Intelligent detection system and method for atmospheric pollution source

Abstract

The invention discloses an atmospheric pollution source intelligent detection system and method, and relates to the technical field of atmospheric pollution detection, and the method comprises the steps: 1, collecting atmospheric pollutant concentration data, space-time position information and meteorological conditions at different time, space and altitude levels; step 2, constructing a pollutant diffusion path prediction model based on the collected pollutant concentration data, space-time position information and meteorological conditions, outputting a pollutant diffusion path prediction value, and predicting the pollutant diffusion path prediction value at a to-be-predicted point; and step 3, constructing a three-dimensional pollution distribution model based on the atmospheric pollutant concentration data, the space-time position information, the meteorological conditions and the derived spatial distance parameters, outputting a pollution concentration prediction value, and predicting the pollution concentration prediction value at the spatial point. According to the method, the risk level of the pollution source is accurately divided, and the problem that the real-time traceability and the classification response capability of the pollution event are limited due to the lack of a systematic data modeling mechanism and a response strategy in the existing method is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of air pollution detection, and in particular to an intelligent detection system and method for air pollution sources. Background Art

[0002] At present, air pollution monitoring mainly relies on fixed monitoring stations or mobile monitoring vehicles to obtain data, which makes it difficult to cover air pollution conditions over a large area and at different altitudes. This leads to insufficient accuracy in locating pollution sources and predicting diffusion trends, making it difficult to accurately identify and quickly respond to pollution sources. Therefore, an intelligent detection system and method for air pollution sources is urgently needed.

[0003] Existing atmospheric pollution detection technologies have limitations in spatial data fusion, pollution diffusion path prediction, and pollution source inversion analysis. They make it difficult to accurately identify the type, location, and risk level of pollution sources. The lack of systematic data modeling mechanisms and response strategies limits the ability to trace pollution incidents in real time and respond to them in a classified manner. Summary of the Invention

[0004] The purpose of this section is to summarize some aspects of the embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the abstract and title of this application to avoid obscuring the purpose of this section, the abstract and the title of the invention, and such simplifications or omissions should not be used to limit the scope of the present invention.

[0005] In view of the above-mentioned problems existing in the existing intelligent detection system and method for air pollution sources, the present invention is proposed.

[0006] Therefore, the purpose of the present invention is to provide an intelligent detection system and method for atmospheric pollution sources, which is suitable for solving the problem that the existing technology lacks a systematic data modeling mechanism and response strategy, which limits the real-time tracing and classification response capabilities of pollution incidents.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] In a first aspect, an embodiment of the present invention provides an intelligent air pollution source detection system and method, characterized by comprising:

[0009] Step 1: Collect atmospheric pollutant concentration data, spatiotemporal location information, and meteorological conditions at different time, space, and altitude levels;

[0010] Step 2: Based on the collected pollutant concentration data, spatiotemporal location information and meteorological conditions, a pollutant diffusion path prediction model is constructed to output the pollutant diffusion path prediction value and predict the pollutant diffusion path prediction value at the prediction point;

[0011] Step 3: Based on the atmospheric pollutant concentration data, spatiotemporal location information, meteorological conditions and derived spatial distance parameters, a three-dimensional pollution distribution model is constructed to output the pollution concentration prediction value and the pollution concentration prediction value at the predicted spatial point;

[0012] Step 4: Combine the pollutant diffusion prediction value and the pollution distribution prediction value to build a pollution source inversion mechanism, output the pollution source probability level score, and determine the pollution source probability level.

[0013] As a preferred solution of the intelligent detection method for air pollution sources described in the present invention, the pollutant diffusion path prediction model is constructed as follows:

[0014] S201: The pollutant concentration data collected in step 1 and the corresponding three-dimensional space coordinates (x j ,y j , z j ), acquisition time t j , release rate per unit time Δψ j , wind speed and wind direction and other environmental factors are uniformly coded to construct a pollution source unit data set;

[0015] Quantitatively calculate the mass Q of pollutants diffused outward from each source point per unit time j , the announcement is as follows:

[0016]

[0017] The extraction of environmental correction factors ensures that pollution source units not only have the original pollution intensity but also dynamically respond to the time-varying characteristics of the atmosphere and urban structure, ensuring the physical rationality and dynamic accuracy of the pollutant diffusion path prediction model;

[0018] Finally, each pollution source unit data item is a six-tuple:

[0019]

[0020] in, is the wind speed vector at the pollution source, which is used for directional modulation of the subsequent pollutant diffusion path prediction model, η j The building blocking coefficient is calculated by analyzing the building density near the pollution source through GIS or BIM data to reflect the degree of inhibition of the upward / lateral spread of pollutants;

[0021] S202: Based on the environmental wind field simulation parameters, terrain characteristics, and the spatiotemporal changes of pollution sources, a spatial attenuation factor λ is introduced for each pollution source unit. j , diffusion weight coefficient k j and source release mass Q j ;

[0022] S203: Preset the three-dimensional space coordinates (x i ,y i , z i );

[0023] S204: Based on the pollution source unit data set, spatial attenuation factor, diffusion weight coefficient and the coordinates of the predicted point, a pollutant diffusion path prediction model is constructed, and the predicted value P of the pollutant diffusion path at the predicted point is output. i , which is expressed as follows:

[0024]

[0025] Among them, P i Represents the predicted value of the pollutant diffusion path at the prediction point i.

