A Geographically Weighted Regression Modeling Method and Device for Spatial Filter Value of CO2 Concentration

Abstract

This invention provides a geographically weighted regression modeling method and apparatus for CO2 concentration spatial filtration. The method includes: constructing a spatial weight matrix based on the geographical locations of carbon dioxide concentration monitoring points, and extracting feature vectors representing global locational relationships from the spatial weight matrix as global spatial feature factors; based on a preset model reflecting local heterogeneity, selecting feature vectors representing local spatial heterogeneity from the global spatial feature factors as local spatial factors; and constructing a carbon dioxide concentration model based on the local spatial factors. This invention constructs a CO2 concentration model by selecting feature vectors representing local heterogeneity from global spatial feature factors representing global spatial relationships as local spatial factors, enabling the constructed model to simultaneously consider both global spatial features and local spatial heterogeneity.

Description

Technical Field

[0001] This invention relates to the field of spatial statistical analysis technology, and in particular to a geographically weighted regression modeling method and apparatus for CO2 concentration spatial filtration values. Background Technology

[0002] Carbon dioxide (CO2), as a major greenhouse gas, has a profound impact on human daily production and life, and plays a key role in global climate change and the greenhouse effect.

[0003] Carbon dioxide exhibits certain patterns in its spatial distribution, influenced by spatial effects, specifically spatial autocorrelation and spatial heterogeneity. The distribution of CO2 is continuous, implying a certain degree of spatial correlation; that is, the CO2 concentration in one region may be similar to that in its neighboring regions. Simultaneously, CO2 distribution is also influenced by various factors, such as topography, meteorological conditions, and carbon emission sources. These factors vary across different geographical locations, leading to spatial heterogeneity in CO2 distribution.

[0004] Traditional CO2 concentration regression models typically consider only one aspect—global spatial autocorrelation or local spatial heterogeneity—rather than both. Therefore, existing CO2 concentration regression models have certain limitations. Summary of the Invention

[0005] This invention provides a geographically weighted regression modeling method and apparatus for CO2 concentration spatial filter values, which addresses the shortcomings of existing CO2 concentration regression models that typically only consider global or local spatial effects. It achieves a CO2 concentration spatial filter value geographically weighted regression modeling method and apparatus that takes into account both global spatial autocorrelation and local spatial heterogeneity.

[0006] This invention provides a geographically weighted regression modeling method for CO2 concentration spatial filter values, comprising:

[0007] A spatial weight matrix is ​​constructed based on the geographical location of carbon dioxide concentration monitoring points, and feature vectors representing global positional relationships are extracted from the spatial weight matrix as global spatial feature factors.

[0008] Based on a pre-defined model of local heterogeneity, feature vectors characterizing local spatial heterogeneity are selected from the global spatial feature factors and used as local spatial factors.

[0009] A carbon dioxide concentration model is constructed based on the aforementioned local spatial factors.

[0010] According to the geographically weighted regression modeling method for CO2 concentration spatial filtration value provided by the present invention, the step of selecting feature vectors characterizing spatial local heterogeneity as local spatial factors from the global spatial feature factors based on a preset model of local heterogeneity specifically includes:

[0011] The global spatial feature factors are used as independent variables of the preset model, and the preset model is calculated in combination with the pre-acquired influence factor data;

[0012] Calculate the correction determination coefficient corresponding to each of the global spatial feature factors;

[0013] If the calculated correction determination coefficient is greater than the current maximum correction determination coefficient, the global spatial feature factor corresponding to the current correction determination coefficient is retained, and the current correction determination coefficient is updated to the current maximum correction determination coefficient; otherwise, the global spatial feature factor corresponding to the current correction determination coefficient is removed, wherein the initial value of the current maximum correction determination coefficient is set to zero.

[0014] The global spatial feature factors retained after iterative calculation are used as the local spatial factors.

[0015] According to the geographically weighted regression modeling method for CO2 concentration spatial filter value provided by the present invention, the step of using the global spatial feature factor as the independent variable of the preset model and combining it with pre-acquired influencing factor data to calculate the preset model specifically includes:

[0016] All the aforementioned global spatial feature factors are randomly sorted and then added to the spatial feature vector set;

[0017] The global spatial feature factors in the spatial feature vector set are sequentially added to the independent variable set of the preset model, and used as independent variables of the preset model. Combined with the pre-acquired influence factor data, the preset model is calculated.

[0018] According to the present invention, a geographically weighted regression modeling method for spatial filtration value of CO2 concentration is provided, wherein the preset model is a geographically weighted regression model, a spatial error model, or a spatial lag model.

[0019] According to the geographically weighted regression modeling method for CO2 concentration spatial filter value provided by the present invention, after the step of constructing a carbon dioxide concentration model based on the local spatial factors, the method further includes:

[0020] A grid calculation was performed on the carbon dioxide concentration model to obtain a carbon dioxide concentration distribution map;

[0021] Outliers in the carbon dioxide concentration distribution map are processed.

