Distributed grid atmospheric environment real-time monitoring system based on internet of things

Abstract

This invention belongs to the field of environmental monitoring technology. It discloses a distributed gridded real-time atmospheric environment monitoring system based on the Internet of Things (IoT). By constructing a micro-topographic surface roughness distribution matrix, calculating the vertical airflow disturbance coefficient and concentration gradient correction factor, and establishing a cross-node spatial concentration transfer correction model, it achieves accurate correction of the topography-height coupling effect. This invention can intelligently identify representative deviation anomaly nodes, dynamically adjust spatial interpolation weight coefficients, generate a topography-corrected three-dimensional pollutant distribution field, and optimize the monitoring grid layout based on spatial variance distribution and concentration gradient abrupt change characteristics. This invention improves the reliability of atmospheric monitoring data under complex terrain conditions and effectively eliminates systematic biases caused by non-standard installation locations.

Description

Technical Field

[0001] This invention relates to the field of environmental monitoring technology, and more specifically, to a distributed gridded real-time atmospheric environment monitoring system based on the Internet of Things. Background Technology

[0002] The current field of atmospheric environmental monitoring faces the problem of representativeness error caused by the combination of installation height and micro-topography. This problem affects the reliability and accuracy of monitoring data in practical applications. In complex urban environments, monitoring equipment is often limited by site conditions and installed on buildings, towers, or high points of terrain at different heights. These non-standardized installation heights make it impossible to directly compare the collected data. Especially in the monitoring of pollutants with significant vertical gradients (such as PM2.5 and NOx), adjacent monitoring points may show concentration changes of more than 20% simply due to differences in installation height. At the same time, the micro-topographic features around the monitoring point (such as high-rise building clusters, open grasslands, water bodies, or mountain barriers) can change the local airflow pattern and generate complex turbulence effects. These factors, coupled with the installation height, further distort the representativeness of the monitoring data. In the actual operation of monitoring networks, even if monitoring points located in urban canyons are installed at the same height, their data representativeness will be fundamentally different from those in open areas. Current technologies lack correction mechanisms that take into account this terrain-height coupling effect and cannot distinguish whether data anomalies are caused by equipment failure or special environmental factors. Furthermore, traditional monitoring networks often employ regular grid layouts or administrative divisions, neglecting the non-uniform distribution of pollutants under complex terrain conditions. This leads to both data blind spots and wasted monitoring resources. This problem is particularly pronounced in mountainous cities or coastal areas due to undulating terrain and complex airflow.

[0003] In view of this, the present invention proposes a distributed gridded real-time atmospheric environment monitoring system based on the Internet of Things to solve the above problems. Summary of the Invention

[0004] To overcome the aforementioned deficiencies of the prior art and to achieve the above objectives, the present invention provides the following technical solution: a distributed gridded real-time atmospheric environment monitoring system based on the Internet of Things, comprising:

[0005] The data acquisition module is used to acquire terrain feature data, installation height data, and measured pollutant concentration data for each monitoring node;

[0006] The analysis module is used to construct a micro-topography surface roughness distribution matrix within a preset range around each monitoring node based on terrain feature data.

[0007] The parameter calculation module is used to calculate the vertical airflow disturbance coefficient and concentration gradient correction factor of each monitoring node based on the coupling relationship between installation height data and surface roughness distribution matrix.

[0008] The model building module is used to acquire terrain elevation difference data between adjacent monitoring nodes and to establish a cross-node spatial concentration transfer correction model based on terrain elevation difference data and wind speed vector data.

[0009] The anomaly node identification module is used to identify representative deviation anomaly nodes in the measured pollutant concentration data by verifying the multipath of the cross-node spatial concentration transfer correction model.

[0010] The matrix dynamic construction module is used to dynamically construct the spatial interpolation weight coefficient matrix of the monitoring nodes based on the deviation amplitude and duration of representative deviation anomaly nodes.

[0011] The distribution field generation module is used to combine the concentration gradient correction factor and the spatial interpolation weight coefficient matrix to generate a topographically corrected three-dimensional distribution field of regional pollutant concentrations.

[0012] The mesh optimization module is used to determine the sparse optimization region and the dense optimization region of the mesh layout by analyzing the spatial variance distribution characteristics and abrupt concentration gradient regions of the three-dimensional distribution field.

[0013] The adjustment and output module is used to adjust the deployment scheme of monitoring nodes based on the location information of sparse optimization area and dense optimization area, and output real-time monitoring data after terrain-height coupling correction.

[0014] The modules are connected via wired and / or wireless means to enable data transmission between them.

[0015] The technical effects and advantages of the distributed gridded real-time atmospheric environment monitoring system based on the Internet of Things of this invention are as follows:

[0016] This invention improves the reliability and scientific rigor of atmospheric environmental monitoring data by addressing the representativeness error caused by the combination of installation height and micro-topography. It achieves intelligent correction of monitoring data under complex urban environments and variable terrain conditions, effectively eliminating systematic biases caused by non-standard installation locations, making monitoring results closer to the actual distribution of pollutants. In complex environments such as mountainous cities, coastal areas, and high-density building areas, this invention provides environmental management departments with a more representative picture of pollutant distribution. By dynamically identifying abnormal monitoring points and adaptively adjusting weights, this invention improves the accuracy and timeliness of pollution event early warnings, providing a scientific basis for emergency response to sudden pollution events. The optimized monitoring network layout breaks the limitations of traditional uniform distribution, achieving a reasonable allocation of monitoring resources to key areas, and significantly reducing construction and maintenance costs while maintaining monitoring accuracy. Attached Figure Description

[0017] Figure 1This is a schematic diagram of the Internet of Things-based distributed gridded real-time atmospheric environment monitoring system of the present invention. Detailed Implementation

[0018] 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, and 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.

[0019] This application provides an example of a distributed gridded real-time atmospheric environment monitoring system based on the Internet of Things. The implementing entities of this system include, but are not limited to, environmental monitoring centers, air pollution prevention and control platforms, smart city management systems, and environmental quality assessment systems, which can be regarded as general computing nodes of this application. The real-time monitoring system includes, but is not limited to, at least one of the following: a cloud data processing engine, a distributed sensor network, and a terrain adaptive computing unit.

[0020] Please see Figure 1 In this embodiment of the invention, the distributed grid-based real-time atmospheric environment monitoring system based on the Internet of Things includes:

[0021] The data acquisition module is used to acquire terrain feature data, installation height data, and measured pollutant concentration data for each monitoring node. Terrain feature data includes key information such as the surface undulations, building distribution, vegetation cover type, and water body distribution around the monitoring nodes, acquired in real time through a digital elevation model and a land cover database. Installation height data records the difference between the actual installation height of each monitoring node and the standard reference height, while measured pollutant concentration data includes real-time gas and particulate matter concentration values ​​collected at each monitoring point. This data provides comprehensive raw material for subsequent analysis, ensuring the adaptability and accuracy of the monitoring system.

[0022] The analysis module is used to construct a micro-topographic surface roughness distribution matrix within a preset range around each monitoring node based on terrain feature data. This matrix includes surface roughness values ​​at different azimuths around the monitoring node, reflecting the surface's resistance characteristics to airflow. Through refined analysis of the surface features around the monitoring node, the analysis module determines the surface roughness in various directions, providing fundamental parameters for subsequent airflow disturbance calculations and concentration corrections.

[0023] The parameter calculation module is used to calculate the vertical airflow disturbance coefficient and concentration gradient correction factor for each monitoring node based on the coupling relationship between installation height data and the surface roughness distribution matrix. The vertical airflow disturbance coefficient describes the turbulence intensity of the airflow under different height and roughness conditions, while the concentration gradient correction factor quantifies the impact of height changes on pollutant concentration distribution. These parameters together constitute the core indicators for terrain correction of monitoring data, providing a scientific basis for the standardization of monitoring data at different heights.

[0024] The model building module is used to acquire topographic elevation difference data between adjacent monitoring nodes. Based on the topographic elevation difference data and wind speed vector data, a cross-node spatial concentration transfer correction model is established. This model considers the impact of topographic elevation changes on pollutant transport, including the uphill deceleration effect and the downhill acceleration effect, accurately describing the spatial transfer law of pollutants under complex terrain conditions, and providing a theoretical basis for anomaly node identification and regional concentration field construction.

[0025] The anomaly node identification module is used to identify representative deviation anomaly nodes in the measured pollutant concentration data through multi-path verification of the cross-node spatial concentration transfer correction model. This module identifies monitoring points with data anomalies by comparing the deviations between theoretical concentration values ​​calculated from different paths and actual monitoring values, ensuring the reliability and representativeness of the monitoring network data and providing data quality assurance for subsequent spatial interpolation.

