A method, system, medium and device for predicting the VOCs emission range of a tower platform

Abstract

The application provides a tower platform VOCs emission range prediction method, system, medium and equipment, belonging to the technical field of pollution tracing, which comprises the following steps: obtaining an initial remote sensing concentration image and regional meteorological data; taking the measured concentration data of the calibration point deployed in the park as a benchmark, correcting the initial remote sensing concentration image in the spatial domain, and generating a corrected concentration field image; inputting the corrected concentration field image and the regional meteorological data into a Gaussian diffusion model, simulating the diffusion trajectory of VOCs and backstepping the emission source position; based on the emission source position and the regional meteorological data, predicting the geographical area affected by the pollution group in a future preset time period; and generating the VOCs emission range according to the emission source position and the predicted geographical area. The application combines remote sensing scanning and ground calibration, eliminates system error and background interference, significantly improves the accuracy of the concentration field data and the tracing accuracy, realizes large-scale rapid monitoring, dynamic diffusion prediction and full-process automatic supervision.

Description

Technical Field

[0001] This invention belongs to the field of pollution source tracing technology, and in particular relates to a method, system, medium and equipment for predicting the emission range of VOCs from a tower platform. Background Technology

[0002] Currently, the monitoring and source tracing technologies for the fugitive emissions of VOCs in industrial parks have mainly formed two major systems: ground-based monitoring networks and remote sensing technology. Ground-based monitoring networks acquire point concentration data by distributing fixed or mobile sensors within the park. This technology is highly mature and can provide relatively accurate local concentration information, serving as a fundamental means of environmental supervision. At the data processing level, simple concentration contour maps or statistical correlation analysis are often used to initially determine the direction of pollution. Some advanced systems attempt to introduce atmospheric diffusion models, such as Gaussian models, Lagrange particle models, or computational fluid dynamics models, combining concentration with meteorological data for reverse source tracing. These technologies collectively constitute the current technical landscape for VOCs emission monitoring, capturing pollution signals from multiple dimensions.

[0003] In existing technologies, monitoring methods still face a trade-off between coverage and accuracy: ground-based monitoring networks have limited spatial resolution, resulting in numerous blind spots and making it easy to miss instantaneous, unorganized fugitive emissions, and they struggle to reflect the two-dimensional concentration field distribution across the entire park; while existing remote sensing technologies, especially single-point path integral measurements, offer wider coverage but cannot directly provide spatially resolved concentration images, and their measurement results are susceptible to atmospheric conditions, background interference, and instrument drift, leading to insufficient quantitative reliability and significant errors when directly used for model inversion. Secondly, source tracing capabilities are weak and practical: existing methods often oversimplify the integration of monitoring data with models, with most systems only capable of rough orientation determination and lacking precise source strength and location inversion capabilities. Simple Gaussian model applications are often based on inaccurate initial concentration fields and do not consider the impact of complex terrain or building complexes; more sophisticated CFD models are computationally expensive and difficult to implement with rapid response. More importantly, when multiple potential fugitive sources exist within a park, existing technologies lack effective multi-source identification and separation mechanisms, failing to accurately analyze the contribution of each source in complex emission scenarios. Finally, existing systems are functionally limited, mostly focusing on monitoring and simple display, lacking integrated capabilities to link source tracing results with real-time prediction, risk assessment, and regulatory decision-making. This makes it difficult to directly translate technological achievements into effective on-site control actions. These shortcomings collectively limit the practical application of existing technologies in the precise source tracing and emission prediction of fugitive VOCs diffusion in industrial parks. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method, system, medium, and device for predicting the VOCs emission range of a tower platform.

[0005] Firstly, the method includes, Acquire initial remote sensing concentration images and regional meteorological data; Based on the measured concentration data from multiple calibration points pre-deployed within the park, the initial remote sensing concentration image is spatially corrected to generate a corrected concentration field image. Based on the corrected concentration field image and regional meteorological data, the diffusion trajectory of VOCs was simulated using a Gaussian diffusion model, and the location of the emission source was inferred. Based on the location of the fugitive source and regional meteorological data, the geographical area affected by the pollution plume in the future within a preset time period is predicted using a Gaussian diffusion model. Based on the location of the emission source and the predicted geographical region, the emission range of VOCs is generated.

[0006] Furthermore, the acquisition of the initial remote sensing concentration image specifically includes acquiring the initial remote sensing concentration map of VOCs by performing horizontal or vertical scanning using infrared remote sensing equipment deployed on a high tower platform in the industrial park.

[0007] Furthermore, the generation of the corrected concentration field image specifically includes, Based on the coordinate mapping relationship, the observed concentration value of the pixel closest to the spatial coordinates of each calibration point is extracted from the initial remote sensing concentration image. Calculate the residual between the observed concentration value and the measured concentration value at each calibration point; Based on the spatial coordinates of all calibration points and their corresponding residuals, a residual field covering the entire initial remote sensing density image area is constructed using a spatial interpolation algorithm. The constructed residual field is used to correct the initial remote sensing concentration image pixel by pixel, generating a corrected concentration field image.

[0008] Furthermore, the spatial interpolation algorithm includes: inverse distance weighted interpolation, Kriging interpolation, and radial basis function interpolation.