[0026] As a preferred solution of the intelligent detection method for air pollution sources described in the present invention, the process of constructing the three-dimensional pollution distribution model is as follows:

[0027] S301: Integrate pollution concentration data from ground monitoring stations and drones C ij , containing its corresponding observation coordinates (x j ,y j , z j ), acquisition time t j and environment variable h ij ;

[0028] S302: Calculate the square of the spatial distance between the target estimation point i and all observation points j Constructing the high impact term 1+α j ·log(1+h ij );

[0029] S303: Introducing weight factor ω j and the spatial attenuation factor b j ;

[0030] S304: Based on the pollution concentration data of the observation point, the square of the spatial distance, the height influence term, the weight factor and the spatial attenuation factor, a three-dimensional pollution distribution model is constructed to output the pollution concentration prediction value D at the prediction point. i , which is expressed as follows:

[0031]

[0032] Among them, D i Represents the predicted value of pollution concentration at spatial point i.

[0033] As a preferred solution of the intelligent detection method for air pollution sources described in the present invention, the pollution source inversion mechanism construction process is as follows:

[0034] S401: Define a set of candidate pollution source locations and calculate the probability level score of each candidate source point as a true pollution source;

[0035] S402: Define variables: S = {s1, s1, ..., s n}, represents the set of all candidate pollution source locations, P i is the predicted value of the pollutant diffusion path at the prediction point i, Indicates s i The diffusion path simulated when it is the source point, D i The predicted value of pollution concentration at spatial point i, Indicates s i The predicted value of pollution concentration when it is a source point, γ and δ represent the weighting coefficients of path score and concentration score, respectively;

[0036] S403: Construct a pollution source probability scoring model and output the pollution source probability level score, which is expressed as follows:

[0037]

[0038] in, Indicates the probability level score of the pollution source, H P (s i ) represents the path inconsistency, H D (s i ) Pollution concentration residual, Z is the normalization factor.

[0039] As a preferred solution of the intelligent detection method for air pollution sources described in the present invention, the probability values of all candidate source points are According to their size, they are divided into multiple level intervals, and the multiple level intervals are as follows:

[0040] Identify the source of pollution (red alert);

[0041] Possible pollution source (orange alert);

[0042] Suspected pollution source (yellow reminder);

[0043] Non-pollution sources (green areas).

[0044] As a preferred solution of the intelligent detection method of air pollution sources described in the present invention, wherein: the pollution source probability level score output by the pollution source probability scoring model is Set the three thresholds as critical threshold T H , risk threshold T M and safety threshold T L ;

[0045] like If the candidate source point falls into the red warning area, the system will execute the first-level response strategy and make a first-level judgment;

[0046] When the candidate source point falls into the red warning area, the diffusion path deviation ΔP of all monitoring points within the radius R1 with the candidate source point as the center is extracted. i , and set the path deviation low threshold δ low and path deviation high threshold δ high ;

[0047] If more than M1% of the monitoring points meet ΔP i <δ low , then the candidate source point is determined to be a concentrated pollution source;

[0048] If more than M2% of the monitoring points meet δ low ≤ΔP i <δ high , then the candidate source point is determined to be a diffuse pollution source;

[0049] Otherwise, the candidate source point is determined to be an abnormal pollution source;

[0050] like If the candidate source point falls into the orange warning area, the system will execute the secondary response strategy and make a secondary judgment;

[0051] When the candidate source point falls into the orange warning area, the time-latitude stability index η is calculated and the volatility high threshold η is set high and volatility low threshold η low ;

[0052] If η>η high , then the candidate source point is determined to be a volatility risk source;

[0053] If η low <η≤η high , then the candidate source point is determined to be a gradual risk source;

[0054] If η≤η low , then the candidate source point is determined to be a stable risk source;

[0055] like If the candidate source point falls into the yellow warning area, the system will execute the three-level response strategy and make three-level judgments;

[0056] When the candidate source point falls into the yellow warning area, the spatial correlation index ρ is calculated and the first spatial correlation threshold ρ is set high , the second spatial correlation threshold ρ low and volatility warning value η y ;

[0057] If ρ>ρ high , then the candidate source point is determined to be a related suspicious source, and the next step is executed;

[0058] When the candidate source point is determined to be a correlated suspicious source, if η>η y , then the candidate source point is determined to fall into the orange warning area;

[0059] If ρ low <ρ≤ρ high , then the candidate source point is determined to be an occasional suspicious source; then proceed to the next step;

[0060] When the candidate source point is determined to be an occasional suspicious source, if the diffusion path deviation ΔP i >δ high , then the candidate source point is determined to fall into the red warning area;

[0061] If ρ≤ρ low , then the candidate source point is determined to be an isolated suspicious source;

[0062] like This indicates that the candidate source point falls into the green area, and the area is marked as the background reference area. The drone path is adjusted to avoid this area.

[0063] As a preferred solution of the intelligent detection method for air pollution sources described in the present invention, the calculation process of the spatial correlation index is as follows:

[0064] During the preset monitoring period, the pollutant concentration data of multiple monitoring points within the preset spatial radius with the candidate pollution source as the center are obtained, and the pollutant concentration value of each monitoring point in a continuous time segment is recorded to form a corresponding pollutant concentration time series data set;

[0065] Normalize the pollutant concentration time series data, calculate the concentration variation of each monitoring point in the corresponding time period in a sliding time window manner, use the mean square error as the time volatility parameter of the concentration change, and obtain the time volatility score value corresponding to each monitoring point;

[0066] The temporal volatility score of each monitoring point is weighted averaged according to the spatial distance from the candidate pollution source point to obtain the temporal latitude stability index of the candidate pollution source point within the preset monitoring period.

