

[0024] Step 3: Determine whether to mark the air pollution monitoring station as a significant pollution point;

[0025] Step 4: Obtain air pollution monitoring values ​​for the city;

[0026] Step 5: Determine the type of air pollution in the city.

[0027] Compared with the prior art, the present invention has the following beneficial effects:

[0028] 1. The method of the present invention uses an LSTM model to accurately predict subsequent air pollution conditions and predict the optical response effect of air pollution data of various types of cities. Combined with the pollution fluctuation of air pollution monitoring points, the air pollution conditions of air pollution monitoring points are comprehensively judged, and the air pollution types of cities are accurately classified according to the air pollution conditions.

[0029] 2. An air pollution monitoring module and an air pollution monitoring point analysis module are set up to conduct in-depth analysis of the data collected from various air pollution monitoring points in the city, and to mark the air pollution monitoring points with significant pollution. An urban air pollution comprehensive module and an air pollution type classification module are set up to further analyze the correlation between pollution areas between significant pollution points, and then comprehensively analyze the air pollution situation in the city, so as to facilitate the environmental protection department to take corresponding control measures according to the air pollution type in the city. Attached Figure Description

[0030] Figure 1 This is a flowchart illustrating the urban air pollution monitoring and management method based on the LSTM model.

[0031] Figure 2 A flowchart for determining the types of air pollution in a city;

[0032] Figure 3 This is a system module diagram of an urban air pollution monitoring and management system based on an LSTM model. Detailed Implementation

[0033] Example 1: Refer to Figure 1 The urban air pollution monitoring and management method based on the LSTM model includes the following steps:

[0034] Step 1: Identify all air pollution monitoring stations in the city, and have these stations collect real-time air pollution data for various types of urban areas;

[0035] Step 2: After each monitoring cycle, obtain the comprehensive air pollution measurement values ​​for each air pollution monitoring station;

[0036] Step 3: Determine whether to mark the air pollution monitoring station as a significant pollution point;

[0037] Step 4: Obtain air pollution monitoring values ​​for the city;

[0038] Step 5: Determine the type of air pollution in the city.

[0039] Example 2: Refer to Figures 2-3 The urban air pollution monitoring and management system based on the LSTM model includes an air pollution monitoring module, an air pollution monitoring point analysis module, an urban air pollution comprehensive module, and an air pollution type classification module.

[0040] Air pollution monitoring module: Identifies all air pollution monitoring points in the city (multiple air pollution monitoring points are usually set up in a city to monitor air pollution). The air pollution monitoring points collect various types of urban air pollution data in real time (the types of air pollution data include sulfur dioxide data, nitrogen dioxide data, carbon monoxide data, methane data, particulate matter data, etc.).

[0041] Air pollution monitoring point analysis module: Set the monitoring cycle (the length of the monitoring cycle is determined according to data requirements, and this application does not limit it). After each monitoring cycle, obtain the comprehensive air pollution value of each air pollution monitoring point, and set the comprehensive air pollution threshold (the comprehensive air pollution threshold is a system preset value). When the comprehensive air pollution value of an air pollution monitoring point is greater than or equal to the comprehensive air pollution threshold, the air pollution monitoring point is marked as a pollution-significant point. When the comprehensive air pollution value of an air pollution monitoring point is less than the comprehensive air pollution threshold, no further processing is performed.

[0042] The method for obtaining comprehensive air pollution values ​​from air pollution monitoring stations is as follows: Obtain the predicted air pollution value Zstg and the air pollution volatility value Aekb for the monitoring period, and then use the formula... The comprehensive air pollution value Ezvm of the air pollution monitoring station is obtained, where q1 is the air pollution prediction coefficient and q2 is the air pollution fluctuation coefficient, with q1 taking the value of 1.58 and q2 taking the value of 1.34.