[0022] According to the geographically weighted regression modeling method for CO2 concentration spatial filter values ​​provided by the present invention, the step of constructing a spatial weight matrix based on the geographical locations of carbon dioxide concentration monitoring points includes:

[0023] A spatial weight matrix is ​​constructed based on the adjacency relationship, inverse distance weighting, or a combination of adjacency relationship and inverse distance weighting, combined with the geographical location of the carbon dioxide concentration data monitoring points.

[0024] According to the geographically weighted regression modeling method for CO2 concentration spatial filtration value provided by the present invention, before the step of selecting feature vectors characterizing spatial local heterogeneity as local spatial factors from the global spatial feature factors based on the preset model of local response heterogeneity, the method further includes:

[0025] Obtain carbon dioxide concentration data for a preset area;

[0026] Surface factors, atmospheric factors, human activity factors, and radiation factors are used as influencing factors to obtain influencing factor data for a preset area;

[0027] The carbon dioxide concentration data and the influencing factor data are preprocessed.

[0028] According to the geographically weighted regression modeling method for CO2 concentration spatial filtration value provided by the present invention, before the step of selecting feature vectors characterizing spatial local heterogeneity as local spatial factors from the global spatial feature factors based on the preset model of local response heterogeneity, the method further includes:

[0029] Calculate the variance inflation factor between each influencing factor data and the carbon dioxide concentration data. If the calculated variance inflation factor is greater than a preset threshold, remove the influencing factor corresponding to the current largest variance inflation factor.

[0030] Repeat the above steps until the variance inflation factor corresponding to each influencing factor after screening is less than the preset threshold.

[0031] This invention also provides a geographically weighted regression modeling method for CO2 concentration spatial filtration values, comprising:

[0032] The global module is used to construct a spatial weight matrix based on the geographical location of carbon dioxide concentration monitoring points, and extract feature vectors representing global positional relationships from the spatial weight matrix as global spatial feature factors.

[0033] The local module is used to select feature vectors representing local spatial heterogeneity from the global spatial feature factors based on a preset model that reflects local heterogeneity.

[0034] A construction module is used to construct a carbon dioxide concentration model based on the local spatial factors.

[0035] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the geographic weighted regression modeling method for spatial filter values ​​of CO2 concentration as described above.

[0036] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the geographic weighted regression modeling method for spatial filter values ​​of CO2 concentration as described above.

[0037] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the geographic weighted regression modeling method for spatial filter values ​​of CO2 concentration as described above.

[0038] The CO2 concentration spatial filter value geographic weighted regression modeling method and apparatus provided by the present invention extracts global spatial feature factors by constructing a global spatial matrix, and extracts local spatial factors that take into account both global and local features from the global spatial feature factors. These factors are then combined with factor data that affect CO2 concentration to perform geographic weighted regression modeling, resulting in a CO2 concentration estimation model. This model can take into account both global spatial features and local spatial heterogeneity. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0040] Figure 1 This is one of the flowcharts of the spatial filter value geographic weighted regression modeling method for CO2 concentration provided by the present invention;

[0041] Figure 2 This is a flowchart used to illustrate the global spatial feature factor screening process in the CO2 concentration spatial filter value geographic weighted regression modeling method provided by this invention;

[0042] Figure 3 This is a flowchart used to illustrate the process of screening influencing factors in the spatial filter value geographic weighted regression modeling method for CO2 concentration provided by this invention;

[0043] Figure 4 This is the second flowchart of the CO2 concentration spatial filter value geographic weighted regression modeling method provided by the present invention;

[0044] Figure 5 This is a schematic diagram of the CO2 spatial filter value geographic weighted regression modeling device provided by the present invention;

[0045] Figure 6 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0047] The following is combined with Figures 1 to 4 This invention introduces a geographically weighted regression modeling method for CO2 concentration spatial filtration values, such as... Figure 1 As shown, it includes:

[0048] Step 101: Construct a spatial weight matrix based on the geographical location of carbon dioxide concentration monitoring points, and extract feature vectors representing global location relationships from the spatial weight matrix as global spatial feature factors.

[0049] Traditional CO2 concentration regression models typically consider only one aspect: global spatial autocorrelation or local spatial heterogeneity, without taking both into account simultaneously. In other words, when processing spatial data, these models either consider the global correlation between data points or the specific heterogeneity of local regions, but do not combine the two, thus failing to fully capture the complex spatial distribution characteristics of CO2.