[0009] Furthermore, the step of using the constructed residual field to perform pixel-by-pixel correction on the initial remote sensing concentration image to generate a corrected concentration field image specifically includes: The residual field is considered as a background concentration field composed of systematic bias and background disturbance; The initial remote sensing concentration image is regarded as a mixed signal image formed by superimposing the target VOCs emission plume signal with this background concentration field; By performing pixel-level differential operations, the background concentration field is stripped from the mixed signal image to extract the target concentration image that mainly reflects the spatial distribution of the target VOCs fume; A corrected concentration field image is generated based on the target concentration image.

[0010] Furthermore, the simulation of VOCs diffusion trajectories and the deduction of their emission source locations specifically includes, The corrected concentration field image is used as the source intensity distribution input for the Gaussian diffusion model; Wind speed and direction from meteorological data are used as the dominant driving parameters of the Gaussian diffusion model. In the Gaussian diffusion model, the hypothesized source location and source strength are adjusted through an iterative optimization algorithm to minimize the difference between the simulated concentration field and the corrected concentration field image. The source location corresponding to this point is then determined as the inversely derived emission source location.

[0011] Furthermore, the geographical area predicted to be affected by the pollution plume within a preset future time period specifically includes, Using the location of the fugitive source as the origin of pollutant release and the prevailing wind direction of regional meteorological data as the axis, a fan-shaped prediction area is constructed. Based on the Gaussian plume diffusion model, the estimated source strength, real-time wind speed and terrain correction coefficient of the escape source location are used as input parameters to simulate the concentration distribution of pollutants along the prevailing wind axis and both sides in the fan-shaped prediction area, and to calculate the movement path and diffusion range of the pollution plume in the future preset time period. The simulated concentration distribution results are compared with environmental safety thresholds to define the predicted risk areas. Image processing techniques are then used to extract the geometric boundaries of the predicted risk areas and generate a sequence of geographic boundary coordinates.

[0012] Secondly, the system includes: a data acquisition module, an image correction module, a location prediction module, a region prediction module, and a scatter range generation module; The data acquisition module is used to acquire initial remote sensing concentration images and regional meteorological data; The image correction module is used to perform spatial domain correction on the initial remote sensing concentration image based on the measured concentration data of multiple calibration points pre-deployed in the park, and generate a corrected concentration field image. The location prediction module is used to simulate the diffusion trajectory of VOCs and infer the location of the emission source based on the corrected concentration field image and regional meteorological data using a Gaussian diffusion model. The regional prediction module is used to predict the geographical area affected by the pollution plume within a preset time period based on the location of the fugitive source and regional meteorological data, using a Gaussian diffusion model. The emission range generation module is used to generate the emission range of VOCs based on the location of the emission source and the predicted geographical area.

[0013] Furthermore, the data acquisition module is specifically used to acquire an initial remote sensing concentration map of VOCs by performing horizontal or vertical scanning using infrared remote sensing equipment deployed on a high tower platform in the industrial park.

[0014] Furthermore, the image correction module is specifically used for, Based on the coordinate mapping relationship, the observed concentration value of the pixel closest to the spatial coordinates of each calibration point is extracted from the initial remote sensing concentration image. Calculate the residual between the observed concentration value and the measured concentration value at each calibration point; Based on the spatial coordinates of all calibration points and their corresponding residuals, a residual field covering the entire initial remote sensing density image area is constructed using a spatial interpolation algorithm. The constructed residual field is used to correct the initial remote sensing concentration image pixel by pixel, generating a corrected concentration field image.

[0015] Furthermore, the spatial interpolation algorithm includes: inverse distance weighted interpolation, Kriging interpolation, and radial basis function interpolation.

[0016] Furthermore, the image correction module is also used for, The residual field is considered as a background concentration field composed of systematic bias and background disturbance; The initial remote sensing concentration image is regarded as a mixed signal image formed by superimposing the target VOCs emission plume signal with this background concentration field; By performing pixel-level differential operations, the background concentration field is stripped from the mixed signal image to extract the target concentration image that mainly reflects the spatial distribution of the target VOCs fume; A corrected concentration field image is generated based on the target concentration image.

[0017] Furthermore, the location prediction module is specifically used for, The corrected concentration field image is used as the source intensity distribution input for the Gaussian diffusion model; Wind speed and direction from meteorological data are used as the dominant driving parameters of the Gaussian diffusion model. In the Gaussian diffusion model, the hypothesized source location and source strength are adjusted through an iterative optimization algorithm to minimize the difference between the simulated concentration field and the corrected concentration field image. The source location corresponding to this point is then determined as the inversely derived emission source location.

[0018] Furthermore, the aforementioned region prediction module is specifically used for: Using the location of the fugitive source as the origin of pollutant release and the prevailing wind direction of regional meteorological data as the axis, a fan-shaped prediction area is constructed. Based on the Gaussian plume diffusion model, the estimated source strength, real-time wind speed and terrain correction coefficient of the escape source location are used as input parameters to simulate the concentration distribution of pollutants along the prevailing wind axis and both sides in the fan-shaped prediction area, and to calculate the movement path and diffusion range of the pollution plume in the future preset time period. The simulated concentration distribution results are compared with environmental safety thresholds to define the predicted risk areas. Image processing techniques are then used to extract the geometric boundaries of the predicted risk areas and generate a sequence of geographic boundary coordinates.

[0019] Thirdly, a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of any of the above-described methods for predicting the VOCs emission range of a tower platform.