[0043] The method for obtaining air pollution prediction values ​​for a monitoring period at air pollution monitoring stations is as follows: First, obtain the stage-specific actual sampling feature sets of various types of air pollution data within the monitoring station. Second, obtain the actual sampling feature prediction models corresponding to each type of air pollution data. Third, input the stage-specific actual sampling feature sets of each type of urban air pollution data into the corresponding actual sampling feature prediction models to output the actual sampling prediction features of each type of urban air pollution data. Fourth, obtain the pollution feature analysis models corresponding to each type of urban air pollution data. Fifth, input the actual sampling prediction features of each type of urban air pollution data into the corresponding pollution feature analysis models to obtain the pollution prediction values ​​of each type of urban air pollution data. Sixth, sum the pollution prediction values ​​of all types of urban air pollution data and take the average. The value is used to obtain the comprehensive predicted value Lcdy. A photochemical relationship diagram is obtained, and the photochemical predicted values ​​Ted of all two types of urban air pollution data with photochemical relationships are obtained from the diagram, where e = 1, 2, ..., E, e is the number of the photochemical predicted value, and E is the total number of photochemical predicted values. The photochemical prediction coefficient is set as fd, where d = 1, 2, 3, ..., d-1, d, f1 < f2 < f3 < ... < fd-1 < fd. Each photochemical prediction coefficient corresponds to a range of photochemical predicted values, including (0, Te1], (Te1, Te2], ..., (Ted-1, Ted]. When Ted ∈ (Te1, Te2], the photochemical prediction coefficient is f2. The formula is then used... The predicted air pollution value Zstg for the monitoring period is obtained from the air pollution monitoring point, where ha is the comprehensive prediction coefficient and the value of ha is 0.81.

[0044] The photochemical prediction values ​​for air pollution data from two types of cities exhibiting a photochemical relationship in the photochemical relationship diagram are obtained as follows: The pollution prediction values ​​for the two types of cities exhibiting a photochemical relationship in the diagram are summed and averaged to obtain the pollution synergy value Kth. The difference between the pollution prediction values ​​of the two types of cities is calculated, and the absolute value is taken to obtain the prediction difference value Bds. The formula is then used to... The photochemical prediction values ​​Ted of air pollution data for two types of cities were obtained, where g1 is the pollution synergy coefficient and g2 is the prediction gap coefficient, with g1 having a value of 0.65 and g2 having a value of 1.31.

[0045] The photochemical relationship diagram contains air pollution data for all types of cities and shows the photochemical relationships between different types of air pollution data. For example, sulfur dioxide data and nitrogen dioxide data can undergo photochemical reactions in the atmosphere. These reactions can further pollute the air and may cause more serious environmental problems. Therefore, sulfur dioxide data and nitrogen dioxide data have a photochemical relationship in the photochemical relationship diagram.

[0046] The method for obtaining the phased actual collection feature set of air pollution data is as follows: Obtain air pollution data of the same type of city for i consecutive monitoring periods before the current time (such as collecting sulfur dioxide data for i consecutive monitoring periods), perform data processing and feature extraction on the i air pollution data of the same type of city (data processing methods include data cleaning, etc.), obtain the data features of the i air pollution data of the same type of city, and combine the data features of the i air pollution data of the same type of city into the phased actual collection feature set of air pollution data of this type.

[0047] Each type of urban air pollution data corresponds to a real-time sampling feature prediction model. All real-time sampling feature prediction models are built based on LSTM models. The only difference between different real-time sampling feature prediction models is the different training data. In the embodiment, taking sulfur dioxide data as an example, the construction process of the real-time sampling feature prediction model for sulfur dioxide data is disclosed: collect a stage real-time sampling feature sets of sulfur dioxide data, build an LSTM model, use the stage real-time sampling feature sets of sulfur dioxide data as training data for the LSTM model, assign a stage real-time sampling feature set to each training data, and the stage real-time sampling feature set is the data feature for predicting sulfur dioxide data in the next monitoring cycle. Divide the training data into a training set, a validation set, and a test set according to a set ratio of 4:1:1, train the training set, validation set, and test set, and complete the training to build the real-time sampling feature prediction model for sulfur dioxide data.