[0026] The matrix dynamic construction module is used to dynamically construct a spatial interpolation weight coefficient matrix for monitoring nodes based on the deviation magnitude and duration of representative deviation anomaly nodes. This matrix reflects the reliability and importance of each monitoring node in the spatial interpolation process. By reducing the weight of anomaly nodes and increasing the contribution of normal nodes, the accuracy and stability of the interpolation results are ensured.

[0027] The distribution field generation module combines the concentration gradient correction factor and the spatial interpolation weighting coefficient matrix to generate a topographically corrected three-dimensional distribution field of regional pollutant concentrations. This three-dimensional distribution field comprehensively displays the spatial distribution characteristics of pollutant concentrations within the study area, considering both the horizontal topographic influence and the vertical concentration gradient, providing an intuitive spatial perspective for pollution monitoring and early warning.

[0028] The mesh optimization module is used to determine the sparse and dense optimization zones of the mesh layout by analyzing the spatial variance distribution characteristics and abrupt concentration gradient regions of the three-dimensional distribution field. This module identifies areas that need enhanced or simplified monitoring by evaluating the effectiveness and necessity of the monitoring mesh, providing a scientific basis for the optimized design of the monitoring network.

[0029] The adjustment and output module is used to adjust the deployment scheme of monitoring nodes based on the location information of sparse and dense optimization zones, and output real-time monitoring data after terrain-height coupling correction. This module realizes dynamic optimization of the monitoring network and final correction of the data, ensuring efficient utilization of monitoring system resources and high-precision output of monitoring data.

[0030] The modules are connected via wired and / or wireless means to enable data transmission between them.

[0031] In this embodiment of the invention, the detailed implementation steps for constructing the micro-topography surface roughness distribution matrix within a preset range around each monitoring node include:

[0032] Centered on the monitoring node, a fan-shaped analysis area is divided within a preset radius, evenly distributed along the eight azimuths of the prevailing wind direction. Fan-shaped analysis is a fundamental step in the refined analysis of topographic features, dividing the environment surrounding the monitoring point into regional units that facilitate quantitative assessment. The analysis process first determines an appropriate preset radius, typically 500-1000 meters, taking into account the extent to which topography influences airflow; then, fan-shaped areas are divided along eight main directions (north, northeast, east, southeast, south, southwest, west, and northwest), with each fan covering a 45° azimuth range. This zoning method considers both the comprehensive influence of topography and the importance of the prevailing wind direction, providing a spatial framework for subsequent roughness analysis.

[0033] Within each sector analysis region, data on building height, vegetation cover type, and water body distribution are extracted to calculate the obstacle density. Obstacle density is a key indicator for quantifying surface roughness, reflecting the degree to which the surface impedes airflow. The calculation process first extracts 3D building data, vegetation classification data, and water body distribution data for each sector from the geographic information database; then, based on the characteristics of different types of obstacles, their obstruction effect on airflow is calculated; finally, the influence of various obstacles is integrated to obtain the obstacle density of the region. The obstacle density calculation uses a weighted area ratio method, i.e.:

[0034] ;

[0035] in, For a fan-shaped area Obstacle density, For the first The horizontal projected area of ​​the obstacle-like object The corresponding weighting coefficients are (0.8-1.0 for buildings, 0.5-0.7 for tall vegetation, 0.2-0.4 for low vegetation, and below 0.1 for water bodies). This represents the total area of ​​the sector.

[0036] The density value typically ranges from 0 to 1, with higher values ​​indicating a rougher surface and a stronger obstruction of airflow. Obstacle density provides a quantitative basis for subsequent roughness value determination, ensuring the objectivity and accuracy of roughness assessment.

[0037] Based on the correspondence between obstacle density and surface type, a pre-defined surface roughness standard table is consulted to obtain the initial roughness values ​​for each sector analysis area. The surface roughness standard table serves as a bridge between obstacle characteristics and aerodynamic parameters, established based on extensive wind tunnel experiments and field observation data. The query process first determines the corresponding roughness level based on the obstacle density of the sector area; then, considering the main surface cover type (such as urban built-up areas, forests, farmland, water bodies, etc.), the roughness values ​​are fine-tuned; finally, the initial roughness values ​​for each sector area are obtained. The standard table typically divides roughness values ​​into several levels between 0.001 and 3.0, with higher values ​​indicating a rougher surface and a stronger disturbance effect on airflow. This step transforms qualitative surface characteristics into quantitative aerodynamic parameters, laying the foundation for subsequent airflow disturbance calculations.

[0038] The initial roughness values ​​are weighted by distance attenuation, with obstacles closer to the monitoring node assigned higher weight coefficients. Distance attenuation weighting is a crucial step in considering the influence of spatial location, reflecting the differentiated impact of obstacles at different distances on the monitoring point. The process first subdivides the sector into three concentric regions: near-field, mid-field, and far-field. Then, based on their distance from the monitoring point, different weight coefficients are assigned to obstacles in different regions. Finally, the weighted roughness values ​​are calculated. The weight coefficients typically employ an exponential decay function to ensure that near-field obstacles receive significantly higher weights, more accurately reflecting the influence of the surrounding environment. This distance attenuation weighting process makes roughness analysis more refined and rational, improving the specificity and accuracy of surface feature characterization.

[0039] The weighted roughness values ​​of each sector are arranged by azimuth to form a surface roughness distribution matrix, which reflects the anisotropic surface characteristics around the monitoring node. The roughness distribution matrix is ​​a mathematical expression of surface characteristics, intuitively reflecting the surface roughness characteristics in different directions around the monitoring point. Matrix construction arranges the weighted roughness values ​​of each sector in azimuth order, forming an eight-dimensional vector or an 8×1 matrix, which fully records the surface heterogeneity around the monitoring point. Each element of the matrix corresponds to a roughness value in a specific azimuth, typically ranging from 0.001 to 3.0, reflecting the surface roughness in that direction. This azimuth-based matrix representation effectively captures the directional variation characteristics of surface roughness, providing a convenient data structure for subsequent parameter calculations considering wind direction effects. As a key parameter of surface-atmosphere interaction, the surface roughness distribution matrix lays the foundation for accurate calculations of atmospheric boundary layer structure and pollutant diffusion characteristics.

[0040] In this embodiment of the invention, the detailed implementation steps for calculating the vertical airflow disturbance coefficient and concentration gradient correction factor for each monitoring node based on the coupling relationship between installation height data and the surface roughness distribution matrix include:

[0041] The difference between the actual installation height of the monitoring node and the standard reference height is recorded as the height deviation. The height deviation is a fundamental parameter for monitoring height normalization, reflecting the degree of deviation between the actual monitoring point and the standard height. The acquisition process first determines the standard reference height, typically set at 10 meters above ground or a fixed height; then, the actual installation height of each monitoring node is measured, including the vertical distance of the sensor relative to the ground; finally, the difference between the two is calculated to obtain the height deviation. A positive value indicates that the monitoring point is above the standard height, and a negative value indicates that it is below the standard height; the larger the value, the more significant the deviation. The height deviation directly affects the vertical representativeness of the monitoring data and is a key input for subsequent correction calculations, ensuring the comparability and consistency of data at different heights.

[0042] Based on the current wind direction angle, the roughness value corresponding to the azimuth is extracted from the surface roughness distribution matrix and denoted as the windward roughness. Windward roughness is a crucial step in considering the influence of wind direction, reflecting the primary impact of upwind surface features on the monitoring point. The extraction process first obtains the current wind direction angle data, typically from real-time observations at the monitoring point or nearby meteorological stations; then, based on the wind direction angle, the corresponding matrix index is determined, and the roughness value for the corresponding azimuth is extracted from the surface roughness distribution matrix; if the wind direction angle falls between two azimuths, the accurate windward roughness is calculated using linear interpolation. This step transforms the approach from static surface feature description to dynamic airflow impact assessment, enabling the roughness parameter to adjust in real-time with wind direction changes, thus improving the relevance and accuracy of subsequent calculations.

[0043] Based on the ratio of altitude deviation to windward roughness, the atmospheric boundary layer thickness correction value at the monitoring node is calculated. Boundary layer thickness correction is a core step considering the surface-altitude coupling effect, quantifying the differentiated impact of surface roughness on airflow characteristics at different altitudes. The calculation process employs a power-law relationship, based on classical boundary layer theory, and considers the coupling effect between altitude deviation and roughness. The formula for calculating the correction value is:

[0044] ;

[0045] in, This is the boundary layer thickness correction value. This is the amount of high deviation. For windward roughness, This is a proportionality coefficient (usually ranging from 0.2 to 0.5). It is a power exponent (usually ranging from 0.15 to 0.3, depending on atmospheric stability).