[0020] Fourthly, an electronic device includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus. Memory, used to store computer programs; When the processor executes the program stored in the memory, it implements any of the steps of the above-described method for predicting the VOCs emission range of the tower platform.

[0021] Compared with the prior art, the present invention has the following advantages: 1. This invention combines infrared remote sensing scanning from a high-tower platform with measured data from ground calibration points to perform spatial domain correction on the initial remote sensing concentration image, eliminating systematic errors and interference from complex environmental backgrounds in remote sensing observations, thus significantly improving the quantitative reliability of concentration field data. Furthermore, based on this, it combines a Gaussian diffusion model with an iterative optimization algorithm for source inversion, greatly improving the accuracy of the inversion of the escape source location. This solves the technical problems of insufficient reliability of remote sensing data and low source tracing accuracy in existing technologies, providing accurate initial input conditions for subsequent diffusion simulation and risk prediction.

[0022] 2. This invention utilizes the high-altitude advantage of a tower platform to deploy infrared remote sensing equipment, enabling rapid and large-scale scanning and monitoring of industrial parks at the square kilometer level. This overcomes the shortcomings of traditional ground-based point monitoring networks, such as limited coverage, insufficient spatial resolution, and numerous monitoring blind spots. Simultaneously, through a Gaussian diffusion model driven by real-time meteorological data, it can dynamically predict the diffusion path and impact range of pollutants within a preset time period. The predicted risk areas are rendered in real-time on an electronic map with highlighted boundaries, providing forward-looking decision support for environmental emergency response and risk avoidance.

[0023] 3. This invention deeply integrates remote sensing image data, ground-measured concentration data, and multi-dimensional meteorological data. Combining spatial interpolation, image processing, and iterative optimization algorithms, it constructs a fully automated processing chain from data acquisition, spatial domain correction, source inversion, regional statistics, diffusion prediction to diagnostic report generation. This chain reduces human intervention and subjective errors, and realizes integrated linkage of monitoring, source tracing, early warning, and decision support. It enables technological achievements to be directly transformed into effective on-site control actions, significantly improving the efficiency and response speed of environmental supervision.

[0024] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures pointed out in the description, claims and drawings. Attached Figure Description

[0025] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0026] Figure 1 A schematic flowchart of a method for predicting the emission range of VOCs from a tower platform according to the present invention is shown.

[0027] Figure 2 A schematic diagram of a VOCs emission range prediction system module for a tower platform according to the present invention is shown. Detailed Implementation

[0028] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] like Figure 1 As shown, this invention proposes a method for predicting the VOCs emission range of a steel tower platform, the steps of which include: S1. Acquire initial remote sensing concentration images and regional meteorological data.

[0030] In this embodiment, the present invention uses an infrared remote sensing device deployed on a high tower platform in an industrial park to perform horizontal or vertical scanning to obtain an initial remote sensing concentration image of VOCs over a large area; wherein, the initial remote sensing concentration image is a two-dimensional matrix, and each pixel contains a geographical location information and a corresponding observed value of VOCs concentration. In this embodiment, the present invention also collects regional meteorological data from the meteorological monitoring station at the scanning time and within a period before and after, including wind speed, wind direction, ambient temperature, relative humidity and atmospheric stability parameters.

[0031] S2. Based on the measured concentration data of multiple calibration points pre-deployed in the park, the initial remote sensing concentration image is spatially corrected to generate a corrected concentration field image.

[0032] S2.1 Based on the coordinate mapping relationship, extract the observed concentration value of the pixel closest to the spatial coordinates of each calibration point from the initial remote sensing concentration image; S2.2 Calculate the residual between the observed concentration value and the measured concentration value at each calibration point; S2.3. Based on the spatial coordinates of all calibration points and their corresponding residuals, a residual field covering the entire initial remote sensing density image area is constructed using a spatial interpolation algorithm. The steps include: Based on the spatial distribution density of the calibration points within the park, appropriate interpolation weights are assigned to different spatial interpolation algorithms. Based on the assigned interpolation weights, the inverse distance weighted interpolation algorithm is selected to construct the residual field; whereby the inverse distance weighted interpolation algorithm is expressed as follows:

[0033] in, The point to be interpolated The residual, It is the residual of the i-th calibration point. is the Euclidean distance from the point to be interpolated to the i-th calibration point, and p is the power parameter.

[0034] Optionally, when constructing the residual field, the present invention may also select different interpolation algorithms according to the specific spatial distribution characteristics and data properties of the calibration points, including Kriging interpolation and radial basis function interpolation; (1) Kriging interpolation method as an alternative: When the residuals of the calibration points exhibit certain structurality and autocorrelation in space, the Kriging interpolation method can be used. This method quantifies spatial correlation by constructing a variogram function, which can provide the best unbiased estimate. It is particularly suitable for situations where the calibration points are unevenly distributed and have obvious spatial trends, and can generate a residual field that is more in line with the laws of geostatistics.

[0035] (2) Radial basis function (RBF) interpolation method: For cases where the background concentration field changes smoothly and continuously, radial basis function interpolation method (such as using Gaussian function, multiple quadratic function, etc. as basis functions) can be used. This method can produce a smooth and continuous residual surface, which helps to avoid introducing unreal local mutations or "bull's eye" effects in the corrected concentration field.