[0050] To balance the impact of global spatial autocorrelation and local spatial heterogeneity on CO2 concentration regression models, this invention proposes a modeling method based on eigenvector spatial filtering-based geographically weighted regression model (ESF-GWR). It posits that the spatial influence on CO2 concentration can be divided into two parts: one is the spatial influence caused by local environmental conditions themselves, and the other is the global spatial influence and its local manifestation.

[0051] Therefore, it is first necessary to extract global spatial feature factors that characterize the global positional relationship of CO2.

[0052] Specifically, a spatial weight matrix is ​​constructed based on the geographical location of CO2 concentration monitoring points, and global spatial feature factors are then selected from the eigenvectors of the spatial weight matrix.

[0053] The geographical location of the CO2 concentration data monitoring point is determined based on the method of obtaining the CO2 concentration data.

[0054] Optionally, when the CO2 concentration data comes from CO2 concentration monitoring stations established on the ground, the coordinates of each station in the study area are used as the geographical locations of the CO2 concentration data monitoring points.

[0055] Furthermore, due to the limited number of ground-based CO2 concentration monitoring stations and their uneven geographical distribution, the development of remote sensing technology has enabled the use of data collected by carbon monitoring satellites and other remote sensing methods to estimate ground-based CO2 concentrations. Carbon monitoring satellites are categorized into two types: active satellites and passive satellites.

[0056] Optionally, when the CO2 concentration data comes from a passive satellite, the geographical location corresponding to the data received by the passive satellite in the study area is used as the geographical location of the CO2 concentration data monitoring point.

[0057] Optionally, when the CO2 concentration data comes from active satellites, the coordinates of the active satellite scan points are used as the geographical locations of the CO2 concentration data monitoring points.

[0058] Compared to passive satellites, active satellites have significant advantages in data accuracy and spatial coverage. Furthermore, active satellite data consists of scanned data with clear spatiotemporal coordinates. Therefore, CO2 concentration data is preferably acquired by active satellites. In this embodiment, CO2 concentration data acquired by the active satellite DQ-1 will be used as an example for explanation.

[0059] It is understood that, in other feasible implementations, the above methods can also be combined to obtain CO2 concentration data. For example, CO2 concentration data can be obtained by using both active satellites and ground-based CO2 concentration monitoring stations. In this case, the scanning location of the active satellites and the coordinates of the CO2 concentration monitoring stations within the study area are used together as the geographical location of the CO2 concentration data monitoring points.

[0060] Based on this, in this embodiment, a spatial weight matrix W1 is constructed according to the projected coordinates of the active satellite scanning data, and then centered to form a symmetric matrix W:

[0061]

[0062] In the formula, I represents an n-dimensional identity matrix, 11 TThis represents an n×n matrix where all elements are equal to 1, and n represents the number of CO2 concentration monitoring points within the study area, which in this embodiment is the number of active satellite scanning points within the study area.

[0063] Furthermore, based on the eigenvalues ​​and eigenvectors of the symmetric matrix W, several global spatial feature factors that best fit the OLS (Ordinary Least Squares) model are selected from the eigenvectors of the symmetric matrix W.

[0064] In one specific implementation, the eigenvalues ​​and eigenvectors of the symmetric matrix W are first calculated, and an initial screening is performed. The eigenvalues ​​are then solved to obtain the eigenvalue λ. i and eigenvector E i There are two conditions for initial screening: one is its eigenvalue λ. i >0; secondly, according to the empirical model, the following conditions must be met:

[0065]

[0066] In the formula, λ i λ represents the feature value to be selected. max This represents the largest eigenvalue, with 0.25 being a set empirical value. This initially selects eigenvectors that meet the criteria as spatial feature vectors.

[0067] Based on this, the spatial feature vectors are further screened, specifically by using stepwise regression to select the spatial feature vectors that yield the best evaluation results as global spatial feature factors.

[0068] Specifically, such as Figure 2 As shown, the spatial feature vectors are arranged in ascending order of their eigenvalues ​​to obtain a spatial feature vector set E. The feature vectors in the spatial feature vector set E are then added sequentially to the independent variable set of the OLS model. Least square regression is performed on the residuals of the OLS model, and the AIC value (Akaike Information Criterion) of the regression equation is calculated.

[0069] Then perform the screening step: select the regression equation Y with the smallest AIC value. K Its corresponding spatial feature vector is e k Then, from the spatial feature vector set E, divide by e k The external eigenvectors are successively added to the regression equation Y. K And calculate the AIC value of the regression equation.

[0070] Determine if the AIC value decreases significantly. If it does, continue the screening process; if it barely decreases, stop adding eigenvectors to the regression equation and record the eigenvectors already added. These eigenvectors are the set of eigenvectors that optimize the evaluation criteria, which are the global spatial feature factors obtained through screening.

[0071] Step 102: Based on the preset model of local heterogeneity, select the feature vectors that characterize local spatial heterogeneity from the global spatial feature factors as local spatial factors.