[0048] If it is a prediction model for the actual sampling characteristics of nitrogen dioxide data, then the stage actual sampling characteristic set of a nitrogen dioxide data is collected, and the stage actual sampling characteristic set is the data characteristic for predicting the nitrogen dioxide data of the next monitoring cycle.

[0049] Each type of urban air pollution data corresponds to a pollution feature analysis model. All actual sampling feature prediction models are built based on deep learning models. The only difference between different pollution feature analysis models is the different training data. In this embodiment, taking sulfur dioxide data as an example, the construction process of the pollution feature analysis model for sulfur dioxide data is disclosed: collect actual sampling prediction features of b sulfur dioxide data, build a deep learning model, use the actual sampling prediction features of sulfur dioxide data as training data for the deep learning model, assign a pollution prediction value to each training data, the value range of the pollution prediction value is (3.0~10.0), the closer the pollution prediction value is to 10.0, the more serious the sulfur dioxide pollution in the next monitoring period, the closer the pollution prediction value is to 3.0, the less serious the sulfur dioxide pollution in the next monitoring period, divide the training data into training set and validation set according to a set ratio of 3:1, train the training set and validation set, and after training is completed, the pollution feature analysis model for sulfur dioxide data is constructed.

[0050] If it is a pollution characteristic analysis model for nitrogen dioxide data, then the actual predicted characteristics of b nitrogen dioxide data are collected. The closer the pollution prediction value is to 10.0, the more serious the nitrogen dioxide pollution will be in the next monitoring period. The closer the pollution prediction value is to 3.0, the less serious the nitrogen dioxide pollution will be in the next monitoring period.

[0051] The method for obtaining the air pollution fluctuation value of an air pollution monitoring station for a monitoring period is as follows: Obtain the air pollution prediction values ​​for j consecutive monitoring periods prior to the current time at the air pollution monitoring station. Perform pairwise matching on all monitoring periods. Calculate the difference between the air pollution prediction values ​​of the two matched monitoring periods and take the absolute value to obtain the matched fluctuation value. Sum all matched fluctuation values ​​and take the average to obtain the matched fluctuation mean BMk. Sort all air pollution prediction values ​​in chronological order according to the monitoring period. Calculate the difference between two adjacent air pollution prediction values ​​after sorting and take the absolute value to obtain the adjacent fluctuation value. Sum all adjacent fluctuation values ​​and take the average to obtain the adjacent fluctuation mean Rsw. Use the formula Aekb=BMk*w1+Rsw*w2 to obtain the air pollution fluctuation value Aekb of the air pollution monitoring station for the monitoring period, where w1 is the matched fluctuation coefficient and w2 is the adjacent fluctuation coefficient, with w1 taking the value of 0.74 and w2 taking the value of 0.68.

[0052] An air pollution monitoring module and an air pollution monitoring point analysis module are set up to conduct in-depth analysis of the data collected from various air pollution monitoring points in the city. The LSTM model is used to accurately predict the subsequent air pollution situation and predict the optical response effect of air pollution data of various types of cities. Combined with the pollution fluctuation of air pollution monitoring points, the air pollution situation of air pollution monitoring points is comprehensively judged, and air pollution monitoring points with significant pollution are marked.

[0053] Urban air pollution comprehensive module: The total number of pollution salients is marked as Ew. All pollution salients are paired into a salient group. The pollution monitoring synergy value of each salient group is obtained. The pollution monitoring synergy values ​​of all salient group groups are summed and averaged to obtain the pollution monitoring synergy mean Hqa. The formula Citd = (Hqa + 1.15) is used. Ew+1.48 The city's air pollution monitoring value, Citd, was obtained.