[0067] Secondly, in order to further address the problem that the existing technology lacks a systematic data modeling mechanism and response strategy, which limits the real-time traceability and classification response capabilities of pollution incidents, the present invention also provides an intelligent detection system for atmospheric pollution sources, the detection system comprising:

[0068] Data acquisition module: used to collect atmospheric pollutant concentration data, spatiotemporal location information and meteorological conditions at different time, space and altitude levels;

[0069] Model construction module: used to construct a pollutant diffusion path prediction model based on the collected pollutant concentration data, spatiotemporal location information and meteorological conditions, output the pollutant diffusion path prediction value, and predict the pollutant diffusion path prediction value at the point to be predicted; used to construct a pollutant diffusion path prediction model based on the collected pollutant concentration data, spatiotemporal location information and meteorological conditions, output the pollutant diffusion path prediction value, and predict the pollutant diffusion path prediction value at the point to be predicted;

[0070] Intelligent early warning module: used to combine the pollutant diffusion prediction value and the pollution distribution prediction value to build a pollution source inversion mechanism, output the pollution source probability level score, and determine the pollution source probability level.

[0071] In a third aspect, an embodiment of the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the intelligent detection method for atmospheric pollution sources as described in the first aspect of the present invention is implemented.

[0072] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, any step of the intelligent detection method for atmospheric pollution sources as described in the second aspect of the present invention is implemented.

[0073] The beneficial effects of the present invention are as follows: By combining the pollution diffusion path prediction model with the three-dimensional pollution distribution model, the system can output the spatial distribution of pollution sources and the degree of pollution impact. It further introduces a pollution source probability scoring mechanism and combines multiple indicators such as path deviation, temporal latitude stability, and spatial correlation to accurately classify the risk level of pollution sources. This solves the problem of the lack of systematic data modeling mechanisms and response strategies in existing methods, which limits the real-time traceability and classification response capabilities of pollution incidents.

[0074] The present invention sets up a multi-level classification mechanism and a three-level response strategy for pollution sources. By automatically identifying the behavioral characteristics of pollution sources (such as concentrated, diffuse, fluctuating, etc.) to match different response measures, the fine classification of pollution sources and the matching of response strategies are achieved. The system not only has identification capabilities, but also has risk management and auxiliary decision-making capabilities. It is particularly suitable for scenarios where pollution incidents occur suddenly, multiple sources intersect, the terrain is complex, or the wind field has significant dynamic changes. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort. Among them:

[0076] Figure 1 This is a schematic diagram of the implementation process of an intelligent detection method for air pollution sources proposed by the present invention;

[0077] Figure 2 This is a schematic diagram of the pollution source probability level judgment process of the intelligent detection method for atmospheric pollution sources proposed by the present invention. DETAILED DESCRIPTION

[0078] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0079] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0080] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0081] Furthermore, the present invention is described in detail with reference to schematic diagrams. For ease of illustration, when describing the embodiments of the present invention, cross-sectional views illustrating device structures may be partially enlarged and not to scale. Furthermore, the schematic diagrams are merely illustrative and should not limit the scope of protection of the present invention. Furthermore, in actual production, the three-dimensional dimensions of length, width, and depth should be included.

[0082] Example 1

[0083] Reference Figure 1-Figure 2 , as an embodiment of the present invention, provides an intelligent detection method for air pollution sources.

[0084] Existing intelligent detection methods for atmospheric pollution sources have the following main problems: limitations in spatial data fusion, pollution diffusion path prediction, and pollution source inversion analysis. It is difficult to finely identify the type, location, and risk level of pollution sources. There is a lack of systematic data modeling mechanisms and response strategies, which restricts the real-time tracing and classification response capabilities of pollution incidents.

[0085] This application provides an effective solution to the above-mentioned problems. Next, we will explain in detail how to implement the intelligent detection method for air pollution sources in combination with multiple embodiments.

[0086] Figure 1 The figure shows a schematic diagram of the overall implementation steps of an intelligent detection method for air pollution sources, including:

[0087] Step 1: Using monitoring nodes deployed at fixed locations on the ground and multiple drones equipped with pollutant sensing modules, we collect atmospheric pollutant concentration data, spatiotemporal location information, and meteorological conditions at different times, locations, and altitudes.

[0088] Specifically, the sensor modules mounted on the drone and the pollutant concentration data collected by them include:

[0089] Laser scattering particle sensor, used to detect the concentration of inhalable particulate matter;

[0090] Sulfur dioxide and nitrogen dioxide sensors: used to detect sulfur dioxide concentration and nitrogen dioxide concentration;

[0091] Carbon monoxide and ozone sensors: used to detect carbon monoxide and ozone concentrations;

[0092] Organic compound sensor: used to monitor the concentration of volatile organic compounds;

[0093] Temperature and humidity sensor: used to provide auxiliary analysis of environmental parameters;

[0094] Air pressure sensor: assists in judging altitude and weather conditions;

[0095] GPS module: used to record the drone’s flight location;

[0096] Altimeter: used to calibrate the flight altitude of the drone;

[0097] Wind speed and direction sensor: used to assess the diffusion trend of wind impact.

[0098] Spatiotemporal location information includes: UAV three-dimensional location coordinates (longitude, latitude, and altitude), sampling time, flight path data, pollutant concentration data at the sampling point (such as the various gases and particulate matter mentioned above), and flight speed and direction (used to assist data fusion);

[0099] Meteorological conditions include meteorological parameters (wind speed, wind direction, temperature and humidity).