[0072] Furthermore, in order to comprehensively consider the impact of local spatial heterogeneity on CO2 concentration, feature vectors characterizing local spatial heterogeneity are selected from global spatial feature factors as local spatial factors, which serve as the basis for the final construction of CO2 concentration.

[0073] Based on this, in order to find suitable local spatial factors among global spatial feature factors, global spatial feature factors are screened based on a preset model.

[0074] Optionally, the preset model can be a geographically weighted regression model, a spatial error model, a spatial lag model, or other models that can reflect local heterogeneity.

[0075] In this embodiment, a geographically weighted regression model is used as an example to illustrate the spatial heterogeneity by using spatially variable coefficients, and to further screen suitable local spatial factors.

[0076] Step 103: Construct a carbon dioxide concentration model based on the local spatial factor.

[0077] After selecting the local spatial factors, a CO2 concentration estimation model (ESF-GWR) that takes into account both global spatial autocorrelation and local heterogeneity can be constructed, as follows:

[0078]

[0079] In the formula, y i This represents the CO2 concentration corresponding to the i-th sampling point, (u i ,v i Let be the coordinates of the i-th sampling point, and β0 be a constant. k (u i ,v i Let x be the k-th regression parameter at the i-th sampling point. ik Let Eα be the value of the independent variable at the i-th sampling point. ik ε is the weight value of the k-th spatial feature at the i-th sampling point, which is calculated based on the local spatial factor. iLet be the random error of the i-th sampling point, satisfying the basic assumptions of zero mean, homoscedasticity, and mutual independence. p represents the number of regression parameters, and q represents the sum of the number of regression parameters and the number of local spatial factors.

[0080] This invention extracts global spatial feature factors by constructing a global spatial matrix, and then extracts local spatial factors that take into account both global and local features from the global spatial feature factors. These local spatial factors are then combined with the factor data that affect CO2 concentration to perform geographically weighted regression modeling, resulting in a CO2 concentration estimation model. This model can take into account both global spatial features and local spatial heterogeneity.

[0081] In the geographically weighted regression modeling method for CO2 concentration spatial filtration value of the present invention, the step of selecting feature vectors representing local spatial heterogeneity from the global spatial feature factors as local spatial factors based on the preset model of local response heterogeneity specifically includes:

[0082] The global spatial feature factors are used as independent variables of the preset model, and the preset model is calculated in combination with the pre-acquired influence factor data;

[0083] Calculate the correction determination coefficient corresponding to each of the global spatial feature factors;

[0084] This embodiment uses a geographic weighted regression model as the preset model for explanation. The selected global spatial feature factors are placed into a dataset in the order of selection to construct a feature vector set E1. The feature vectors in E1 are then added to the independent variables of the geographic weighted regression model in sequence, and geographic weighted regression is calculated together with the pre-acquired influence factor data.

[0085] For each global spatial feature factor, calculate the corrected coefficient of determination (Adj.R.) of its corresponding regression equation. 2 ), which serves as the correction determination coefficient corresponding to the global spatial feature factor.

[0086] Among them, the influencing factors characterize the relevant factors that affect CO2 concentration. These can include surface factors such as air temperature, geothermal flux, and land cover, atmospheric factors such as air pressure and wind speed, human activity factors, and radiation factors. The specific factors can be selected and determined based on the geographical location and surrounding environment of the study area.

[0087] Data on the selected influencing factors were collected in advance and preprocessed together with the acquired CO2 concentration data to obtain the influencing factor data used for calculation in the geographically weighted regression model.

[0088] Understandably, the selection of influencing factors and the model calculation method can be set accordingly depending on the preset model used.

[0089] If the calculated correction determination coefficient is greater than the current maximum correction determination coefficient, the global spatial feature factor corresponding to the current correction determination coefficient is retained, and the current correction determination coefficient is updated to the current maximum correction determination coefficient; otherwise, the global spatial feature factor corresponding to the current correction determination coefficient is removed, wherein the initial value of the current maximum correction determination coefficient is set to zero.

[0090] The global spatial feature factors retained after iterative calculation are used as the local spatial factors.

[0091] Furthermore, the optimal global spatial feature factor is found through iterative comparison as the local spatial factor.

[0092] Specifically, the evaluation indicators will first be initialized, which means... The value is set to 0.

[0093] When using global spatial feature factors as independent variables in geographic weighted regression calculations, their corresponding Adj.R. 2 With the current largest correction determination coefficient The values ​​are compared.

[0094] If the current global spatial feature factor Adj.R 2 greater than If the value is found, the current global spatial feature factor is saved and updated. The value of .

[0095] Otherwise, remove the current global spatial feature factor. Repeat the above steps until all feature vectors in E1 are traversed. Finally, use the saved feature vectors as the filtered local spatial factors that can reflect the local heterogeneity of space.