[0054] The method for obtaining the pollution monitoring synergy value of a significant point group is as follows: Calculate the distance difference between two significant pollution points in the group to obtain the pollution monitoring distance Ls; calculate the difference between the comprehensive air pollution monitoring values ​​of two significant pollution points in the group to obtain the significant distance value Xt; and then use the formula... The pollution monitoring synergy value Fru for this significant point group was obtained.

[0055] Air pollution type classification module: Set high and low values ​​for air pollution monitoring (both are system preset values, with the high value being greater than the low value). When the city's air pollution monitoring value is greater than or equal to the high value, the city's air pollution type is marked as severe air pollution. When the city's air pollution monitoring value is less than or equal to the low value, the city's air pollution type is marked as no air pollution. When the city's air pollution monitoring value is between the high and low values, the city's air pollution type is marked as normal air pollution.

[0056] The system includes a comprehensive urban air pollution module and an air pollution type classification module. This module further analyzes the correlation between pollution areas and significant pollution points, thereby comprehensively analyzing the city's air pollution situation. Based on the air pollution situation, the system accurately classifies the city's air pollution types, making it easier for environmental protection departments to take corresponding control measures according to the type of air pollution in the city.

[0057] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0058] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0059] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0060] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0061] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0062] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0063] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0064] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A city air pollution monitoring and management system based on an LSTM model, characterized in that, It includes an air pollution monitoring module, an air pollution monitoring point analysis module, an urban air pollution comprehensive module, and an air pollution type classification module; The air pollution monitoring module is used to identify all air pollution monitoring points in the city, and the air pollution monitoring points collect air pollution data of various types of cities in real time. The air pollution monitoring point analysis module acquires the comprehensive air pollution value of each air pollution monitoring point after each monitoring cycle. Based on the comparison between the comprehensive air pollution value and the comprehensive air pollution threshold, it determines whether to mark the air pollution monitoring point as a significant pollution point. The method for obtaining comprehensive air pollution values ​​from air pollution monitoring stations is as follows: Obtain the predicted air pollution value Zstg and the air pollution volatility value Aekb for the monitoring period, and then use the formula... The comprehensive air pollution value Ezvm of the air pollution monitoring station is obtained, where q1 is the air pollution prediction coefficient and q2 is the air pollution wave coefficient. The method for obtaining air pollution fluctuation values ​​for a monitoring period at an air pollution monitoring station is as follows: Obtain the predicted air pollution values ​​for j consecutive monitoring periods prior to the current time at the monitoring station. Perform pairwise matching on all monitoring periods. Calculate the difference between the predicted air pollution values ​​of the two matched monitoring periods and take the absolute value to obtain the matched fluctuation value. Sum all matched fluctuation values ​​and take the average to obtain the matched fluctuation mean BMk. Sort all air pollution prediction values ​​sequentially according to the chronological order of the monitoring periods. Calculate the difference between two adjacent air pollution prediction values ​​after sorting and take the absolute value to obtain the adjacent fluctuation value. Sum all adjacent fluctuation values ​​and take the average to obtain the adjacent fluctuation mean Rsw. Use the formula... The air pollution fluctuation value Aekb of the air pollution monitoring point for the monitoring period is obtained, where w1 is the matching fluctuation coefficient and w2 is the adjacent fluctuation coefficient. The method for obtaining air pollution prediction values ​​for a monitoring period at air pollution monitoring stations is as follows: Obtain the stage-specific actual sampling feature sets of various types of air pollution data within the monitoring stations; obtain the actual sampling feature prediction models corresponding to each type of air pollution data; input the stage-specific actual sampling feature sets of each type of urban air pollution data into the corresponding actual sampling feature prediction models; output the actual sampling prediction features of each type of urban air pollution data; obtain the pollution feature analysis models corresponding to each type of urban air pollution data; input the actual sampling prediction features of each type of urban air pollution data into the corresponding pollution feature analysis models; and obtain... Pollution prediction values ​​for air pollution data from various city types are obtained. These prediction values ​​are summed and averaged to obtain a comprehensive prediction value, Lcdy. A photochemical relationship diagram is then generated, and the photochemical prediction values ​​Ted for all two city types with a photochemical relationship are retrieved from the diagram. e = 1, 2, ..., E, where e is the photochemical prediction value number and E is the total number of photochemical prediction values. The photochemical prediction coefficient is set as fd, where d = 1, 2, 3, ..., d-1, d, f1 < f2 < f3 < ... < fd-1 < fd. The formula is then used... The predicted air pollution value Zstg for the monitoring period is obtained from the air pollution monitoring point, where ha is the comprehensive prediction coefficient; The photochemical prediction values ​​for air pollution data from two types of cities exhibiting a photochemical relationship in the photochemical relationship diagram are obtained as follows: The pollution prediction values ​​for the two types of cities exhibiting a photochemical relationship in the diagram are summed and averaged to obtain the pollution synergy value Kth. The difference between the pollution prediction values ​​of the two types of cities is calculated, and the absolute value is taken to obtain the prediction difference value Bds. The formula is then used to... The photochemical prediction values ​​Ted of two types of urban air pollution data were obtained, where g1 is the pollution synergy coefficient and g2 is the prediction gap coefficient; The urban air pollution integrated module is used to acquire urban air pollution monitoring values; The method for obtaining urban air pollution monitoring values ​​is as follows: The total number of significant pollution points is labeled Ew. All significant pollution points are paired into groups, and the pollution monitoring synergy value for each group is obtained. The pollution monitoring synergy values ​​of all groups are summed and averaged to obtain the pollution monitoring synergy mean value Hqa. This is then calculated using the formula... The city's air pollution monitoring value, Citd, was obtained. The method for obtaining the pollution monitoring synergy value of a significant point group is as follows: Calculate the distance difference between two significant pollution points in the group to obtain the pollution monitoring distance Ls; calculate the difference between the comprehensive air pollution monitoring values ​​of two significant pollution points in the group to obtain the significant distance value Xt; and then use the formula... The pollution monitoring synergy value Fru for this significant point group was obtained; The air pollution type classification module determines the air pollution type of a city based on the comparison results of air pollution monitoring values ​​with high and low air pollution monitoring values.