[0100] Step 2: Based on the collected pollutant concentration data, spatiotemporal location information and meteorological conditions, a pollutant diffusion path prediction model is constructed to output the pollutant diffusion path prediction value and predict the pollutant diffusion path prediction value at the prediction point;

[0101] The process of building a pollutant diffusion path prediction model is as follows:

[0102] S201: The pollutant concentration data collected in step 1 and the corresponding three-dimensional space coordinates (x j ,y j , z j ), acquisition time t j , release rate per unit time Δψ j , wind speed and wind direction and other environmental factors are uniformly coded to construct a pollution source unit data set;

[0103] It should be noted that the construction of the pollution source unit dataset includes the following:

[0104] The ground monitoring nodes and the drones equipped with sensors jointly collect pollutant concentration data at multiple spatial pointsψ j And the corresponding space coordinates (x j ,y j , z j ), sampling time interval Δt j and sampling volume ΔV j ;

[0105] Quantitatively calculate the mass Q of pollutants diffused outward from each source point per unit time j , the announcement is as follows:

[0106]

[0107] The extraction of environmental correction factors ensures that pollution source units not only have the original pollution intensity but also dynamically respond to the time-varying characteristics of the atmosphere and urban structure, ensuring the physical rationality and dynamic accuracy of the pollutant diffusion path prediction model;

[0108] Finally, each pollution source unit data item is a six-tuple:

[0109]

[0110] in, is the wind speed vector at the pollution source, which is used for directional modulation of the subsequent pollutant diffusion path prediction model, η j The building blocking coefficient is the building density near the pollution source point analyzed through GIS or BIM data to reflect the degree of inhibition of the upward / lateral spread of pollutants.

[0111] S202: Based on the environmental wind field simulation parameters, terrain characteristics, and the spatiotemporal changes of pollution sources, a spatial attenuation factor λ is introduced for each pollution source unit. j (dynamically adjusted with wind speed and building density, etc.), diffusion weight coefficient k j (Integrating wind direction superposition effect, terrain ventilation resistance and other coefficients) and source release mass Q j ;

[0112] S203: Preset the three-dimensional space coordinates (x i ,y i , z i );

[0113] S204: Based on the pollution source unit data set, spatial attenuation factor, diffusion weight coefficient and the coordinates of the predicted point, a pollutant diffusion path prediction model is constructed, and the predicted value P of the pollutant diffusion path at the predicted point is output. i , which is expressed as follows:

[0114]

[0115] Among them, P i Represents the predicted value of the pollutant diffusion path at the prediction point i.

[0116] Step 3: Based on the atmospheric pollutant concentration data, spatiotemporal location information, meteorological conditions and derived spatial distance parameters, a three-dimensional pollution distribution model is constructed to output the pollution concentration prediction value and the pollution concentration prediction value at the predicted spatial point;

[0117] The process of constructing the three-dimensional pollution distribution model is as follows:

[0118] S301: Integrate pollution concentration data from ground monitoring stations and drones C ij , containing its corresponding observation coordinates (x j ,y j , z j ), acquisition time t j and environment variable h ij ;

[0119] S302: Calculate the square of the spatial distance between the target estimation point i and all observation points j Constructing the high impact term 1+α j ·log(1+h ij );

[0120] S303: Introducing weight factor ω j and the spatial attenuation factor b j ;

[0121] S304: Based on the pollution concentration data of the observation point, the square of the spatial distance, the height influence term, the weight factor and the spatial attenuation factor, a three-dimensional pollution distribution model is constructed to output the pollution concentration prediction value D at the prediction point. i , which is expressed as follows:

[0122]

[0123] Among them, D i Represents the predicted value of pollution concentration at spatial point i.

[0124] Step 4: Combine the pollutant diffusion prediction value and the pollution distribution prediction value to build a pollution source inversion mechanism, output the pollution source probability level score, and determine the pollution source probability level;

[0125] The process of constructing the pollution source inversion mechanism is as follows:

[0126] S401: Define a set of candidate pollution source locations and calculate the probability level score of each candidate source point as a true pollution source;

[0127] S402: Define variables: S = {s1, s1, ..., s n}, represents the set of all candidate pollution source locations, P i is the predicted value of the pollutant diffusion path at the prediction point i, Indicates s i The diffusion path simulated when it is the source point, D i The predicted value of pollution concentration at spatial point i, Indicates s i The predicted value of pollution concentration when it is a source point, γ and δ represent the weighting coefficients of path score and concentration score, respectively;

[0128] S403: Construct a pollution source probability scoring model and output the pollution source probability level score, which is expressed as follows:

[0129]

[0130] in, Indicates the probability level score of the pollution source, H P (s i ) represents the path inconsistency, H D (s i ) Pollution concentration residual, Z is the normalization factor.

[0131] The probability values of all candidate source points According to their size, they are divided into multiple level intervals, which are as follows:

[0132] Identify the source of pollution (red alert);

[0133] Possible pollution source (orange alert);

[0134] Suspected pollution source (yellow reminder);

[0135] Non-pollution source (green area);

[0136] In the present invention, the candidate source points are divided into four intervals with reference to the standards for grading environmental pollution events (especially serious / serious / large / general) in the National Emergency Plan for Environmental Emergencies.

[0137] Figure 2 A schematic diagram of a pollution source probability level judgment process of an intelligent detection method for air pollution sources is shown, including:

[0138] Pollution source probability level score output by the pollution source probability scoring model Set the three thresholds as critical threshold T H (used to distinguish whether the candidate pollution source point reaches the demarcation standard requiring the highest level response (red alert)), risk threshold T M and safety threshold T L ;

[0139] like If the candidate source point falls into the red warning area, the system will execute the first-level response strategy and make a first-level judgment;

[0140] When the candidate source point falls into the red warning area, the diffusion path deviation ΔP of all monitoring points within the radius R1 with the candidate source point as the center is extracted. i , and set the path deviation low threshold δ low and path deviation high threshold δ high ;

[0141] If more than M1% of the monitoring points meet ΔP i <δ low , then the candidate source point is determined to be a concentrated pollution source;

[0142] If more than M2% of the monitoring points meet δ low ≤ΔP i <δ high , then the candidate source point is determined to be a diffuse pollution source;

[0143] On the contrary (that is, when the candidate source point falls into the red warning area and is neither a concentrated pollution source nor a diffuse pollution source), the candidate source point is determined to be an abnormal pollution source;