[0096] In the geographically weighted regression modeling method for CO2 concentration spatial filtration value of the present invention, the step of using the global spatial feature factor as the independent variable of the preset model and combining it with pre-acquired influencing factor data to calculate the preset model specifically includes:

[0097] All the aforementioned global spatial feature factors are randomly sorted and then added to the spatial feature vector set;

[0098] The global spatial feature factors in the spatial feature vector set are sequentially added to the independent variable set of the preset model, and used as independent variables of the preset model. Combined with the pre-acquired influence factor data, the preset model is calculated.

[0099] Due to the cyclical comparison process, The value is based on the current global spatial feature factor Adj.R 2The values ​​are updated in real time; therefore, the order of global spatial feature factors will have a certain impact on the screening results.

[0100] Therefore, before the cyclic comparison step, the global spatial feature factors are first randomly sorted, and then added to a dataset in the order of random sorting to construct a spatial feature vector set E2. When performing the cyclic comparison step, a feature vector is selected from E2 for calculation in turn.

[0101] In the spatial filtration value geographic weighted regression modeling method for CO2 concentration of the present invention, after the step of constructing a carbon dioxide concentration model based on the local spatial factors, the method further includes:

[0102] A grid calculation was performed on the carbon dioxide concentration model to obtain a carbon dioxide concentration distribution map;

[0103] When obtaining the CO2 concentration estimation model, the correlation coefficients of various factors at various locations throughout the study area were also obtained.

[0104] Each location point in the study area is converted into a raster image, and the raster values ​​are set as the coefficient values ​​of each factor. The extracted spatial feature vectors are interpolated into a raster image covering the entire study area with the same resolution as the variables. Then, raster calculations are performed, multiplying the factor raster image by the factor correlation coefficient and summing the results. Finally, the intercept term is added to obtain the estimated CO2 concentration, resulting in a carbon dioxide concentration distribution map.

[0105] Outliers in the carbon dioxide concentration distribution map are processed.

[0106] Due to errors, a small number of areas may have negative CO2 concentrations in the obtained carbon dioxide concentration distribution map. There may also be a few abnormally high CO2 concentrations, which need to be processed. A negative concentration value indicates a low CO2 concentration at that point, so the CO2 concentration value is set to 0. A threshold is also set for the CO2 concentration values; any CO2 concentration exceeding this threshold is adjusted to that threshold and then visualized to obtain a continuous ground-level CO2 concentration distribution map.

[0107] In the CO2 concentration spatial filter value geographic weighted regression modeling method of the present invention, the step of constructing a spatial weight matrix based on the geographic location of carbon dioxide concentration monitoring points includes:

[0108] A spatial weight matrix is ​​constructed based on the adjacency relationship, inverse distance weighting, or a combination of adjacency relationship and inverse distance weighting, combined with the geographical location of the carbon dioxide concentration data monitoring points.

[0109] Optionally, a spatial weight matrix can be constructed based on adjacency, inverse distance weighting, or a combination of adjacency and inverse distance weighting.

[0110] In one feasible implementation, a spatial weight matrix W1 is constructed based on adjacency relationships. First, the Thiessen polygons of multiple CO2 concentration monitoring points within the study area are used. Furthermore, a global spatial weight matrix can be constructed based on Bishop adjacency, Rook adjacency, or Queen adjacency methods.

[0111] In another feasible implementation, a spatial weight matrix W1 is constructed based on inverse distance weighting. First, adjacency relationships are established based on the distances between multiple CO2 concentration monitoring points within the study area. Then, a kernel function is used to convert the distances into weight values, resulting in the global spatial weight matrix. Kernel functions can be categorized into exponential, Gaussian, and spherical models, with specific formulas as follows:

[0112]

[0113] In the formula, i and j represent position points i and j, respectively, and W i,j The adjacency (weight) between locations i and j is represented by r, and r represents the maximum distance between edges in the minimum spanning tree of all monitoring points.

[0114] In another feasible implementation, a composite approach is used, such as a K-nearest neighbor matrix and a spatial weight matrix W1 constructed based on adjacency and distance. First, a threshold K is determined, and then the K nearest neighbor monitoring points of each monitoring point are searched to construct an adjacency matrix W1. i,j The adjacency (weight) between locations i and j is represented; a weight matrix is ​​constructed based on the adjacency relationship and distance. First, a threshold d is determined, and other monitoring points whose distance to each monitoring point does not exceed the threshold d are searched. These two points are considered to be adjacent. W i,j =1; otherwise, they are not adjacent, W i,j =0, and construct the spatial weight matrix W1 in this way.

[0115] In the geographically weighted regression modeling method for CO2 concentration spatial filtration value of the present invention, before the step of selecting feature vectors representing spatial local heterogeneity as local spatial factors from the global spatial feature factors, the method further includes:

[0116] Obtain carbon dioxide concentration data for a preset area;

[0117] The preset region is the research area where a CO2 concentration estimation model needs to be built.