[0036] S2.4. The initial remote sensing concentration image is corrected pixel-by-pixel using the constructed residual field to generate a corrected concentration field image. The steps include: The residual field is considered as a background concentration field composed of systematic bias and background disturbance; The initial remote sensing concentration image is regarded as a mixed signal image formed by superimposing the target VOCs emission plume signal with this background concentration field; By employing pixel-level differential operations, the background concentration field is extracted from the mixed signal image, revealing the target concentration image that primarily reflects the spatial distribution of the target VOCs fume. The pixel-level differential operation is expressed as follows: I target (x, y) = I initial (x, y) R (x, y); Among them, I initial (x, y) represents the initial remote sensing concentration image, R(x, y) represents the residual field, and I target (x, y) represents the extracted target concentration image; A corrected concentration field image is generated based on the target concentration image.

[0037] S3. Input the corrected concentration field image and meteorological data into the Gaussian diffusion model to simulate the diffusion trajectory of VOCs and infer the location of their emission sources. The steps include: a. Use the corrected concentration field image as the source intensity distribution input for the Gaussian diffusion model; b. Use wind speed and direction from meteorological data as the dominant driving parameters of the Gaussian diffusion model; c. In the Gaussian diffusion model, the hypothesized source location and source strength are adjusted through an iterative optimization algorithm to minimize the difference between the simulated concentration field and the corrected concentration field image. The source location corresponding to this point is then determined as the inversely derived emission source location.

[0038] In this embodiment, when multiple potential escape sources exist, the present invention adjusts the assumed source positions and source intensities through an iterative optimization algorithm to minimize the difference between the concentration field simulated by the Gaussian diffusion model and the corrected concentration field image. Based on the distribution of high-concentration regions in the corrected concentration field image, multiple candidate escape source regions are initially identified, and an initial assumed source position and source intensity are set for each candidate escape source region. The Gaussian diffusion model simultaneously optimizes the source positions and source intensities of all candidate escape sources, and the final output of the converged composite concentration field has the smallest overall difference with the corrected concentration field image.

[0039] In this embodiment, the present invention initially identifies multiple candidate emission source regions based on the distribution of high-concentration regions in the corrected concentration field image. Specifically, this includes identifying multiple candidate emission source regions from the distribution of high-concentration regions in the corrected concentration field image, performing image segmentation on the corrected concentration field image, identifying connected regions with concentration values ​​exceeding a preset threshold, calculating the centroid coordinates, area, and average concentration of each connected region, and selecting regions that meet preset conditions from all connected regions based on the area and average concentration as candidate emission source regions.

[0040] Alternatively, when identifying multiple candidate escape source regions from the corrected concentration field image, in addition to threshold-based image segmentation methods, the following alternative algorithms can be used, including: 1. Machine Learning Clustering Algorithm Alternatives: Unsupervised machine learning algorithms, such as K-means clustering or density-based DBSCAN clustering, can be used. These algorithms can automatically cluster based on pixel concentration values ​​and spatial coordinates, dividing the concentration field into different potential source regions. The DBSCAN algorithm is particularly good at identifying clusters of arbitrary shapes and can effectively eliminate noise points, making it suitable for situations where the escape sources have irregular shapes.

[0041] 2. Alternative Local Extremum (Peak) Detection Algorithm: A local maximum detection algorithm from image processing (e.g., combining Gaussian smoothing and peak finding) can be used to directly locate multiple concentration peaks in the concentration field image. The local regions where these peaks are located are then directly initialized as candidate escape source regions. This method is more direct and efficient for point source identification.

[0042] S4. Based on the location of the fugitive source and real-time wind direction data, predict the geographical area affected by the pollution plume within a preset time period in the future.

[0043] Optionally, the present invention also renders the predicted area on an electronic map as a highlighted boundary.

[0044] In this embodiment, the present invention reverse-engineers the location of potential emission sources of pollutants based on predictions within a preset time period in the future, uses this location as the release origin, and constructs a fan-shaped prediction area with the prevailing wind direction of real-time wind direction data as the axis; wherein, the angle of the fan-shaped prediction area is dynamically determined by the statistical value of historical wind direction fluctuations, and can be further fine-tuned by introducing real-time atmospheric stability parameters.

[0045] In this embodiment, the present invention also uses the Gaussian plume diffusion model as the key input parameters, including the estimated source strength of the escape source location, real-time wind speed, and terrain correction coefficient, to simulate the concentration distribution of pollutants in the fan-shaped prediction area, along the prevailing wind axis and both sides, and to calculate the movement path and diffusion range of the pollution plume in the future preset time period.

[0046] Optionally, during the Gaussian plume diffusion model simulation, real-time background concentration data synchronously accessed from the environmental monitoring network is overlaid for calculation to make the predicted values ​​closer to the real environment. Then, the concentration distribution results obtained from the simulation are compared with the environmental safety threshold to accurately define the predicted risk area. Image processing technology is then used to extract the geometric boundary of the area exceeding the standard, generating a set of high-precision geographic boundary coordinate sequences. Finally, the predicted risk area is rendered in real time on the corresponding layer of the electronic map in the form of a highlighted boundary.

[0047] Optionally, the present invention also makes a comprehensive judgment based on information such as the peak concentration, the area exceeding the standard, and whether it covers sensitive locations within the risk area, provides multi-level risk level prompts, and automatically pushes a report containing a predicted trend chart, key coordinates, and level to the monitoring terminal.