[0144] Specifically, the first-level response strategy operates as follows:

[0145] If the candidate source point is a centralized pollution source, the first-level response strategy performs the following operations:

[0146] The drone grid sampling mode was activated, and sampling points were arranged in a 100-meter × 100-meter grid. Sampling continued for 15 minutes at each grid center point. Simultaneously, a ground mobile monitoring vehicle was dispatched to implement surrounding sampling within 200 meters downwind of the pollution source, and the sampling interval was set to 5 minutes per time. A vehicle-mounted lidar system was used to perform vertical scanning of suspected emission outlets. The scanning angle was set to 0-90 degrees with a step accuracy of 0.1 degrees. A portable gas chromatography-mass spectrometry was used to perform qualitative analysis of characteristic pollutants. The sampling flow rate was controlled at 500 mL / min. According to Article 24 of the Air Pollution Prevention and Control Law, the companies involved were ordered to immediately implement a 50% production restriction measure.

[0147] If the candidate source is a diffuse pollution source, the first-level response strategy performs the following operations:

[0148] Deploy UAVs to conduct fan-shaped trajectory sampling along the dominant wind direction, with the fan angle set to 60 degrees and the flight altitude gradients of 50 meters, 100 meters, and 150 meters. Ground-based monitoring vehicles move along the pollution diffusion axis at a speed of 20 km / h for sampling, and the data collection frequency is set to 10 seconds / time. Run the CALPUFF atmospheric diffusion model, with input parameters including terrain data (accuracy of 30 meters), hourly meteorological data, and pollution source strength data. Perform reverse deduction calculations, set the time step to 15 minutes, and output a pollution source probability distribution map. Multifunctional dust suppression vehicles carry out spraying operations in the polluted area, with a spray volume of 3L / m 2 h standard implementation. Coordinate with enterprises within a 3-kilometer upwind area to implement staggered production, reducing production load to 70% of normal for at least 24 hours.

[0149] It should be noted that the CALPUFF atmospheric diffusion model is derived from the regulatory model recommended by the EPA. As an existing technology, it is used in the present invention for path inversion of diffuse pollution sources. By inputting real-time monitoring data and meteorological parameters, it outputs a probability distribution map of pollution sources. Compared with the traditional Gaussian model, it can accurately handle the influence of complex terrain and support long-distance transmission simulation (>50km).

[0150] If the candidate source point is an abnormal pollution source, the first-level response strategy performs the following operations:

[0151] The drone's spiral stereo sampling was launched, and the flight radius was gradually expanded from 50 meters to 300 meters, with an altitude gradient of 20 meters, 50 meters, and 100 meters. The GF-4 satellite remote sensing data was connected, the spatial resolution was increased to 50 meters, and the revisit period was shortened to 15 minutes. The sudden environmental event tracing model was run, and the input parameters included pollutant fingerprint characteristics, meteorological field data, and chemical enterprise database. The hazardous chemical feature fingerprint library was called for spectral comparison, and the matching threshold was set at 85%. According to the "National Emergency Plan for Environmental Sudden Events", a Level III response was initiated, and a warning area with a radius of 500 meters was demarcated. Four fixed monitoring points and two mobile monitoring points were set up. Non-emergency response personnel were prohibited from entering the warning area, and the duration was not less than 6 hours.

[0152] It should be noted that the sudden environmental event tracing model is a special model developed based on the "Technical Specifications for Emergency Monitoring of Sudden Environmental Events" (HJ 589-2021). As an existing technology, it can provide rapid tracing capabilities for abnormal pollution sources in the present invention and predict the scope of impact.

[0153] like If the candidate source point falls into the orange warning area, the system will execute the secondary response strategy and make a secondary judgment;

[0154] When the candidate source point falls into the orange warning area, the time-latitude stability index η is calculated and the volatility high threshold η is set high and volatility low threshold η low ;

[0155] If η>η high , then the candidate source point is determined to be a volatility risk source;

[0156] If η low <η≤η high , then the candidate source point is determined to be a gradual risk source;

[0157] If η≤η low , then the candidate source point is determined to be a stable risk source.

[0158] Specifically, the secondary response strategy operates as follows:

[0159] If the candidate source point is a volatility risk source, the secondary response strategy performs the following operations:

[0160] One fixed monitoring point is added 200 meters upwind and 500 meters downwind of the pollution source, and the sampling frequency is increased to once an hour. The real-time data of the enterprise's production conditions is connected, and the data collection interval is set to 1 minute. A correlation model between production parameters and emission concentrations is established, and the correlation coefficient threshold is set to 0.7. Enterprise equipment with a correlation coefficient exceeding the threshold is subject to key inspections. According to the "Regulations on the Management of Pollutant Discharge Permits", a notice of rectification within a time limit is issued, and the rectification period shall not exceed 15 working days. Enterprises are required to install online monitoring equipment for working conditions, with a data transmission frequency of not less than 5 minutes / time, and directly connected to the monitoring platform of the environmental protection department.

[0161] It should be noted that the correlation model between production parameters and emission concentrations belongs to the existing technology. Its application in the present invention can achieve accurate identification of fluctuating risk sources and provide a quantitative basis for environmental law enforcement.

[0162] If the candidate source point is a gradual risk source, the secondary response strategy performs the following operations:

[0163] Key sampling is carried out every day during the two pollution peak periods of 08:00-10:00 in the morning and 18:00-20:00 in the evening. Each sampling lasts no less than 30 minutes. The sampling points are arranged at 10 meters, 50 meters, and 100 meters outside the factory boundary of the enterprise. The equipment aging assessment model is adopted. The input parameters include service life (accurate to the month), maintenance records, and energy consumption data. The output is the equipment health index. Equipment with an index below 60 points is required to be immediately shut down for maintenance. The enterprise must submit a detailed maintenance plan within 7 working days, clarifying the maintenance content, time nodes and responsible persons. The environmental protection department implements monthly supervisory monitoring, and the monitoring frequency is increased from 1 time / month to 2 times / month.