[0046] The correction value reflects the changes in boundary layer structure under the combined effects of height and roughness, which directly affects the intensity of airflow disturbance and pollutant distribution characteristics in the vertical direction, providing a theoretical basis for subsequent calculation of the vertical disturbance coefficient.

[0047] By combining atmospheric boundary layer thickness correction values ​​with real-time temperature stratification data, the vertical perturbation coefficient of airflow is calculated using the Moning-Obukhov similarity theory. The vertical perturbation coefficient is a key parameter describing the intensity of vertical mixing and reflects the atmospheric turbulence characteristics under different conditions. The calculation process first acquires real-time temperature stratification data to determine atmospheric stability; then, the Moning-Obukhov similarity theory is applied, along with boundary layer thickness correction values, to calculate the vertical perturbation coefficient. The calculation formula is as follows:

[0048] ;

[0049] in, This is the vertical airflow disturbance coefficient. is the von Kármán constant (usually taken as 0.4). This is the friction speed (related to the roughness in the windward direction). In order to monitor altitude, The Moning-Obukhoff length (calculated from temperature stratification data). This is the stability correction function.

[0050] The unit of the disturbance coefficient is usually m² / s. The larger the value, the stronger the vertical mixing and the more significant the vertical diffusion of pollutants. The vertical disturbance coefficient of airflow directly affects the distribution characteristics of pollutants in the vertical direction. It is the theoretical basis for concentration gradient correction and an important indicator for assessing the vertical representativeness of monitoring data.

[0051] Based on the vertical airflow disturbance coefficient and the molecular diffusion coefficient of pollutants, a vertical concentration decay equation is established. A concentration gradient correction factor is extracted from this equation to normalize measured concentrations at non-standard altitudes to the standard reference altitude. Concentration gradient correction is the final step in standardizing data at different altitudes, ensuring the comparability and consistency of monitoring data. The vertical concentration decay equation considers the combined effects of airflow disturbance and molecular diffusion, conforming to classical diffusion theory. The equation is in the following form:

[0052] ;

[0053] in, For height The concentration of pollutants at that location For reference height Concentration at that location, This is the vertical airflow disturbance coefficient. This represents the dry deposition rate of pollutants (related to the molecular diffusion coefficient).

[0054] From this equation, the concentration gradient correction factor can be extracted as:

[0055] ;

[0056] Correction factor It is the height Concentration at that location converted to standard reference height The concentration multiplier, applied by the formula is: This factor takes into account the combined effects of vertical perturbation, molecular characteristics, and altitude differences. It can effectively correct for systematic biases in monitoring data at different altitudes, ensure the consistency and comparability of monitoring network data, and lay a data foundation for subsequent spatial analysis.

[0057] In this embodiment of the invention, the detailed implementation steps for establishing a cross-node spatial concentration transfer correction model based on terrain elevation difference data and wind speed vector data include:

[0058] The terrain elevation gradient between adjacent monitoring nodes is calculated, and upward and downward elevation paths are marked. Elevation gradient analysis is the first step in assessing the impact of terrain, quantifying the degree to which terrain changes affect airflow. The calculation process first acquires the three-dimensional coordinate data of adjacent monitoring nodes, including horizontal position and altitude; then, it calculates the horizontal distance and vertical height difference between the nodes; finally, it derives the elevation gradient, which is the ratio of vertical height difference to horizontal distance. Positive values ​​indicate an uphill path from the starting point to the ending point, while negative values ​​indicate a downhill path. The elevation gradient directly affects the uphill deceleration or downhill acceleration effect of airflow and is the foundational data for subsequent transport effect assessments. By clearly marking upward and downward paths, a classification basis is provided for different types of terrain effect analysis, improving the accuracy and applicability of the model.

[0059] Wind speed vector data along the path connecting adjacent monitoring nodes is acquired, and the component of the wind speed vector in the path direction is calculated, denoted as the effective transmission wind speed. Effective transmission wind speed is a key parameter for assessing pollutant transmission capacity, reflecting the actual wind intensity that promotes pollutant movement along a specific path. The calculation process first acquires wind speed vector data on or near the path, including wind speed magnitude and wind direction angle; then, the wind speed vector is projected onto the direction of the node connection to obtain the component of the wind speed in the path direction; positive values ​​indicate that the wind direction is consistent with the path direction, and negative values ​​indicate that the wind direction is opposite to the path direction. Effective transmission wind speed directly determines the transmission efficiency and direction of pollutants and is a core input parameter for constructing the transmission model. This vector projection method considers the angular relationship between wind direction and the transmission path, reflecting the actual contribution of wind to pollutant transmission more accurately than simply using wind speed magnitude.

[0060] Based on the topographic elevation gradient and effective transport wind speed, the intensity of the uphill deceleration effect or downhill acceleration effect of airflow is determined. The assessment of topographic effects is a crucial step in model construction, quantifying the impact of topographic changes on airflow and pollutant transport. The assessment process first analyzes the combined relationship between elevation gradient and effective transport wind speed; then, based on fluid dynamics principles, the intensity of the uphill deceleration or downhill acceleration effect is evaluated. Positive wind speeds produce a deceleration effect on uphill paths and an acceleration effect on downhill paths; opposing wind speeds produce the opposite effect. The intensity calculation considers the combined influence of gradient magnitude, wind speed intensity, and path length, and the relative intensity of the effect is represented by a dimensionless parameter. This physics-based assessment method ensures the model's adaptability and scientific rationality under different topographic conditions, providing a theoretical foundation for subsequent transport factor calculations.

[0061] Based on the intensity of the effect, the dilution enhancement factor or cumulative enhancement factor of pollutants during cross-node transport is calculated. Calculating the enhancement factor is a core step in quantifying the impact of topographic effects, transforming qualitative physical effects into quantitative correction parameters. The calculation process distinguishes between two different scenarios: dilution and accumulation. Deceleration during uphill driving typically leads to increased cumulative pollutant concentration, while acceleration during downhill driving leads to increased dilution. The formula for calculating the enhancement factor is:

[0062] ;

[0063] in, As an enhancing factor, For elevation gradient, To effectively transmit wind speed, This is a proportionality coefficient (adjusted according to the characteristics of the pollutants, typically ranging from 0.5 to 2.0). This is a sign function that returns the sign of the argument.

[0064] when A positive value indicates a headwind uphill or a tailwind downhill, resulting in a cumulative effect. The value is greater than 1; when A negative value indicates a dilution effect, either uphill with the wind or downhill with the wind. The value is less than 1. The enhancement factor directly affects the concentration change of pollutants during transport and is a core parameter of the cross-node transport model, providing a quantitative basis for accurately assessing the impact of topography on the spatial distribution of pollutants.

[0065] A cross-node spatial concentration transfer correction model is constructed by combining dilution enhancement factors or cumulative enhancement factors with an exponential decay function of path distance. This model characterizes the impact of topographic relief on the spatial migration of pollutants. The transfer correction model is a comprehensive expression integrating the influence of multiple factors, fully describing the spatial transport patterns of pollutants under topographic conditions. The model construction first considers the influence of enhancement factors to reflect the immediate effect of topography; then, it introduces the exponential decay of path distance to represent natural dilution during transport; finally, it combines the two effects to form a complete transfer model. The model formula is:

[0066] ;

[0067] in, End point The predicted concentration, Starting node The measured concentration, As an enhancing factor, This is the attenuation coefficient (related to the nature of pollutants and meteorological conditions). For nodes and The path distance between them.

[0068] This model considers both the local effects of topography and the general law of distance attenuation, enabling it to accurately predict the spatial transport characteristics of pollutants under complex terrain conditions. It provides a theoretical tool for identifying anomalous nodes and constructing concentration distribution fields. The model's physical mechanism is clear, its parameters are easy to obtain, and it has strong applicability, adapting to different terrain conditions and pollutant types, thus serving as a crucial support for the system's core functions.

[0069] In this embodiment of the invention, the detailed implementation steps for identifying representative deviation anomaly nodes in the measured pollutant concentration data through multi-path validation of the cross-node spatial concentration transfer correction model include:

[0070] For each monitoring node, a set of verification nodes is formed by selecting multiple neighboring nodes. Constructing the verification node set is a fundamental step in multipath verification, ensuring the comprehensiveness and representativeness of the verification. The construction process first determines an appropriate neighborhood range, typically selecting nodes within a radius of 3-5 kilometers; then, based on spatial distribution characteristics, at least 4-8 neighboring nodes in different directions are selected; finally, a verification node set is formed, covering the main directional and distance gradients around the target node. The selection of verification nodes considers both spatial representativeness, ensuring that terrain features in different directions are covered, and data quality, prioritizing nodes with stable historical performance. A well-constructed set of verification nodes provides a reliable data foundation for subsequent multipath verification and is a prerequisite for identifying anomalous nodes.