[0048] In another embodiment of the invention, in steps S3 and S4, the Gaussian diffusion model is preferred due to its high computational efficiency, but it can be replaced in specific application scenarios, including... 1. Lagrange particle model replacement: For industrial parks with complex terrain or significant building disturbance, the Lagrange particle model can be used to replace the Gaussian model. This model characterizes pollutant diffusion by simulating the random motion of a large number of virtual particles in the wind field. It can more accurately handle non-uniform flow fields and complex boundary conditions, thereby improving the source tracing accuracy in complex environments.

[0049] 2. Computational Fluid Dynamics (CFD) Model Alternative: In situations where extremely high-precision source analysis of a specific region is required, CFD models can be used for local fine simulation. CFD models can solve the three-dimensional Navier-Stokes equations and accurately reproduce the wind field and turbulence structure around the building, which can be used to verify and deepen the preliminary source tracing results of the Gaussian model.

[0050] S5. Generate the VOCs emission range based on the location of the emission source and the predicted geographical area.

[0051] In another embodiment of the invention, the invention also responds to a user's selection of a specific area on a map interface, dynamically calculates and displays statistical information on the VOCs concentration within the selected area.

[0052] In this embodiment, the present invention supports users to select a specific area on an electronic map interface, dynamically extract the concentration values ​​of all pixels in the selected area from the corrected concentration field image or real-time monitoring data, and quickly calculate its statistical information, including: average concentration, maximum concentration and concentration variance.

[0053] In another embodiment of the present invention, the present invention also extracts key information from the result database of the above steps and integrates it to generate a structured diagnostic report; wherein, the report content includes the location coordinates of the escape source, confidence assessment, concentration statistics of the user-selected area, the range of the predicted risk area, the duration of the impact, and the main affected points.

[0054] Optionally, the diagnostic report can be exported in formats such as Word and PDF, and can be automatically sent to the relevant responsible persons via email or messaging system.

[0055] like Figure 2 As shown, the present invention also proposes a VOCs emission range prediction system for a tower platform, which includes a data acquisition module, an image correction module, a location prediction module, a region prediction module, and an emission range generation module.

[0056] 1. The data acquisition module is used to acquire initial remote sensing concentration images and regional meteorological data.

[0057] In this embodiment, the data acquisition module performs horizontal or vertical scanning using an infrared remote sensing device deployed on a tower platform in the industrial park to acquire an initial remote sensing concentration image of VOCs over a large area; wherein, the initial remote sensing concentration image is a two-dimensional matrix, and each pixel contains a geographical location information and the corresponding observed VOCs concentration value. In this embodiment, the data acquisition module also collects regional meteorological data from the meteorological monitoring station at the scanning time and for a period of time before and after, including wind speed, wind direction, ambient temperature, relative humidity, and atmospheric stability parameters. 2. The image correction module is used to perform spatial domain correction on the initial remote sensing concentration image based on measured concentration data from multiple pre-deployed calibration points within the park, generating a corrected concentration field image, which includes: (1) Based on the coordinate mapping relationship, extract the observed concentration values ​​of the pixels closest to the spatial coordinates of each calibration point from the initial remote sensing concentration image; (2) Extract the observed concentration values ​​of the pixels corresponding to the spatial coordinates of each calibration point from the initial remote sensing concentration image; (3) Calculate the residual between the observed concentration value and the measured concentration value at each calibration point; (4) Based on the spatial coordinates of all calibration points and the corresponding residuals, a residual field covering the entire initial remote sensing density image area is constructed using a spatial interpolation algorithm. The steps include: Based on the spatial distribution density of the calibration points within the park, appropriate interpolation weights are assigned to different spatial interpolation algorithms. Based on the assigned interpolation weights, the inverse distance weighted interpolation algorithm is selected to construct the residual field; the formula for the inverse distance weighted interpolation algorithm is:

[0058] in, The point to be interpolated The residual, It is the residual of the i-th calibration point. is the Euclidean distance from the point to be interpolated to the i-th calibration point, and p is the power parameter.

[0059] Optionally, when constructing the residual field, the present invention may also select different interpolation algorithms according to the specific spatial distribution characteristics and data properties of the calibration points, including Kriging interpolation and radial basis function interpolation; Kriging interpolation as an alternative: When the residuals of the calibration points exhibit certain structural and autocorrelation characteristics in space, the Kriging interpolation method can be used. This method quantifies spatial correlation by constructing a variogram, providing the best unbiased estimate. It is particularly suitable for situations where the calibration points are unevenly distributed and have obvious spatial trends, and can generate a residual field that is more consistent with geostatistical laws. Radial basis function (RBF) interpolation alternative: For cases where the background concentration field changes relatively smoothly and continuously, radial basis function interpolation (such as using Gaussian functions, multiple quadratic functions, etc. as basis functions) can be used. This method can produce a smooth and continuous residual surface, which helps to avoid introducing unrealistic local abrupt changes or "bull's-eye" effects into the corrected concentration field.