[0164] It should be noted that the equipment aging degree assessment model belongs to the existing technology, and its application in the present invention can implement equipment health diagnosis for gradual risk sources.

[0165] If the candidate source is a stable risk source, the secondary response strategy performs the following operations:

[0166] Two fixed monitoring points are set up 100 meters upwind and 300 meters downwind of the pollution source. Supervisory monitoring is carried out on the 5th, 15th and 25th of each month, with continuous sampling for 24 hours each time. The online monitoring data of the enterprise in the past three months is retrieved, and the effectiveness is audited according to the HJ 75-2017 standard, focusing on checking the time periods when the hourly average pollutant emission concentration exceeds the limit by 50%, and checking the data missing rate. In the months when the rate exceeds 5%, the enterprise is required to submit an explanatory report, and a 72-hour continuous performance test is conducted on the pollution control facilities. The test indicators include treatment efficiency (calculated according to GB / T 16157), energy consumption ratio and operational stability. The enterprise is required to improve the "one enterprise, one file" environmental protection management file, submit a self-monitoring report once a month, and submit a rectification plan for unqualified equipment within 15 working days. After the rectification is completed, a 72-hour continuous verification monitoring is carried out.

[0167] like If the candidate source point falls into the yellow warning area, the system will execute the three-level response strategy and make three-level judgments;

[0168] When the candidate source point falls into the yellow warning area, the spatial correlation index ρ is calculated and the first spatial correlation threshold ρ is set high , the second spatial correlation threshold ρ low and volatility warning value η y ;

[0169] During the preset monitoring period, the pollutant concentration data of multiple monitoring points within the preset spatial radius with the candidate pollution source as the center are obtained, and the pollutant concentration value of each monitoring point in a continuous time segment is recorded to form a corresponding pollutant concentration time series data set;

[0170] It should be noted that the preset monitoring period is 24 to 72 hours. According to EPA standards such as the "Technical Assistance Document for the Reporting of Daily Air Quality–Air Quality Index (EPA-454 / B-18-007)", the diffusion and superposition effects of atmospheric pollutants (such as PM2.5, NO2, and SO2) can form a significant change pattern within 24 to 72 hours, which is the optimal time scale for capturing the behavioral characteristics of pollution sources. The present invention involves the analysis of pollution source behavior patterns, which requires consideration of both sudden pollution and gradual pollution. Therefore, 48 hours is selected as the standard period that takes into account both response speed and the adequacy of pollution evolution. This can capture sudden pollution sources while not missing gradual pollution sources due to a short time period.

[0171] The preset spatial radius range is 300 meters to 800 meters. According to the "Technical Regulations for the Application of Atmospheric Pollutant Diffusion Models" (HJ2.2-2018) and the "CALPUFF User Guide", in urban environments, the main impact range of typical pollution sources is generally concentrated in the area 300 to 1000 meters around the source point, and the pollution concentration changes within 500 meters are most sensitive.

[0172] Normalize the pollutant concentration time series data, and calculate the concentration variation of each monitoring point in the corresponding time period using a sliding time window. Use the mean square error as the time volatility parameter of the concentration change to obtain the time volatility score value corresponding to each monitoring point.

[0173] The temporal volatility score of each monitoring point is weighted averaged according to the spatial distance from the candidate pollution source point to obtain the temporal latitude stability index of the candidate pollution source point within the preset monitoring period;

[0174] If ρ>ρ high , then the candidate source point is determined to be a related suspicious source, and the next step is executed;

[0175] When the candidate source point is determined to be a correlated suspicious source, if η>η y , then the candidate source point is determined to fall into the orange warning area;

[0176] If ρ low <ρ≤ρ high , then the candidate source point is determined to be an occasional suspicious source; then proceed to the next step;

[0177] When the candidate source point is determined to be an occasional suspicious source, if the diffusion path deviation ΔP i >δ high , then the candidate source point is determined to fall into the red warning area;

[0178] If ρ≤ρ low , then the candidate source point is determined to be an isolated suspicious source;

[0179] Specifically, the three-level response strategy operates as follows:

[0180] If the candidate source point is a related suspicious source, the three-level response strategy performs the following operations:

[0181] Simultaneous sampling was carried out at the suspected pollution source and its three nearest adjacent monitoring points. The sampling time error was controlled within ±2 minutes, the sampling duration was set to 1 hour, and the sampling flow rate was stabilized at 1.0L / min. Correlation analysis of pollutant components was performed, and the company's raw and auxiliary material usage records were verified, with a focus on comparing the purchase volume and actual usage of characteristic pollutants. In accordance with the "Environmental Administrative Penalty Measures", a formal environmental inquiry letter was issued, requiring the company to make a written reply within 3 working days, and "double random" inspections were simultaneously initiated, randomly selecting law enforcement personnel and inspection times.

[0182] If the candidate source point is an occasional suspicious source, the three-level response strategy performs the following operations:

[0183] Surprise sampling will be carried out on randomly selected working days and non-working days respectively. The sampling time will not be notified in advance. The sampling point will be set at the factory boundary downwind of the pollution source. The sampling time will be no less than 2 hours. The time series consistency of the enterprise's production log and the automatic monitoring data will be compared. The time alignment accuracy is required to reach ±5 minutes. The enterprise shall submit a situation statement with the official seal within 5 working days, and the explanatory materials shall be accompanied by relevant supporting materials.