[0071] By utilizing a cross-node spatial concentration transfer correction model, the theoretical concentration value of the target monitoring node is calculated based on the measured concentration data of each node in the validation node set. The theoretical concentration estimation is the core step in the validation process, establishing a reference standard through model prediction to provide a basis for anomaly identification. The estimation process first acquires the real-time concentration data of each node in the validation node set; then, the cross-node spatial concentration transfer correction model is applied to estimate the theoretical concentration of the target node from each validation node; for each validation path, elevation changes, wind speed conditions, and distance factors are considered to calculate the enhancement factor and attenuation function; finally, the theoretical concentration value corresponding to each validation path is obtained. This multi-path estimation method utilizes comprehensive information from the surrounding environment and considers the differentiated influence of terrain features in different directions, providing a more comprehensive and reliable theoretical reference and laying the foundation for deviation analysis.

[0072] The deviation rate between the theoretical concentration value and the measured concentration value at the target monitoring node is calculated and denoted as the single-path deviation rate. Deviation rate calculation is a direct method for quantifying the degree of anomaly, providing a quantitative assessment of the anomaly of node data. The calculation process compares the estimated theoretical concentration value with the measured concentration value at the target node for each validation path, calculating the relative deviation. The formula is:

[0073] ;

[0074] in, To verify the node To the target node Single path deviation rate, For nodes The measured concentration, For node-based Estimated nodes The theoretical concentration.

[0075] A positive deviation rate indicates that the measured value is higher than the theoretical value, which may indicate a local pollution source or an overestimation of the data; a negative deviation rate indicates that the measured value is lower than the theoretical value, which may indicate a monitoring equipment malfunction or an underestimation of the data. The single-path deviation rate provides fine-grained anomaly information and is the basis for subsequent consistency analysis.

[0076] The concentration consistency index is calculated by taking the single-path deviation rate of all validation paths in the statistical validation node set and then calculating the ratio of its standard deviation to the mean. This consistency index calculation is a crucial step in assessing the reliability of anomalies, distinguishing between systematic bias and random fluctuations. The calculation process first aggregates the single-path deviation rates of all validation paths to form a deviation sample set; then, it calculates the statistical characteristics of this sample set, including the mean, standard deviation, and coefficient of variation (the ratio of standard deviation to the mean); finally, it yields the concentration consistency index, quantifying the degree of consistency between the prediction results of different paths. A smaller consistency index value indicates more consistent predictions across paths and a higher reliability of anomaly identification; a larger value indicates more dispersed prediction results, potentially influenced by complex environmental factors or model uncertainties. This index effectively distinguishes between real anomalies and model errors, improving the accuracy and reliability of anomaly identification.

[0077] When the concentration consistency index is lower than the preset consistency threshold and the absolute value of the single-path deviation rate is greater than the preset deviation threshold, the target monitoring node is marked as a representative deviation anomaly node. Anomaly node marking is the final determination in the verification process, based on multiple conditions to ensure the accuracy and reliability of the judgment. The marking process first sets reasonable threshold parameters: the consistency threshold is typically 0.3-0.5, and the deviation threshold is typically 20%-30%. Then, for each monitoring node, its consistency index and deviation magnitude are comprehensively evaluated. When both conditions are met simultaneously (low consistency index and large deviation magnitude), the node is marked as an anomaly node. This dual-condition judgment requires both high consistency of verification results to exclude false positives caused by environmental complexity and a deviation magnitude that significantly exceeds the normal fluctuation range to ensure the importance of the anomaly. The marked representative deviation anomaly nodes will undergo special processing in subsequent spatial interpolation and concentration field construction to ensure the overall reliability and representativeness of the monitoring network data.

[0078] In this embodiment of the invention, the detailed implementation steps for dynamically constructing the spatial interpolation weight coefficient matrix of monitoring nodes based on the deviation amplitude and duration of representative deviation anomaly nodes include:

[0079] For representative nodes exhibiting abnormal deviations, the proportion of times they are marked as abnormal within a preset sliding time window is recorded as the abnormality frequency. Abnormality frequency statistics are a fundamental step in assessing the persistence of abnormalities, distinguishing between temporary fluctuations and systemic problems. The statistical process first determines an appropriate sliding time window, typically 24 hours or longer, covering the complete daily variation cycle; then, it backtracks the historical marking records of nodes within the window period, calculating the number of time points marked as abnormal; finally, it calculates the abnormality frequency, which is the ratio of the number of abnormal time points to the total number of time points. The frequency value ranges from 0 to 1; a higher value indicates a more persistent abnormality and a more severe node problem. This time-window-based frequency analysis effectively distinguishes between temporary abnormalities and long-term problems, providing crucial time-dimensional information for subsequent reliability assessments and ensuring the stability and rationality of weight adjustments.

[0080] The product of the average deviation magnitude and the anomaly frequency of representative deviation anomaly nodes is calculated and denoted as the node reliability penalty factor. The penalty factor calculation is a comprehensive step in quantifying node reliability, combining the intensity and persistence of the anomaly. The calculation process first calculates the average deviation magnitude of the node during the anomaly period, reflecting the severity of the anomaly; then, the average deviation magnitude is multiplied by the anomaly frequency to obtain a comprehensive score; finally, normalization is applied to form the final penalty factor. The penalty factor considers both the magnitude (intensity dimension) and the frequency (time dimension) of the anomaly, comprehensively assessing the reliability of node data. This multi-dimensional reliability quantification method can more accurately identify and handle different types of anomaly nodes, providing a scientific basis for the fine-tuning of weighting coefficients.

[0081] For non-abnormal nodes, the degree of contamination impact is calculated based on their spatial distance from representative deviation abnormal nodes, and this is denoted as the neighborhood reliability correction factor. Neighborhood reliability correction is a crucial step in considering the impact of anomaly propagation, reflecting the spatial correlation of the anomaly effect. The calculation process first determines the attenuation model of spatial impact, typically using an exponential decay function of distance; then, the distance between each non-abnormal node and each abnormal node is calculated; finally, considering both distance and the penalty factor of abnormal nodes, the neighborhood correction factor is calculated. The correction factor ranges from 0 to 1; a lower value indicates a greater impact from the anomaly, and the greater the need for reliability adjustment. This correction method, which considers spatial correlation, effectively addresses the spatial propagation problem of anomaly impact, avoids excessive influence from nodes near the anomaly region during interpolation, and improves the accuracy and smoothness of spatial interpolation.

[0082] Based on the node reliability penalty factor and the neighborhood reliability correction factor, initial weight coefficients are assigned to each monitoring node. Initial weight assignment is the core step in constructing the interpolation matrix, directly determining the contribution of each node in spatial interpolation. The assignment process first determines the basic weight model, typically using an inverse distance weighting or Gaussian weighting function; then, it differentiates nodes based on their type: for anomalous nodes, their basic weights are reduced based on the penalty factor; for non-anomalous nodes, their basic weights are adjusted based on the neighborhood correction factor; finally, normalization ensures that the sum of all weights is 1. The initial weight assignment fully considers the reliability of the nodes themselves and the influence of the surrounding environment, effectively reducing the negative impact of anomalous nodes while maintaining the spatial continuity and smoothness of the interpolation results, providing a basic weight framework for terrain representativeness assessment.

[0083] The initial weighting coefficients are multiplied by the terrain representativeness score of the monitoring node. The terrain representativeness score is determined by the reciprocal of the terrain complexity surrounding the monitoring node, generating the final spatial interpolation weighting coefficient matrix. Terrain representativeness fusion is the final step in weight construction, considering the influence of terrain features on representativeness at the node location. The fusion process first calculates the terrain complexity of each node, typically quantified by the standard deviation of the surrounding terrain undulations or a complexity index; then, it calculates the terrain representativeness score, using the reciprocal relationship of complexity to ensure that nodes in simple terrain areas receive higher representativeness scores; finally, the initial weights are multiplied by the representativeness scores to generate the final weighting coefficient matrix. This weighting adjustment method, which considers terrain factors, effectively improves the contribution of nodes in flat and open areas, reduces the influence of nodes in complex terrain areas, better conforms to the physical laws of atmospheric diffusion, and ensures the scientific rationality and spatial representativeness of the interpolation results. The final spatial interpolation weighting coefficient matrix comprehensively integrates information from three dimensions: data reliability, spatial correlation, and terrain representativeness, providing crucial support for the construction of high-precision concentration fields.