[0060] (5) The initial remote sensing concentration image is corrected pixel by pixel using the constructed residual field to generate a corrected concentration field image. The steps include: The residual field is considered as a background concentration field composed of systematic bias and background disturbance; The initial remote sensing concentration image is regarded as a mixed signal image formed by superimposing the target VOCs emission plume signal with this background concentration field; By employing pixel-level differential operations, the background concentration field is extracted from the mixed signal image, revealing the target concentration image that primarily reflects the spatial distribution of the target VOCs fume. The pixel-level differential operation is expressed as follows: I target (x, y) = I initial (x, y) R (x, y); Among them, I initial (x, y) represents the initial remote sensing concentration image, R(x, y) represents the residual field, and I target (x, y) represents the extracted target concentration image; A corrected concentration field image is generated based on the target concentration image.

[0061] 3. The location prediction module is used to input the corrected concentration field image and meteorological data into the Gaussian diffusion model to simulate the diffusion trajectory of VOCs and infer the location of their emission sources. Its steps include: a. Use the corrected concentration field image as the source intensity distribution input for the Gaussian diffusion model; b. Use wind speed and direction from meteorological data as the dominant driving parameters of the Gaussian diffusion model; c. In the Gaussian diffusion model, the hypothesized source location and source strength are adjusted through an iterative optimization algorithm to minimize the difference between the simulated concentration field and the corrected concentration field image. The source location corresponding to this point is then determined as the inversely derived emission source location.

[0062] In this embodiment, when multiple potential escape sources exist, the location prediction module adjusts the assumed source locations and intensities through an iterative optimization algorithm to minimize the difference between the concentration field simulated by the Gaussian diffusion model and the corrected concentration field image. Based on the distribution of high-concentration regions in the corrected concentration field image, multiple candidate escape source regions are initially identified, and an initial assumed source location and intensity are set for each candidate escape source region. The Gaussian diffusion model simultaneously optimizes the source locations and intensities of all candidate escape sources, and the final output of the converged composite concentration field has the smallest overall difference with the corrected concentration field image.

[0063] In this embodiment, the location prediction module initially identifies multiple candidate escape source regions based on the distribution of high-concentration regions in the corrected concentration field image. Specifically, this includes identifying multiple candidate escape source regions from the distribution of high-concentration regions in the corrected concentration field image, performing image segmentation on the corrected concentration field image, identifying connected regions with concentration values ​​exceeding a preset threshold, calculating the centroid coordinates, area, and average concentration of each connected region, and selecting regions that meet preset conditions from all connected regions based on the area and average concentration as candidate escape source regions.

[0064] Alternatively, when identifying multiple candidate escape source regions from the corrected concentration field image, in addition to threshold-based image segmentation methods, the following alternative algorithms can be used, including: (1) Machine learning clustering algorithm alternative: Unsupervised machine learning algorithms, such as K-means clustering or density-based DBSCAN clustering algorithm, can be used. These algorithms can automatically cluster according to the concentration value and spatial coordinates of pixels, dividing the concentration field into different potential source regions. The DBSCAN algorithm is particularly good at identifying clusters of arbitrary shapes and can effectively eliminate noise points, making it suitable for cases where the shape of the escape source is irregular.

[0065] (2) Alternative local extremum (peak) detection algorithm: Local maximum detection algorithms in image processing (e.g., combining Gaussian smoothing and peak finding) can be used to directly locate multiple concentration peaks in the concentration field image, and the local regions where these peaks are located can be directly initialized as candidate escape source regions. This method is more direct and efficient for point source identification.

[0066] 4. The regional prediction module is used to predict the geographical area affected by the pollution plume within a preset time period in the future, based on the location of the fugitive source and real-time wind direction data.

[0067] Optionally, the region prediction module also renders the predicted geographic region on the electronic map with highlighted boundaries.

[0068] In this embodiment, the regional prediction module reverse-engineers the location of potential emission sources of pollutants based on predictions within a preset time period in the future, uses this location as the release origin, and constructs a fan-shaped prediction area with the prevailing wind direction of real-time wind direction data as the axis. The angle of this fan-shaped prediction area is dynamically determined by the statistical value of historical wind direction fluctuations, and can be further fine-tuned by introducing real-time atmospheric stability parameters.

[0069] In this embodiment, the regional prediction module also uses the Gaussian plume diffusion model as a basis, taking the estimated source strength of the escape source location, real-time wind speed and terrain correction coefficient as key input parameters to simulate the concentration distribution of pollutants in the fan-shaped prediction area, along the prevailing wind axis and both sides, and calculate the movement path and diffusion range of the pollution plume in the future preset time period.

[0070] Optionally, during the Gaussian plume diffusion model simulation, real-time background concentration data synchronously accessed from the environmental monitoring network is overlaid for calculation to make the predicted values ​​closer to the real environment. Then, the concentration distribution results obtained from the simulation are compared with the environmental safety threshold to accurately define the predicted risk area. Image processing technology is then used to extract the geometric boundary of the area exceeding the standard, generating a set of high-precision geographic boundary coordinate sequences. Finally, the predicted risk area is rendered in real time on the corresponding layer of the electronic map in the form of a highlighted boundary.

[0071] Optionally, the regional prediction module also makes a comprehensive judgment based on the peak concentration, the area exceeding the standard, and whether it covers sensitive locations within the risk area, provides multi-level risk level prompts, and automatically pushes a report containing prediction trend charts, key coordinates, and levels to the monitoring terminal.