2. The urban air pollution monitoring and management system based on the LSTM model according to claim 1, characterized in that, The method for obtaining the phased real-time feature set of air pollution data is as follows: obtain air pollution data of the same type of city for i consecutive monitoring periods before the current time from the air pollution monitoring point, perform data processing and feature extraction on the i air pollution data of the same type of city to obtain the data features of the i air pollution data of the same type of city, and combine the data features of the i air pollution data of the same type of city to form the phased real-time feature set of air pollution data of this type.

3. The urban air pollution monitoring and management system based on the LSTM model according to claim 1, characterized in that, Based on the comparison results of air pollution monitoring values ​​with high and low air pollution monitoring values, the air pollution type of a city is determined. Specifically, high and low air pollution monitoring values ​​are set. When the air pollution monitoring value of a city is greater than or equal to the high air pollution monitoring value, the city's air pollution type is marked as severe air pollution. When the air pollution monitoring value of a city is less than or equal to the low air pollution monitoring value, the city's air pollution type is marked as no air pollution. When the air pollution monitoring value of a city is between the high and low air pollution monitoring values, the city's air pollution type is marked as normal air pollution.

4. A method for urban air pollution monitoring and management based on an LSTM model, applied to the urban air pollution monitoring and management system based on an LSTM model as described in any one of claims 1-3, characterized in that, Includes the following steps: Step 1: Identify all air pollution monitoring stations in the city, and have these stations collect real-time air pollution data for various types of urban areas; Step 2: After each monitoring cycle, obtain the comprehensive air pollution measurement values ​​for each air pollution monitoring station; Step 3: Determine whether to mark the air pollution monitoring station as a significant pollution point; Step 4: Obtain air pollution monitoring values ​​for the city; Step 5: Determine the type of air pollution in the city.

Images