[0084] In this embodiment of the invention, the detailed implementation steps for generating a topographically corrected three-dimensional distribution field of regional pollutant concentrations by combining a concentration gradient correction factor and a spatial interpolation weighting coefficient matrix include:

[0085] By utilizing a concentration gradient correction factor, the measured pollutant concentration data from each monitoring node are normalized to a unified standard reference height, resulting in height-normalized concentration data. Height normalization is a fundamental step in constructing comparable datasets, eliminating systematic biases caused by different installation heights. The normalization process first acquires real-time concentration data and installation height information for each monitoring node; then, it applies the concentration gradient correction factor to convert concentration data from different heights into equivalent concentrations at a standard reference height (typically 10 meters above ground); for anomalous nodes, additional adjustments or special processing can be performed to ensure the accuracy of the normalization results. This height-unified processing solves the problem of inconsistent installation heights in the monitoring network, ensuring the comparability and consistency of the spatial interpolation data basis, which is a prerequisite for constructing an accurate concentration field.

[0086] A three-dimensional mesh is constructed within the region. The mesh is adaptively divided horizontally according to the monitoring node density and vertically according to the atmospheric mixing layer height. 3D mesh construction is a crucial step in spatial discretization, providing a spatial framework for subsequent interpolation calculations. The construction process first determines the spatial extent of the study area, covering all monitoring nodes with appropriate boundaries. Then, an adaptive mesh is constructed horizontally, using a finer mesh in high-density monitoring areas and a coarser mesh in sparse areas to ensure a balance between computational efficiency and accuracy. Vertically, a stratification scheme is set according to the local atmospheric mixing layer height, typically using finer stratification near the ground. The horizontal resolution of the adaptive mesh is dynamically adjusted between 50 and 500 meters, while the vertical resolution is 1-5 meters near the ground, gradually increasing to 10-50 meters with increasing altitude. This multi-scale meshing method ensures both detailed representation of key areas and improves overall computational efficiency, providing a reasonable spatial framework for constructing large-scale, high-resolution concentration fields.

[0087] For each grid cell, a corresponding weight is extracted from the spatial interpolation weight coefficient matrix based on its 3D spatial distance to each monitoring node. Weight extraction is a preparatory step for spatial interpolation, determining the degree of influence of each monitoring point on the target grid. The extraction process first calculates the 3D Euclidean distance between the grid cell and each monitoring node; then, based on the distance and node characteristics, the basic weight is obtained from the spatial interpolation weight coefficient matrix; for anomalous nodes, an additional attenuation coefficient is applied to further reduce their influence. This weight extraction method based on spatial distance and node characteristics fully considers the dual impact of spatial relationships and data quality, ensuring a reasonable allocation of the contribution of each node during the interpolation process, laying the foundation for accurate spatial interpolation.

[0088] A three-dimensional spatial interpolation of highly normalized concentration data was performed using an inverse distance weighting method combined with a spatial interpolation weighting coefficient matrix. Spatial interpolation is a core step in constructing the concentration field, expanding discrete monitoring data into a continuous spatial distribution. The interpolation process employed a modified inverse distance weighting method, comprehensively considering spatial distance, node weights, and terrain factors. The interpolation formula is as follows:

[0089] ;

[0090] in, For grid points interpolated concentration, For monitoring nodes The overall weight (combining inverse distance ratio and weight coefficient matrix). For nodes Highly normalized concentration.

[0091] The comprehensive weight calculation takes into account the inverse relationship between spatial distance and the characteristic weights of nodes, and is in the form of:

[0092] ;

[0093] in, For nodes The weighting coefficients, For point To the node Spatial distance, This is the distance decay exponent (usually taken as 2).

[0094] This interpolation method, which combines weighted coefficients, utilizes the classic inverse distance principle while incorporating the influence of node reliability and terrain representativeness. It can more accurately reconstruct the spatial distribution of pollutants, reduce interference from outlier data and complex terrain, and improve the accuracy and reliability of the interpolation results.

[0095] By combining the interpolation results with the inverse operation of the concentration gradient correction factor corresponding to the 3D grid voxels, the true concentration values ​​of each voxel are restored, forming a 3D distribution field of regional pollutant concentrations. True concentration restoration is the final step in constructing the distribution field, recovering the actual concentration distribution at different heights. The restoration process first calculates the concentration gradient correction factor corresponding to each voxel in the 3D grid, considering the influence of vertical height and surface characteristics; then, the standard height concentration obtained from interpolation is inversely converted to restore the actual concentration values ​​at different heights; finally, the concentration data of all voxels are integrated to form a complete 3D distribution field. This height restoration process ensures the physical rationality of the 3D distribution field in the vertical direction, accurately reflecting the characteristics of pollutant variation with height, and providing comprehensive spatial data support for subsequent analysis and applications. The final generated 3D distribution field of regional pollutant concentrations comprehensively integrates information from three dimensions: monitoring data, topographic influence, and height variation, intuitively displaying the complete spatial distribution of pollutants within the study area, providing a scientific basis for pollution monitoring, source apportionment, and diffusion prediction.

[0096] In this embodiment of the invention, the detailed implementation steps for determining the sparse optimization region and the dense optimization region of the mesh layout by analyzing the spatial variance distribution characteristics and abrupt concentration gradient regions of the three-dimensional distribution field include:

[0097] Horizontal slicing is performed on the three-dimensional distribution field to extract a two-dimensional concentration distribution map of the standard reference height layer. Horizontal slicing is an effective method to simplify three-dimensional data and extract planar distribution features at key heights. The process first determines the standard reference height, typically 10 meters above the ground, consistent with the conventional monitoring height; then, data at this height layer is extracted from the three-dimensional distribution field to form a two-dimensional concentration distribution map; simultaneously, the original resolution and coordinate information are preserved to ensure the accuracy of subsequent analysis. The two-dimensional distribution map of the standard height layer retains the spatial variation characteristics in the horizontal direction while simplifying the data structure, facilitating efficient analysis and processing. This step provides a standardized data platform for spatial feature assessment, ensuring the consistency and comparability of optimized analysis.

[0098] On a two-dimensional concentration distribution map, a sliding window is used to calculate the spatial variance of concentration in each region. Spatial variance calculation is a core step in assessing regional homogeneity, quantifying the degree of fluctuation in local concentration distribution. The calculation process first determines an appropriate sliding window size, typically a 5×5 to 11×11 grid, balancing local characteristics and statistical stability. Then, the window is slid across the two-dimensional distribution map, and the statistical variance of the concentration values ​​within each window is calculated. Finally, a spatial variance distribution map is generated, visually displaying the degree of concentration fluctuation in each region. Regions with large spatial variance values ​​typically indicate uneven concentration distribution, potentially due to local pollution sources or complex terrain influences; regions with small values ​​indicate uniform concentration distribution and relatively stable environments. This spatial analysis method based on local statistics effectively identifies the heterogeneity of regional concentration distribution, providing an important basis for grid optimization.

[0099] Regions with spatial variance below a preset lower limit are marked as uniform concentration regions, corresponding to sparse optimization regions. Uniform region identification is a fundamental step in sparse optimization, determining areas where monitoring resources can be appropriately reduced. The identification process first sets a reasonable lower limit threshold for variance, typically the lower quartile of the overall variance distribution; then, on the spatial variance distribution map, all continuous regions below the threshold are marked; finally, morphological processing is performed to eliminate scattered small regions and retain the main uniform blocks. Uniform regions are characterized by slow concentration distribution changes, small spatial gradients, and a single monitoring point can represent a large area, making them suitable for sparse optimization of the monitoring grid. This adaptive identification method based on data characteristics avoids the uncertainty of subjective settings, ensures the scientific nature and data-driven nature of the optimization process, and provides an objective basis for efficient resource allocation.

[0100] The horizontal concentration gradient field of a two-dimensional concentration distribution map is calculated to identify regions where the concentration gradient value exceeds a preset gradient threshold, which are then designated as abrupt concentration gradient change regions. Gradient calculation is a crucial step in identifying regions of drastic change, capturing the spatial discontinuity of concentration distribution. The calculation process first applies a gradient operator (such as the Sobel or Prewitt operator) to the two-dimensional concentration distribution map to calculate the horizontal gradient vector for each grid point; then, the gradient magnitude is calculated to quantify the absolute magnitude of the rate of change; finally, high gradient regions are marked according to a preset threshold (usually the upper quartile of the gradient distribution). Abrupt gradient change regions typically represent concentration jumps caused by pollution fronts, emission source boundaries, or topography. These regions exhibit drastic concentration changes and require a higher density of monitoring points for accurate capture. This gradient analysis-based method directly identifies key areas of spatial change, providing precise spatial location for monitoring network optimization and ensuring sufficient monitoring density in critical areas.