[0072] In another embodiment of the present invention, the Gaussian diffusion model is preferred in the location prediction module and the region prediction module due to its high computational efficiency, but it can be replaced in specific application scenarios, including... (1) Lagrange particle model replacement: For industrial parks with complex terrain or significant building disturbance, the Lagrange particle model can be used to replace the Gaussian model. This model characterizes pollutant diffusion by simulating the random motion of a large number of virtual particles in the wind field. It can more accurately handle non-uniform flow fields and complex boundary conditions, thereby improving the source tracing accuracy in complex environments.

[0073] (2) Computational Fluid Dynamics (CFD) Model Alternative: In situations where extremely high-precision source analysis of a specific region is required, a CFD model can be used for local fine simulation. The CFD model can solve the three-dimensional Navier-Stokes equations and accurately reproduce the wind field and turbulent structure around the building. It can be used as a verification and deepening of the preliminary source tracing results of the Gaussian model.

[0074] 5. The fugitive range generation module is used to generate the VOCs fugitive range based on the location of the fugitive source and the predicted geographical area.

[0075] In another embodiment of the present invention, the system also responds to the user's selection of a specific area on the map interface, dynamically calculates and displays statistical information on the VOCs concentration within the selected area.

[0076] In this embodiment, the system supports users to select a specific area on the electronic map interface, dynamically extract the concentration values ​​of all pixels in the selected area from the corrected concentration field image or real-time monitoring data, and quickly calculate its statistical information, including: average concentration, maximum concentration and concentration variance.

[0077] In another embodiment of the present invention, the system also extracts key information from the result database of the above steps and integrates it to generate a structured diagnostic report; wherein the report content includes the location coordinates of the escape source, confidence assessment, concentration statistics of the user-selected area, the range of the predicted risk area, the duration of the impact, and the main affected points.

[0078] Optionally, the diagnostic report can be exported in formats such as Word and PDF, and can be automatically sent to the relevant responsible persons via email or messaging system.

[0079] Based on the above disclosure, the present invention also provides an electronic device. The electronic device of this embodiment includes at least one processor and at least one storage medium electrically connected to the processor. The storage medium is electrically connected to the processor, wherein the storage medium stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method described above.

[0080] Based on the same inventive concept, the present invention also provides a storage medium storing instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the method as described above.

[0081] The foregoing description and accompanying drawings fully illustrate embodiments of the invention to enable those skilled in the art to practice them. Other embodiments may include structural and other changes. The embodiments represent only possible variations. Individual components and functions are optional unless explicitly required, and the order of operation may vary. Some portions and features of some embodiments may be included or substituted for portions and features of other embodiments. Embodiments of the invention are not limited to the structures described above and shown in the accompanying drawings, and various modifications and changes may be made without departing from their scope. The scope of the invention is limited only by the appended claims.

Claims

1. A method for predicting the emission range of VOCs from a steel tower platform, characterized in that, The method includes, Acquire initial remote sensing concentration images and regional meteorological data; Based on the measured concentration data from multiple calibration points pre-deployed within the park, the initial remote sensing concentration image is spatially corrected to generate a corrected concentration field image. Based on the corrected concentration field image and regional meteorological data, the diffusion trajectory of VOCs was simulated using a Gaussian diffusion model, and the location of the emission source was inferred. Based on the location of the fugitive source and regional meteorological data, the geographical area affected by the pollution plume in the future within a preset time period is predicted using a Gaussian diffusion model. Based on the location of the emission source and the predicted geographical region, the emission range of VOCs is generated.

2. The method for predicting the VOCs emission range of a tower platform according to claim 1, characterized in that, The acquisition of the initial remote sensing concentration image specifically includes obtaining the initial remote sensing concentration map of VOCs by performing horizontal or vertical scanning using infrared remote sensing equipment deployed on a high tower platform in the industrial park.

3. The method for predicting the VOCs emission range of a tower platform according to claim 1, characterized in that, The generation of the corrected concentration field image specifically includes, Based on the coordinate mapping relationship, the observed concentration value of the pixel closest to the spatial coordinates of each calibration point is extracted from the initial remote sensing concentration image. Calculate the residual between the observed concentration value and the measured concentration value at each calibration point; Based on the spatial coordinates of all calibration points and their corresponding residuals, a residual field covering the entire initial remote sensing density image area is constructed using a spatial interpolation algorithm. The constructed residual field is used to correct the initial remote sensing concentration image pixel by pixel, generating a corrected concentration field image.

4. The method for predicting the VOCs emission range of a tower platform according to claim 3, characterized in that, The spatial interpolation algorithm includes: inverse distance weighted interpolation, Kriging interpolation, and radial basis function interpolation.

5. The method for predicting the VOCs emission range of a tower platform according to claim 3, characterized in that, The step of using the constructed residual field to perform pixel-by-pixel correction on the initial remote sensing concentration image to generate a corrected concentration field image specifically includes: The residual field is considered as a background concentration field composed of systematic bias and background disturbance; The initial remote sensing concentration image is regarded as a mixed signal image formed by superimposing the target VOCs emission plume signal with this background concentration field; By performing pixel-level differential operations, the background concentration field is stripped from the mixed signal image to extract the target concentration image that mainly reflects the spatial distribution of the target VOCs fume; A corrected concentration field image is generated based on the target concentration image.