[0118] Optionally, CO2 concentration data for the preset area during the study period can be obtained at least through CO2 concentration monitoring stations, passive satellites, active satellites, or a combination thereof.

[0119] In this embodiment, the CO2 concentration data is remote sensing data from the DQ-1 atmospheric environment monitoring satellite.

[0120] Surface factors, atmospheric factors, human activity factors, and radiation factors are used as influencing factors to obtain influencing factor data for a preset area;

[0121] Optionally, surface factors may include at least air temperature, normalized vegetation index, geothermal flux, and land cover.

[0122] Optionally, atmospheric factors may include at least air pressure and wind speed.

[0123] Optionally, human activity factors may include at least population density and CO2 emissions from fossil fuel combustion.

[0124] Optionally, the radiation factor may include at least longwave radiation, shortwave radiation, and solar surface radiation.

[0125] The above factors all reflect the spatial distribution of ground CO2 concentration to some extent. The specific factor data can be selected as the influencing factors based on the climate characteristics and data quality of the study area.

[0126] Data on the selected impact factors were collected within the study area and during the study period to obtain impact factor data.

[0127] The carbon dioxide concentration data and the influencing factor data are preprocessed.

[0128] Preprocessing of CO2 concentration data and influencing factor data includes quality checks on active satellite scan data and removal of obvious outliers.

[0129] Preprocessing includes removing null and outlier values, projection transformation, and unification of temporal and spatial resolution.

[0130] Optionally, the raw CO2 data is a concentration with a time field, from which daily, monthly, quarterly, and annual average concentration values ​​can be calculated sequentially. The time resolution of the other influencing factors is also unified to the same time resolution as the CO2 data.

[0131] Optionally, regarding spatial resolution, the remaining influencing factors need to be converted to a consistent spatial resolution, which can be achieved by methods such as resampling or spatial interpolation.

[0132] Both fossil fuel combustion CO2 emissions and DQ-1 satellite CO2 data are point data, which can be converted into raster data covering the entire study area using a fishing net statistical method.

[0133] Data from different spatial reference frames need to be projected and transformed to unify the spatial reference frames.

[0134] After preprocessing, the data of each factor at the CO2 satellite scan points are extracted to screen for local spatial factors and to construct a CO2 concentration estimation model.

[0135] In the geographically weighted regression modeling method for CO2 concentration spatial filtration value of the present invention, before the step of selecting feature vectors representing spatial local heterogeneity as local spatial factors from the global spatial feature factors, the method further includes:

[0136] Calculate the variance inflation factor between each influencing factor data and the carbon dioxide concentration data. If the calculated variance inflation factor is greater than a preset threshold, remove the influencing factor corresponding to the current largest variance inflation factor.

[0137] Repeat the above steps until the variance inflation factor corresponding to each influencing factor after screening is less than the preset threshold.

[0138] Furthermore, in order to simplify the process of screening local spatial factors and the construction of the CO2 concentration estimation model, the extracted influencing factors are first subjected to a collinearity test to screen out some influencing factors.

[0139] Specifically, such as Figure 3 As shown, firstly, a candidate set of impact factors is constructed from the data of each impact factor. Then, data from this set are selected sequentially and collinearity tests are performed with CO2 concentration data. The variance inflation factor (VIF) corresponding to each impact factor is calculated. If there is an impact factor with a VIF greater than a preset threshold, the impact factor corresponding to the largest VIF needs to be removed. The above steps are repeated until the VIF corresponding to each impact factor is less than the preset threshold.

[0140] In this embodiment, the preset threshold is determined to be 10 based on experience.

[0141] Based on this, the construction process of a complete carbon dioxide concentration estimation model, ESF-SWR, is as follows: Figure 4 As shown.

[0142] Furthermore, for the completed CO2 concentration estimation model, the R-squared value of the model can also be calculated. 2 Adjusted R 2 The root mean square error (RMSE) and the Akaike information criterion (AIC) were used as evaluation indicators to verify the accuracy of the proposed CO2 concentration estimation model that takes into account both global spatial autocorrelation and local spatial heterogeneity.

[0143]

[0144] In the formula, y iy is the observed CO2 concentration at monitoring point i, and y is the average value of the observed data. is the CO2 concentration of i predicted by the model, and n is the number of monitoring points.

[0145]

[0146] In the formula, p is the number of independent variables, and R0 is the number of independent variables. 2 and Adj.R 2 The value range is [0,1], and the larger the value, the higher the accuracy of the model.

[0147]

[0148] In the formula, the smaller the RMSE, the higher the model accuracy.

[0149]

[0150] In the formula, the smaller the AIC, the higher the model accuracy.