[0101] Connectivity analysis was performed on regions with abrupt changes in concentration gradients to extract continuous high-gradient areas as optimization zones. These optimization zones characterize areas with drastic changes in pollutant concentrations where increased monitoring density is needed. Connectivity analysis is a crucial step in region integration, organizing discrete high-gradient points into meaningful regional units. The analysis process first applies a connected component labeling algorithm to gradient abrupt change points to identify connected regions formed by adjacent points; then, morphological processing is performed, including opening and closing operations, noise reduction, and hole filling; finally, based on area thresholds, sufficiently large continuous regions are retained and marked as optimization zones. This region extraction method, which considers spatial connectivity, avoids interference from scattered points, focuses on major areas of change, and improves the practicality and operability of the optimization scheme. The scientific identification of optimization zones ensures that monitoring resources are concentrated in the most needed areas, improving the system's responsiveness and accuracy in detecting pollution changes, and providing data support for the refined management of the monitoring network.

[0102] In this embodiment of the invention, the detailed implementation steps for adjusting the deployment scheme of monitoring nodes based on the location information of the sparse optimization region and the dense optimization region, and outputting real-time monitoring data after terrain-height coupling correction, include:

[0103] Redundancy assessment is performed on existing monitoring nodes within the sparse optimization zone. Nodes with influence below a preset sparsity threshold are selected, and a list of nodes to be adjusted by frequency reduction is generated. Redundancy assessment is the first step in optimizing node layout, identifying nodes with limited contribution for resource release. The assessment process first analyzes the influence of each node in the sparse optimization zone on the concentration field construction, typically using a removal test: temporarily removing the target node, recalculating the concentration distribution, and comparing the magnitude of change; then, based on the comparison of influence with the preset threshold, nodes with high redundancy are selected; finally, a list of nodes to be adjusted by frequency reduction is generated, including nodes recommended for removal and nodes recommended for reduced sampling frequency. This redundancy assessment method based on influence ensures the data-driven nature of the optimization process, avoids the uncertainty of subjective judgment, makes resource adjustments more accurate and reasonable, and maintains the overall performance of the monitoring network.

[0104] This method identifies monitoring blind spots within the encryption optimization zone that are more than a preset distance threshold from existing monitoring nodes. Within these blind spots, candidate locations for enhanced encryption nodes are determined based on terrain accessibility. Blind spot identification is a fundamental step in network encryption, identifying the unmonitored areas requiring enhanced monitoring. The identification process first calculates the distance from each grid point within the encryption optimization zone to the nearest monitoring node; then, it compares this distance with a preset distance threshold, marking continuous areas exceeding the threshold as monitoring blind spots; finally, candidate point analysis is performed within these blind spots, considering practical conditions such as terrain accessibility, transportation convenience, and power supply to determine feasible locations for encryption nodes. This location optimization method, which comprehensively considers spatial requirements and practical constraints, ensures both the spatial coverage of the monitoring network and the feasibility of implementation, thus improving the practical value and implementability of the optimization scheme.

[0105] Based on the node frequency reduction adjustment list and the locations of candidate encryption nodes, a deployment adjustment plan is formed that includes information on node removal, frequency reduction, and addition. Integrating this adjustment plan is a crucial step in optimization decision-making, balancing resource release and increased demand. The integration process first assesses the overall resource balance to ensure that the number of nodes removed or reduced in frequency matches the resource requirements of new nodes; then, a phased implementation plan is developed, prioritizing key areas and outstanding issues; finally, a complete adjustment plan is formed, detailing the handling method and implementation timeline for each node. This systematic approach to plan integration ensures overall coordination and efficient resource utilization during the optimization process, avoids network instability that may result from fragmented adjustments, and provides a feasible action guide for the continuous optimization of the monitoring system.

[0106] The concentration gradient correction factor and spatial interpolation weight coefficient matrix of each monitoring node are updated according to the deployment adjustment plan to generate an adjusted comprehensive correction parameter set. Parameter updating is the technical support for achieving the optimization effect, ensuring that data processing and network layout are optimized synchronously. The update process first recalculates the surrounding environmental characteristics of each node based on the adjusted node layout; then, the concentration gradient correction factor is updated to reflect the latest state of node height and surface characteristics; simultaneously, the spatial interpolation weight coefficient matrix is ​​reconstructed to adapt to the new node distribution and reliability assessment; finally, a complete correction parameter set is formed to support subsequent data processing and concentration field construction. This dynamic parameter updating mechanism ensures that the monitoring system can adapt to environmental changes and network adjustments, maintain the timeliness and accuracy of data processing, and provide technical support for high-quality monitoring data output.

[0107] A comprehensive set of correction parameters is applied to the real-time pollutant concentration data from each monitoring node to obtain preliminary corrected concentration data. Preliminary correction is a fundamental step in data processing, employing standardized parameters for system calibration. The correction process first acquires the raw real-time data from each monitoring node; then, it applies concentration gradient correction factors from the correction parameter set for high-level standardization; simultaneously, it considers the influence of terrain features and performs spatial representativeness correction; finally, it yields preliminary corrected concentration data, eliminating systemic biases caused by installation height and terrain differences. This systematic correction method ensures the comparability and consistency of data collected under different conditions, providing a standardized data foundation for subsequent quality control and data fusion.

[0108] By comparing the initially corrected concentration data with the predicted values ​​from the cross-node spatial concentration transfer correction model, outlier data points with deviations exceeding a preset threshold are identified and subjected to secondary correction. Secondary correction is a deeper step in data quality control, addressing residual anomalies and fluctuations. The correction process first applies the cross-node spatial concentration transfer correction model to extrapolate theoretical expected values ​​based on data from surrounding nodes; then, it compares this with the initially corrected data to calculate the degree of deviation; for data points with deviations exceeding the threshold, they are classified and processed according to their deviation characteristics: systematic deviations are corrected linearly, random fluctuations are smoothed, and abnormal spikes are replaced; finally, high-quality data after secondary correction is generated. This model-validated secondary correction method effectively improves data reliability and consistency, reduces the impact of abnormal fluctuations and systematic deviations, and ensures the scientific value and application credibility of the monitoring data.

[0109] A quality identification code containing the correction type and reliability level is added to the secondary corrected concentration data, forming real-time monitoring data after topographic-height coupling correction and output. Quality identification is a value-added step in data output, providing transparent processing information and reliability assessment. The identification process first determines the correction type based on the data processing history, including height correction, topographic correction, and anomaly handling; then, based on the reliability of the correction process and data consistency, the data reliability level is assessed, typically divided into high, medium, and low levels; finally, the correction type and reliability information are encoded into a quality identification code and appended to the data record. This output method with metadata improves data transparency and traceability, enabling users to select appropriate data according to their needs and quality requirements, while providing important reference information for further data analysis and application. Monitoring data that has undergone comprehensive correction and quality identification not only accurately reflects the actual state of the atmospheric environment but also provides reliable quality assurance, offering high-value data support for environmental monitoring, pollution control, and scientific research.

[0110] In this embodiment of the invention, the detailed implementation steps for distance attenuation weighting of the initial roughness value include:

[0111] Within each sector analysis region, obstacles are categorized into near-field, mid-field, and far-field obstacles based on their distance from the monitoring node. Distance partitioning is a fundamental step in refined weighting, differentiating the degree of influence based on spatial location. The partitioning process first determines reasonable distance thresholds, typically defining the 0-200 meter range centered on the monitoring node as the near field, the 200-500 meter range as the mid-field, and the 500-1000 meter range as the far field. Then, based on these thresholds, obstacles within the sector region are classified and labeled. The classification process considers the spatial distribution and continuity of obstacles to ensure the rationality and representativeness of the partitioning. This distance-based hierarchical processing method reflects the attenuation law of obstacle influence with distance, providing a spatial framework for subsequent refined weight allocation and ensuring the spatial representativeness and physical rationality of roughness calculations.

[0112] For near-field obstacles, a first weighting coefficient is assigned, determined based on the ratio of the obstacle's height to the monitoring node's installation height. Assigning near-field weights is a crucial step considering direct impact, emphasizing the key role of adjacent obstacles. The assignment process first measures the actual height and horizontal coverage of near-field obstacles; then, it calculates the ratio of the obstacle's height to the monitoring node's installation height as the basis for weight calculation; finally, a weighting function is applied to convert this ratio into a first weighting coefficient within the range of 0.7-1.0. The weighting function typically employs a piecewise linear or sigmoid function to ensure the smoothness and rationality of weight changes. This height-ratio-based weighting method fully considers the actual impact of obstacles on the airflow at the monitoring point, particularly the windbreak effect and turbulence generation, providing a physical basis for accurate roughness calculation and improving the model's adaptability to complex environments.