6. The method for predicting the VOCs emission range of a tower platform according to claim 1, characterized in that, The simulation of VOCs diffusion trajectories and the deduction of their emission source locations specifically include, The corrected concentration field image is used as the source intensity distribution input for the Gaussian diffusion model; Wind speed and direction from meteorological data are used as the dominant driving parameters of the Gaussian diffusion model. In the Gaussian diffusion model, the hypothesized source location and source strength are adjusted through an iterative optimization algorithm to minimize the difference between the simulated concentration field and the corrected concentration field image. The source location corresponding to this point is then determined as the inversely derived emission source location.

7. The method for predicting the VOCs emission range of a tower platform according to claim 1, characterized in that, The geographical areas predicted to be affected by the pollution plume within a preset future time period specifically include, Using the location of the fugitive source as the origin of pollutant release and the prevailing wind direction of regional meteorological data as the axis, a fan-shaped prediction area is constructed. Based on the Gaussian plume diffusion model, the estimated source strength, real-time wind speed and terrain correction coefficient of the escape source location are used as input parameters to simulate the concentration distribution of pollutants along the prevailing wind axis and both sides in the fan-shaped prediction area, and to calculate the movement path and diffusion range of the pollution plume in the future preset time period. The simulated concentration distribution results are compared with environmental safety thresholds to define the predicted risk areas. Image processing techniques are then used to extract the geometric boundaries of the predicted risk areas and generate a sequence of geographic boundary coordinates.

8. A VOCs emission range prediction system for a steel tower platform, characterized in that, The system includes: a data acquisition module, an image correction module, a location prediction module, a region prediction module, and a spillover range generation module; The data acquisition module is used to acquire initial remote sensing concentration images and regional meteorological data; The image correction module is used to perform spatial domain correction on the initial remote sensing concentration image based on the measured concentration data of multiple calibration points pre-deployed in the park, and generate a corrected concentration field image. The location prediction module uses a Gaussian diffusion model to simulate the diffusion trajectory of VOCs and infer the location of the emission source based on the corrected concentration field image and regional meteorological data. The regional prediction module is used to predict the geographical area affected by the pollution plume within a preset time period based on the location of the fugitive source and regional meteorological data, using a Gaussian diffusion model. The emission range generation module is used to generate the emission range of VOCs based on the location of the emission source and the predicted geographical area.

9. The VOCs emission range prediction system for iron tower platforms according to claim 8, characterized in that, The data acquisition module is specifically used to acquire an initial remote sensing concentration map of VOCs by performing horizontal or vertical scanning using infrared remote sensing equipment deployed on a high tower platform in the industrial park.

10. The VOCs emission range prediction system for iron tower platforms according to claim 8, characterized in that, The image correction module is specifically used for, Based on the coordinate mapping relationship, the observed concentration value of the pixel closest to the spatial coordinates of each calibration point is extracted from the initial remote sensing concentration image. Calculate the residual between the observed concentration value and the measured concentration value at each calibration point; Based on the spatial coordinates of all calibration points and their corresponding residuals, a residual field covering the entire initial remote sensing density image area is constructed using a spatial interpolation algorithm. The constructed residual field is used to correct the initial remote sensing concentration image pixel by pixel, generating a corrected concentration field image.

11. The VOCs emission range prediction system for a steel tower platform according to claim 10, characterized in that, The spatial interpolation algorithm includes: inverse distance weighted interpolation, Kriging interpolation, and radial basis function interpolation.

12. The VOCs emission range prediction system for a steel tower platform according to claim 10, characterized in that, The image correction module is also used for, The residual field is considered as a background concentration field composed of systematic bias and background disturbance; The initial remote sensing concentration image is regarded as a mixed signal image formed by superimposing the target VOCs emission plume signal with this background concentration field; By performing pixel-level differential operations, the background concentration field is stripped from the mixed signal image to extract the target concentration image that mainly reflects the spatial distribution of the target VOCs fume; A corrected concentration field image is generated based on the target concentration image.

13. The VOCs emission range prediction system for iron tower platforms according to claim 8, characterized in that, The location prediction module is specifically used for, The corrected concentration field image is used as the source intensity distribution input for the Gaussian diffusion model; Wind speed and direction from meteorological data are used as the dominant driving parameters of the Gaussian diffusion model. In the Gaussian diffusion model, the hypothesized source location and source strength are adjusted through an iterative optimization algorithm to minimize the difference between the simulated concentration field and the corrected concentration field image. The source location corresponding to this point is then determined as the inversely derived emission source location.

14. The VOCs emission range prediction system for a steel tower platform according to claim 8, characterized in that, The aforementioned region prediction module is specifically used for, Using the location of the fugitive source as the origin of pollutant release and the prevailing wind direction of regional meteorological data as the axis, a fan-shaped prediction area is constructed. Based on the Gaussian plume diffusion model, the estimated source strength, real-time wind speed and terrain correction coefficient of the escape source location are used as input parameters to simulate the concentration distribution of pollutants along the prevailing wind axis and both sides in the fan-shaped prediction area, and to calculate the movement path and diffusion range of the pollution plume in the future preset time period. The simulated concentration distribution results are compared with environmental safety thresholds to define the predicted risk areas. Image processing techniques are then used to extract the geometric boundaries of the predicted risk areas and generate a sequence of geographic boundary coordinates.

15. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the VOCs emission range prediction method for any one of the tower platforms described in claims 1-7.

16. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; Memory, used to store computer programs; When the processor executes the program stored in the memory, it implements the steps of the method for predicting the VOCs emission range of the tower platform as described in any one of claims 1-7.

Images