[0184] If the candidate source point is an isolated suspicious source, the three-level response strategy performs the following operations:

[0185] Three control sampling points are set up within a radius of 100 meters from the suspected pollution source. The sampling points are arranged in an equilateral triangle. Synchronous sampling is adopted. The sampling periods cover three typical periods: morning (08:00-10:00), noon (13:00-15:00), and evening (18:00-20:00). The sampling flow rate is controlled at 1.0±0.1L / min, and the single sampling time is ≥60 minutes. The ground meteorological station data for the last 72 hours is retrieved, and the point is included in the list of "special attention points". Random inspections are carried out once a week (for 1 month), and key monitoring is carried out during meteorologically sensitive periods. Case files are established and the file retention period is no less than 3 years.

[0186] like This indicates that the candidate source point falls into the green area, and the area is marked as the background reference area. The drone path is adjusted to avoid this area.

[0187] For example, suppose a city's environmental protection department deploys an intelligent detection system for air pollution sources in an industrial zone. One day, the system detects an abnormally high PM2.5 concentration in the area through ground monitoring stations and drone networks, and immediately initiates the intelligent pollution source detection process. The system first integrates monitoring data, including PM2.5 concentration, three-dimensional spatial coordinates, wind speed, wind direction and other environmental parameters, to construct a pollutant diffusion path prediction model. Model analysis shows that pollutants mainly diffuse to the northwest and form a high-concentration cluster near an industrial park. Subsequently, the system combines the three-dimensional pollution distribution model to calculate the pollution concentration prediction value of each spatial point and locks in several high-probability pollution source candidate points. Through the pollution source inversion mechanism, the system assigns probability scores to the candidate points and finally determines that a point near a chemical plant is a "confirmed pollution source" (red alert). Based on the system's warning information, the environmental protection department immediately sends personnel to the site for verification and finds that the chemical plant's waste gas treatment equipment has a leakage problem. After emergency repairs, the PM2.5 concentration gradually returns to normal levels and the system warning is lifted.

[0188] In summary, by combining the pollution diffusion path prediction model with the three-dimensional pollution distribution model, the system can output the spatial distribution of pollution sources and the degree of pollution impact. Furthermore, a pollution source probability scoring mechanism is introduced, and multiple indicators such as path deviation, temporal and latitude stability, and spatial correlation are combined to accurately classify the risk level of pollution sources. This solves the problem of the lack of systematic data modeling mechanisms and response strategies in existing methods, which limits the real-time traceability and classification response capabilities of pollution incidents.

[0189] The present invention sets up a multi-level classification mechanism for pollution sources and a three-level response strategy. By automatically identifying the behavioral characteristics of pollution sources (such as concentrated, diffuse, and fluctuating types), different response measures are matched to achieve fine classification of pollution sources and matching of response strategies. It not only has identification capabilities, but also has risk management and decision-making support capabilities.

[0190] Example 2, an embodiment of the present invention, provides an intelligent detection system for air pollution sources, including:

[0191] Data acquisition module: used to collect atmospheric pollutant concentration data, spatiotemporal location information and meteorological conditions at different time, space and altitude levels;

[0192] Model construction module: used to construct a pollutant diffusion path prediction model based on the collected pollutant concentration data, spatiotemporal location information and meteorological conditions, output the pollutant diffusion path prediction value, and predict the pollutant diffusion path prediction value at the point to be predicted; used to construct a pollutant diffusion path prediction model based on the collected pollutant concentration data, spatiotemporal location information and meteorological conditions, output the pollutant diffusion path prediction value, and predict the pollutant diffusion path prediction value at the point to be predicted;

[0193] Intelligent early warning module: used to combine the pollutant diffusion prediction value and the pollution distribution prediction value to build a pollution source inversion mechanism, output the pollution source probability level score, and determine the pollution source probability level.

[0194] Example 3 is an embodiment of the present invention, which is different from the previous embodiment in that:

[0195] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, and other media that can store program code.

[0196] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0197] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering, or processing in another suitable manner as necessary, and then stored in a computer memory.

[0198] It should be understood that various parts of the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application-specific integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0199] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. An intelligent detection method for air pollution sources, characterized in that: include: Step 1: Collect atmospheric pollutant concentration data, spatiotemporal location information, and meteorological conditions at different time, space, and altitude levels; Step 2: Based on the collected pollutant concentration data, spatiotemporal location information and meteorological conditions, a pollutant diffusion path prediction model is constructed to output the pollutant diffusion path prediction value and predict the pollutant diffusion path prediction value at the prediction point; Step 3: Based on the atmospheric pollutant concentration data, spatiotemporal location information, meteorological conditions and derived spatial distance parameters, a three-dimensional pollution distribution model is constructed to output the pollution concentration prediction value and the pollution concentration prediction value at the predicted spatial point; Step 4: Combine the pollutant diffusion prediction value and the pollution distribution prediction value to build a pollution source inversion mechanism, output the pollution source probability level score, and determine the pollution source probability level.

2. The method for intelligent detection of air pollution sources according to claim 1, characterized in that: The pollutant diffusion path prediction model construction process is as follows: S201: uniformly encode the pollutant concentration data collected in step 1, the corresponding three-dimensional spatial coordinates, the collection time, the release rate per unit time, the wind speed and the wind direction to construct a pollution source unit data set; S202: Based on the environmental wind field simulation parameters, terrain characteristics, and the spatiotemporal variation of pollution sources, a spatial attenuation factor, a diffusion weight coefficient, and source release mass are introduced for each pollution source unit; S203: Preset the three-dimensional spatial coordinates (longitude, latitude, altitude) of the point to be predicted; S204: Based on the pollution source unit data set, spatial attenuation factor, diffusion weight coefficient and the coordinates of the predicted point, a pollutant diffusion path prediction model is constructed, and the predicted value P of the pollutant diffusion path at the predicted point is output. i .