[0113] For obstacles in the middle field, a second weighting coefficient is assigned, which is the product of the first weighting coefficient and the distance attenuation factor. The middle field weighting calculation is an extension step that considers secondary effects, balancing obstacle characteristics and distance factors. The calculation process first determines an appropriate distance attenuation factor, typically using a linear or exponential decay function to reflect the weakening of influence with distance. Then, the attenuation factor is multiplied by the first weighting coefficient of similar obstacles in the near field to obtain the second weighting coefficient for the middle field obstacle, which usually falls within the range of 0.3-0.7. This weighting calculation method considering distance attenuation accurately reflects the influence characteristics of obstacles at medium distances, taking into account both the inherent characteristics of the obstacles and the influence of spatial position, making roughness calculations more refined and reasonable, and improving the model's adaptability and accuracy to complex environments.

[0114] For far-field obstacles, a third weighting coefficient is assigned, calculated using an exponential decay function. Determining the far-field weights is a supplementary step that considers background influence, capturing the overall contribution of the distant environment. The determination process directly applies the exponential decay function, calculating the weighting coefficients based on the actual distance from the obstacle to the monitoring point, typically in the following form:

[0115] ;

[0116] in, The third weighting coefficient for far-field obstacles. The first weighting coefficient is the baseline for the corresponding type of obstacle. The distance from the obstacle to the monitoring point. This is the attenuation coefficient (typically ranging from 0.001 to 0.005, depending on the scale and characteristics of the study area).

[0117] The third weighting coefficient typically falls within the range of 0.1-0.3, reflecting the relatively weak influence of far-field obstacles. This weighting calculation method, which employs exponential decay, conforms to the physical law that the influence of atmospheric flow decreases with distance, ensuring a reasonable contribution of the far-field environment to roughness and improving the spatial integrity and accuracy of roughness calculation.

[0118] The weighted roughness value of the sector analysis region is obtained by summing the products of the roughness contribution values ​​of obstacles at each distance and their corresponding weighting coefficients. Roughness comprehensive calculation is the final step in distance-weighted calculation, integrating the differentiated contributions of obstacles at different distances. The calculation process first determines the basic roughness contribution values ​​of various obstacles, usually obtained through table lookup or empirical formulas; then, the contribution values ​​are multiplied by the corresponding weighting coefficients to obtain the weighted contribution; finally, all weighted contributions are summed to obtain the final roughness value of the sector region. This multi-level weighted summation method comprehensively considers the differentiated influences of near-field, mid-field, and far-field obstacles, making roughness calculations more accurate and reasonable, better reflecting the surface characteristics in complex environments, and providing a reliable parameter basis for subsequent airflow disturbance and pollutant transport analysis.

[0119] In this embodiment of the invention, the detailed implementation steps for calculating the theoretical concentration value of the target monitoring node based on the measured concentration data of each node in the verification node set using the cross-node spatial concentration transfer correction model include:

[0120] For each validation node in the validation node set, the path topographic elevation difference, path distance, and wind speed vector along the path between it and the target monitoring node are obtained. Path parameter acquisition is a fundamental step in model application, collecting key physical features of the transmission path. The acquisition process first determines the precise geographic coordinates of the validation and target nodes; then, elevation information along the connecting path is extracted using a digital elevation model, and the elevation difference between the starting and ending points is calculated; simultaneously, the actual path distance between nodes is measured; finally, wind speed vector data along the path, including wind speed magnitude and wind direction angle, is obtained from meteorological stations or flow field models. These path parameters comprehensively describe the physical environment of pollutant transmission, including spatial distance, topographic relief, and aerodynamics, providing the necessary input conditions for cross-node concentration transfer calculations and ensuring the relevance and accuracy of the model application.

[0121] Based on the cross-node spatial concentration transfer correction model, the dilution enhancement factor or cumulative enhancement factor of concentration transfer from the verification node to the target monitoring node is calculated. The calculation of the transfer factor is a core step in the model application, quantifying the impact of terrain and wind speed on the transfer process. The calculation process first determines the path type, identifying uphill with the wind, uphill against the wind, downhill with the wind, or downhill against the wind based on the elevation difference and wind direction. Then, according to the corresponding physical mechanisms, the path-specific enhancement factor is calculated; uphill deceleration usually leads to a cumulative effect, with a factor greater than 1; downhill acceleration leads to a dilution effect, with a factor less than 1. The calculation considers the combined effects of elevation gradient, wind speed intensity, and path characteristics, accurately reflecting the modulating effect of terrain conditions on pollutant transfer. This physics-based transfer correction method improves the accuracy and reliability of cross-node concentration estimation, especially its applicability under complex terrain conditions, providing a scientific basis for accurately assessing the theoretical concentration of the target node.

[0122] The concentration contribution of the validation node to the target monitoring node is obtained by multiplying the measured concentration value of the validation node by the dilution enhancement factor or the cumulative enhancement factor. Concentration contribution calculation is a direct step in theoretical extrapolation, converting the measured value into the expected contribution. The calculation process first acquires the real-time measured concentration data of the validation node; then, the enhancement factor calculated in the previous step is applied for propagation correction; finally, the single-source contribution prediction of the validation node to the target node is obtained. This concise and direct calculation method transforms the complex physical process into operable numerical processing, maintaining the clarity of the physical meaning while improving the convenience and efficiency of the calculation, laying the foundation for the comprehensive evaluation of multi-source contributions.

[0123] The concentration contribution values ​​of all validation nodes in the validation node set are weighted and averaged. The weights are determined by normalizing the inverse of the path distance between the validation node and the target monitoring node. Multi-source synthesis is the final step in theoretical extrapolation, integrating the differentiated contributions of each validation node. The synthesis process first determines a reasonable weight calculation method, typically using the inverse of the distance as the basic weight to reflect the importance of near-source contributions; then, the weights are normalized to ensure the sum is 1; finally, the weighted average is calculated to obtain the final theoretical concentration of the target node. The weight calculation formula is:

[0124] ;

[0125] in, To verify the normalized weights of node i, For nodes The path distance to the target node is calculated, with the denominator being the sum of the reciprocals of the distances to all verification nodes.

[0126] This distance-based weighted averaging method reasonably balances the contribution importance of different verification nodes, ensuring the spatial representativeness and physical rationality of the theoretical calculation results, and providing a reliable reference standard for the identification of abnormal nodes.

[0127] The weighted average result is used as the theoretical concentration value for the target monitoring node, reflecting the expected concentration value extrapolated from surrounding nodes. Theoretical value confirmation is the concluding step of the verification process, establishing a benchmark for anomaly detection. The confirmation process formally defines the comprehensive result of the weighted average as the theoretical concentration value for the target node, serving as the expected value that should be observed at that location under normal circumstances. The theoretical value considers not only the measured data from each verification node, but also the influence of topography and meteorology through a transfer model, and the importance of spatial relationships through weighting. It is a reasonable expectation that integrates multiple factors, possessing high reference value and credibility. Comparative analysis of theoretical and measured concentration values ​​can effectively identify abnormal monitoring behaviors and data quality issues, providing a scientific basis for quality control and data reliability assessment of the monitoring network.

[0128] This invention achieves high-precision atmospheric environment monitoring under complex terrain conditions through data acquisition, analysis, parameter calculation, model building, anomaly identification, dynamic matrix construction, distribution field generation, grid optimization, and output adjustment. The terrain-height coupling correction method of this invention accurately considers the impact of terrain features on monitoring data, effectively identifies abnormal monitoring nodes, and provides a systematic solution for air pollution monitoring and environmental quality assessment.