[0151] In addition, the model can be cross-validated. Optionally, a 10-fold cross-validation method can be used to evaluate the model's prediction accuracy for CO2 concentration in non-scanned areas.

[0152] Specifically, the acquired CO2 concentration data and influencing factor data are randomly divided into 10 parts. Nine parts are selected as the training set for each iteration, and the model is built using the method described above. The remaining part is used as the test set. The CO2 concentration data and influencing factor data from the test set are substituted into the model obtained from the training set as independent variables, and the root mean square error (RMSE) of the test set is calculated. After each set of data has been tested once, the mean of the 10 RMSEs is calculated, which is the result of cross-validation. The smaller the RMSE, the higher the prediction accuracy of the model, the stronger its robustness, and the wider its applicability.

[0153] The following describes the CO2 concentration spatial filter value geographic weighted regression modeling device provided by the present invention. The CO2 concentration spatial filter value geographic weighted regression modeling device described below can be referred to in correspondence with the CO2 concentration spatial filter value geographic weighted regression modeling method described above.

[0154] like Figure 5 As shown, the spatial filter value geographic weighted regression modeling device for CO2 concentration includes a global module 501, a local module 502, and a construction module 503:

[0155] The global module 501 is used to construct a spatial weight matrix based on the geographical location of carbon dioxide concentration data monitoring points, and extract feature vectors representing global positional relationships from the spatial weight matrix as global spatial feature factors.

[0156] Traditional CO2 concentration regression models typically consider only one aspect: global spatial autocorrelation or local spatial heterogeneity, without taking both into account simultaneously. In other words, when processing spatial data, these models either consider the global correlation between data points or the specific heterogeneity of local regions, but do not combine the two, thus failing to fully capture the complex spatial distribution characteristics of CO2.

[0157] To balance the impact of global spatial autocorrelation and local spatial heterogeneity on CO2 concentration regression models, this invention proposes a modeling method based on eigenvector spatial filtering-based geographically weighted regression model (ESF-GWR). It posits that the spatial influence on CO2 concentration can be divided into two parts: one is the spatial influence caused by local environmental conditions themselves, and the other is the global spatial influence and its local manifestation.

[0158] Therefore, it is first necessary to extract global spatial feature factors that characterize the global positional relationship of CO2.

[0159] Specifically, a spatial weight matrix is ​​constructed based on the geographical location of CO2 concentration monitoring points, and global spatial feature factors are then selected from the eigenvectors of the spatial weight matrix.

[0160] The geographical location of the CO2 concentration data monitoring point is determined based on the method of obtaining the CO2 concentration data.

[0161] Optionally, when the CO2 concentration data comes from CO2 concentration monitoring stations established on the ground, the coordinates of each station in the study area are used as the geographical locations of the CO2 concentration data monitoring points.

[0162] Furthermore, due to the limited number of ground-based CO2 concentration monitoring stations and their uneven geographical distribution, the development of remote sensing technology has enabled the use of data collected by carbon monitoring satellites and other remote sensing methods to estimate ground-based CO2 concentrations. Carbon monitoring satellites are categorized into two types: active satellites and passive satellites.

[0163] Optionally, when the CO2 concentration data comes from a passive satellite, the geographical location corresponding to the data received by the passive satellite in the study area is used as the geographical location of the CO2 concentration data monitoring point.

[0164] Optionally, when the CO2 concentration data comes from active satellites, the coordinates of the active satellite scan points are used as the geographical locations of the CO2 concentration data monitoring points.

[0165] Compared to passive satellites, active satellites have significant advantages in data accuracy and spatial coverage methods. Furthermore, active satellite data consists of scanned data with clear spatiotemporal coordinates. Therefore, CO2 concentration data is preferably acquired by active satellites. In this embodiment, CO2 concentration data acquired by the active satellite DQ-1 will be used as an example for explanation.

[0166] It is understood that, in other feasible implementations, the above methods can also be combined to obtain CO2 concentration data. For example, CO2 concentration data can be obtained by using both active satellites and ground-based CO2 concentration monitoring stations. In this case, the scanning location of the active satellites and the coordinates of the CO2 concentration monitoring stations within the study area are used together as the geographical location of the CO2 concentration data monitoring points.

[0167] Based on this, in this embodiment, a spatial weight matrix W1 is constructed according to the projected coordinates of the active satellite scanning data, and then centered to form a symmetric matrix W:

[0168]

[0169] In the formula, I represents an n-dimensional identity matrix, 11 T This represents an n×n matrix where all elements are equal to 1, and n represents the number of CO2 concentration monitoring points within the study area, which in this embodiment is the number of active satellite scanning points within the study area.

[0170] Furthermore, based on the eigenvalues ​​and eigenvectors of the symmetric matrix W, several global spatial feature factors that best fit the OLS (Ordinary Least Squares) model are selected from the eigenvectors of the symmetric matrix W.