3. The method for intelligent detection of air pollution sources according to claim 1, characterized in that: The three-dimensional pollution distribution model construction process is as follows: S301: Integrate pollution concentration data from ground monitoring stations and drones, including their corresponding observation coordinates (longitude, latitude, altitude), collection time, and environmental variables; S302: Calculate the square of the spatial distance between the target estimation point and all observation points, and construct a height influence term; S303: configuring a weight factor and a spatial attenuation factor for each observation point to adjust its contribution to the three-dimensional pollution distribution model; S304: Based on the pollution concentration data of the observation point, the square of the spatial distance, the height influence term, the weight factor and the spatial attenuation factor, a three-dimensional pollution distribution model is constructed to output the pollution concentration prediction value D at the prediction point. i .

4. The method for intelligent detection of air pollution sources according to claim 3, characterized in that: The pollution source inversion mechanism construction process is as follows: S401: Define a set of candidate pollution source locations and calculate the probability level score of each candidate source point as a true pollution source; S402: Input the predicted value of the pollutant diffusion path, the simulated diffusion path value, the predicted value of the pollution concentration, and the simulated concentration value, and set the weighting coefficients of the path score and the concentration score; S403: Construct a pollution source probability scoring model and output pollution source probability level scores 5. The method for intelligent detection of air pollution sources according to claim 4, characterized in that: The probability values of all candidate source points According to their size, they are divided into multiple level intervals, and the multiple level intervals are as follows: Identify the source of pollution (red alert); Possible pollution source (orange alert); Suspected pollution source (yellow reminder); Non-pollution sources (green areas).

6. The method for intelligent detection of air pollution sources according to claim 5, characterized in that: Pollution source probability level score output by the pollution source probability scoring model Set the three thresholds as critical threshold T H , risk threshold T M and safety threshold T L ; like If the candidate source point falls into the red warning area, the system will execute the first-level response strategy and make a first-level judgment; When the candidate source point falls into the red warning area, the diffusion path deviation ΔP of all monitoring points within the radius R1 with the candidate source point as the center is extracted. i , and set the path deviation low threshold δ low and path deviation high threshold δ high ; If more than M1% of the monitoring points meet ΔP i <δ low , then the candidate source point is determined to be a concentrated pollution source; If more than M2% of the monitoring points meet δ low ≤ΔP i <δ high , then the candidate source point is determined to be a diffuse pollution source; Otherwise, the candidate source point is determined to be an abnormal pollution source; like If the candidate source point falls into the orange warning area, the system will execute the secondary response strategy and make a secondary judgment; When the candidate source point falls into the orange warning area, the time-latitude stability index η is calculated and the volatility high threshold η is set high and volatility low threshold η low ; If η>η high , then the candidate source point is determined to be a volatility risk source; If η low <η≤η high , then the candidate source point is determined to be a gradual risk source; If η≤η low , then the candidate source point is determined to be a stable risk source; like If the candidate source point falls into the yellow warning area, the system will execute the three-level response strategy and make three-level judgments; When the candidate source point falls into the yellow warning area, the spatial correlation index ρ is calculated and the first spatial correlation threshold ρ is set high , the second spatial correlation threshold ρ low and volatility warning value η y ; If ρ>ρ high , then the candidate source point is determined to be a related suspicious source, and the next step is executed; When the candidate source point is determined to be a correlated suspicious source, if η>η y , then the candidate source point is determined to fall into the orange warning area; If ρ low <ρ≤ρ high , then the candidate source point is determined to be an occasional suspicious source; then proceed to the next step; When the candidate source point is determined to be an occasional suspicious source, if the diffusion path deviation ΔP i >δ high , then the candidate source point is determined to fall into the red warning area; If ρ≤ρ low , then the candidate source point is determined to be an isolated suspicious source; like This indicates that the candidate source point falls into the green area, and the area is marked as the background reference area. The drone path is adjusted to avoid this area.

7. The intelligent detection method for air pollution sources according to claim 6, characterized in that: The calculation process of the spatial correlation index is as follows: During the preset monitoring period, the pollutant concentration data of multiple monitoring points within the preset spatial radius with the candidate pollution source as the center are obtained, and the pollutant concentration value of each monitoring point in a continuous time segment is recorded to form a corresponding pollutant concentration time series data set; Normalize the pollutant concentration time series data, calculate the concentration variation of each monitoring point in the corresponding time period in a sliding time window manner, use the mean square error as the time volatility parameter of the concentration change, and obtain the time volatility score value corresponding to each monitoring point; The temporal volatility score of each monitoring point is weighted averaged according to the spatial distance from the candidate pollution source point to obtain the temporal latitude stability index of the candidate pollution source point within the preset monitoring period.

8. An intelligent air pollution source detection system, based on the intelligent air pollution source detection method according to any one of claims 1 to 7, characterized in that: The detection system comprises: Data acquisition module: used to collect atmospheric pollutant concentration data, spatiotemporal location information and meteorological conditions at different time, space and altitude levels; Model construction module: used to construct a pollutant diffusion path prediction model based on the collected pollutant concentration data, spatiotemporal location information and meteorological conditions, output the pollutant diffusion path prediction value, and predict the pollutant diffusion path prediction value at the point to be predicted; used to construct a pollutant diffusion path prediction model based on the collected pollutant concentration data, spatiotemporal location information and meteorological conditions, output the pollutant diffusion path prediction value, and predict the pollutant diffusion path prediction value at the point to be predicted; Intelligent early warning module: used to combine the pollutant diffusion prediction value and the pollution distribution prediction value to build a pollution source inversion mechanism, output the pollution source probability level score, and determine the pollution source probability level.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the intelligent detection method for air pollution sources 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 a processor, the steps of the intelligent detection method for air pollution sources according to any one of claims 1 to 7 are implemented.

Images