[0129] The above are merely preferred embodiments 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 foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0130] It should be noted that all formulas in this manual are calculated by removing dimensions and taking their numerical values. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0131] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A distributed grid atmospheric environment real-time monitoring system based on the Internet of Things, characterized in that, include: The data acquisition module is used to acquire terrain feature data, installation height data, and measured pollutant concentration data for each monitoring node; The analysis module is used to construct a micro-topography surface roughness distribution matrix within a preset range around each monitoring node based on the terrain feature data, including: Centered on the monitoring node, a fan-shaped analysis area is divided within a preset radius. The fan-shaped analysis area is evenly distributed according to the eight azimuth angles of the prevailing wind direction. Within each sector analysis area, data on the height of surface buildings, vegetation cover type, and water body distribution are extracted, and the obstacle density of each sector analysis area is calculated. Based on the correspondence between obstacle density and surface type, a preset surface roughness standard table is queried to obtain the initial roughness value of each sector analysis area; The initial roughness value is then subjected to distance attenuation weighting. The surface roughness distribution matrix is ​​formed by arranging the weighted roughness values ​​of each sector analysis region according to the azimuth angle. The parameter calculation module is used to calculate the vertical airflow disturbance coefficient and concentration gradient correction factor for each monitoring node based on the coupling relationship between the installation height data and the surface roughness distribution matrix, including: The difference between the actual installation height of the monitoring node and the standard reference height is recorded as the height deviation. Based on the current wind direction angle, the roughness value of the corresponding azimuth is extracted from the surface roughness distribution matrix and recorded as the windward roughness. Based on the ratio of the height deviation to the windward roughness, the atmospheric boundary layer thickness correction value at the monitoring node is calculated; The vertical airflow disturbance coefficient is calculated by combining the atmospheric boundary layer thickness correction value with real-time temperature stratification data. Based on the vertical airflow disturbance coefficient and the molecular diffusion coefficient of pollutants, a vertical concentration decay equation is established, and the concentration gradient correction factor is extracted from the vertical concentration decay equation. The model building module is used to acquire terrain elevation difference data between adjacent monitoring nodes, and to establish a cross-node spatial concentration transfer correction model based on the terrain elevation difference data and wind speed vector data. An anomaly node identification module is used to identify representative deviation anomaly nodes in the measured pollutant concentration data by performing multi-path verification of the cross-node spatial concentration transfer correction model. The matrix dynamic construction module is used to dynamically construct the spatial interpolation weight coefficient matrix of the monitoring nodes based on the deviation amplitude and duration of the representative deviation anomaly nodes. The distribution field generation module is used to combine the concentration gradient correction factor and the spatial interpolation weight coefficient matrix to generate a topographically corrected three-dimensional distribution field of regional pollutant concentration. The mesh optimization module is used to determine the sparse optimization region and the dense optimization region of the mesh layout by analyzing the spatial variance distribution characteristics and concentration gradient abrupt change regions of the three-dimensional distribution field. The adjustment and output module is used to adjust the deployment scheme of monitoring nodes according to the location information of the sparse optimization area and the dense optimization area, and output the real-time monitoring data after terrain-height coupling correction.

2. The system according to claim 1, characterized in that, The step of establishing a cross-node spatial concentration transfer correction model based on the terrain elevation difference data and wind speed vector data includes: Calculate the terrain elevation gradient between adjacent monitoring nodes and mark the elevation rise path and elevation fall path; Obtain wind speed vector data along the path connecting adjacent monitoring nodes, calculate the component of the wind speed vector in the path direction, and denote it as the effective transmitted wind speed; Based on the terrain elevation gradient and the effective transmission wind speed, determine the intensity of the uphill deceleration effect or downhill acceleration effect of the airflow. Based on the intensity of the effect, calculate the dilution enhancement factor or cumulative enhancement factor of the pollutant during the cross-node transport process; The cross-node spatial concentration transfer correction model is constructed by combining the dilution enhancement factor or the cumulative enhancement factor with the exponential decay function of the path distance.

3. The system according to claim 1, characterized in that, The process of identifying representative deviation anomaly nodes in the measured pollutant concentration data through multi-path validation of the cross-node spatial concentration transfer correction model includes: For each monitoring node, select multiple neighboring nodes around it to form a set of verification nodes; Using the cross-node spatial concentration transfer correction model, the theoretical concentration value of the target monitoring node is calculated based on the measured concentration data of each node in the verification node set. Calculate the deviation rate between the theoretical concentration value and the measured concentration value at the target monitoring node, and record it as the single-path deviation rate; The single-path deviation rate corresponding to all verification paths in the set of verification nodes is statistically analyzed, and the ratio of its standard deviation to the mean is calculated and denoted as the concentration consistency index. When the concentration consistency index is less than the preset consistency threshold and the absolute value of the single path deviation rate is greater than the preset deviation threshold, the target monitoring node is marked as the representative deviation abnormal node.

4. The system according to claim 1, characterized in that, The method of dynamically constructing a spatial interpolation weight coefficient matrix for monitoring nodes based on the deviation amplitude and duration of the representative deviation anomaly nodes includes: For the representative deviation abnormal nodes, the proportion of times they are marked as abnormal within a preset sliding time window is counted and recorded as the abnormal frequency. The product of the average deviation amplitude and the abnormal frequency of the representative deviation abnormal nodes is calculated and denoted as the node reliability penalty factor. For non-abnormal nodes, the degree of contamination is calculated based on their spatial distance from the representative deviation abnormal nodes, and is denoted as the neighborhood reliability correction factor. Based on the node reliability penalty factor and the neighborhood reliability correction factor, an initial weight coefficient is assigned to each monitoring node; The initial weight coefficients are multiplied by the terrain representativeness score of the monitoring node, which is determined by the reciprocal of the terrain complexity around the monitoring node, to generate the final spatial interpolation weight coefficient matrix.

5. The system according to claim 1, characterized in that, The step of combining the concentration gradient correction factor with the spatial interpolation weighting coefficient matrix to generate a terrain-corrected three-dimensional distribution field of regional pollutant concentrations includes: Using the concentration gradient correction factor, the measured pollutant concentration data of each monitoring node are normalized to a unified standard reference height to obtain highly normalized concentration data. A three-dimensional mesh element is established within the region. For each mesh element, the corresponding weight is extracted from the spatial interpolation weight coefficient matrix based on its three-dimensional spatial distance from each monitoring node. The distance-inverse weighting method combined with the spatial interpolation weight coefficient matrix is ​​used to perform three-dimensional spatial interpolation on the highly normalized concentration data. The interpolation results are combined with the inverse operation of the concentration gradient correction factor corresponding to the three-dimensional grid element to restore the true concentration value of each element, thus forming a three-dimensional distribution field of pollutant concentration in the region.

6. The system according to claim 1, characterized in that, The process of analyzing the spatial variance distribution characteristics and abrupt concentration gradient regions of the three-dimensional distribution field to determine the sparse optimization region and the dense optimization region of the mesh layout includes: The three-dimensional distribution field is horizontally sliced ​​to extract a two-dimensional concentration distribution map of the standard reference height layer; On the two-dimensional concentration distribution map, the spatial variance of concentration in each region is calculated using a sliding window. The region whose concentration spatial variance is less than a preset lower limit of variance is marked as a uniform concentration region, and the uniform concentration region corresponds to the sparse optimization region. Calculate the horizontal concentration gradient field of the two-dimensional concentration distribution map, identify regions where the concentration gradient value is greater than a preset gradient threshold, and record them as the concentration gradient abrupt change regions. Connectivity analysis is performed on the concentration gradient abrupt change region, and continuous high gradient regions are extracted as the encryption optimization region.

7. The system according to claim 1, characterized in that, The step of adjusting the deployment scheme of monitoring nodes based on the location information of the sparse optimization region and the dense optimization region, and outputting real-time monitoring data after terrain-height coupling correction, includes: Redundancy assessment is performed on existing monitoring nodes in the sparse optimization region, nodes with influence below the preset sparse threshold are selected, and a list of nodes for frequency reduction adjustment is generated. Identify monitoring blind spots within the encryption optimization zone that are more than a preset distance threshold from existing monitoring nodes, and determine the locations of candidate encryption nodes within the monitoring blind spots based on terrain accessibility. Based on the node frequency reduction adjustment list and the candidate encryption node positions, a deployment adjustment scheme is formed that includes node removal, frequency reduction and addition information; The concentration gradient correction factor and the spatial interpolation weight coefficient matrix of each monitoring node are updated according to the deployment adjustment scheme to generate the adjusted comprehensive correction parameter set. The comprehensive correction parameter set is applied to the real-time pollutant concentration data of each monitoring node to obtain preliminary corrected concentration data; By comparing the preliminary corrected concentration data with the predicted values ​​of the cross-node spatial concentration transfer correction model, abnormal data points with deviations exceeding a preset threshold are identified and secondary corrections are performed. A quality identification code containing the correction type and reliability level is added to the concentration data after secondary correction, forming real-time monitoring data after terrain-height coupling correction and outputting it.

8. The system according to claim 3, characterized in that, The step of using the cross-node spatial concentration transfer correction model to calculate the theoretical concentration value of the target monitoring node based on the measured concentration data of each node in the verification node set includes: For each verification node in the set of verification nodes, obtain the path terrain elevation difference, path distance, and wind speed vector along the path between it and the target monitoring node; Based on the cross-node spatial concentration transfer correction model, calculate the dilution enhancement factor or cumulative enhancement factor when the concentration of the verification node is transferred to the target monitoring node. Multiply the measured concentration value of the verification node by the dilution enhancement factor or the cumulative enhancement factor to obtain the concentration contribution value of the verification node to the target monitoring node. The concentration contribution values ​​of all verification nodes in the verification node set are weighted and averaged. The weighted average result is used as the theoretical concentration value of the target monitoring node.

Images