[0171] Local module 502 is used to select feature vectors representing local spatial heterogeneity from the global spatial feature factors as local spatial factors based on a preset model that reflects local heterogeneity.

[0172] Furthermore, in order to comprehensively consider the impact of local spatial heterogeneity on CO2 concentration, feature vectors characterizing local spatial heterogeneity are selected from global spatial feature factors as local spatial factors, which serve as the basis for the final construction of CO2 concentration.

[0173] Based on this, in order to find suitable local spatial factors among global spatial feature factors, global spatial feature factors are screened based on a preset model.

[0174] Optionally, the preset model can be a geographically weighted regression model, a spatial error model, a spatial lag model, or other models that can reflect local heterogeneity.

[0175] In this embodiment, a geographically weighted regression model is used as an example to illustrate the spatial heterogeneity by using spatially variable coefficients, and to further screen suitable local spatial factors.

[0176] Module 503 is used to construct a carbon dioxide concentration model based on the local spatial factor.

[0177] After selecting the local spatial factors, a CO2 concentration estimation model (ESF-GWR) that takes into account both global spatial autocorrelation and local heterogeneity can be constructed, as follows:

[0178]

[0179] In the formula, y i This represents the CO2 concentration corresponding to the i-th sampling point, (u i ,v i Let be the coordinates of the i-th sampling point, and β0 be a constant. k (u i ,v i Let x be the k-th regression parameter at the i-th sampling point. ik Let Eα be the value of the independent variable at the i-th sampling point. ik ε is the weight value of the k-th spatial feature at the i-th sampling point, which is calculated based on the local spatial factor. i Let be the random error of the i-th sampling point, satisfying the basic assumptions of zero mean, homoscedasticity, and mutual independence. p represents the number of regression parameters, and q represents the sum of the number of regression parameters and the number of local spatial factors.

[0180] This invention extracts global spatial feature factors by constructing a global spatial matrix, and then extracts local spatial factors that take into account both global and local features from the global spatial feature factors. These local spatial factors are then combined with the factor data that affect CO2 concentration to perform geographically weighted regression modeling, resulting in a CO2 concentration estimation model. This model can take into account both global spatial features and local spatial heterogeneity.

[0181] Figure 6 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 6As shown, the electronic device may include a processor 610, a communication interface 620, a memory 630, and a communication bus 640. The processor 610, communication interface 620, and memory 630 communicate with each other via the communication bus 640. The processor 610 can call logical instructions in the memory 630 to execute a geographic weighted regression modeling method for CO2 concentration spatial filtering. This method includes: constructing a spatial weight matrix based on the geographic locations of carbon dioxide concentration monitoring points, and extracting feature vectors representing global positional relationships from the spatial weight matrix as global spatial feature factors; based on a preset model reflecting local heterogeneity, selecting feature vectors representing local spatial heterogeneity from the global spatial feature factors as local spatial factors; and constructing a carbon dioxide concentration model based on the local spatial factors.

[0182] Furthermore, the logical instructions in the aforementioned memory 630 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0183] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the geographic weighted regression modeling method for spatial filter values ​​of CO2 concentration provided by the above methods. The method includes: constructing a spatial weight matrix based on the geographic locations of carbon dioxide concentration data monitoring points, and extracting feature vectors representing global location relationships from the spatial weight matrix as global spatial feature factors; selecting feature vectors representing local spatial heterogeneity from the global spatial feature factors based on a preset model reflecting local heterogeneity as local spatial factors; and constructing a carbon dioxide concentration model based on the local spatial factors.

[0184] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements a geographic weighted regression modeling method for spatial filter values ​​of CO2 concentration provided by the methods described above. The method includes: constructing a spatial weight matrix based on the geographic locations of carbon dioxide concentration data monitoring points, and extracting feature vectors representing global positional relationships from the spatial weight matrix as global spatial feature factors; selecting feature vectors representing local spatial heterogeneity from the global spatial feature factors based on a preset model reflecting local heterogeneity as local spatial factors; and constructing a carbon dioxide concentration model based on the local spatial factors.

[0185] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0186] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0187] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A CO2 concentration spatial filter value geographically weighted regression modeling method, characterized in that, The application relates to a method for constructing a carbon dioxide concentration model. The method comprises the following steps: a spatial weight matrix is constructed based on the geographical positions of carbon dioxide concentration data monitoring points, and a feature vector representing global position relationship is extracted from the spatial weight matrix as a global spatial feature factor, wherein the geographical positions of the carbon dioxide concentration data monitoring points are determined based on scanning point coordinates of active satellites; a feature vector representing spatial local heterogeneity is screened from the global spatial feature factor as a local spatial factor based on a preset model of reaction local heterogeneity; a carbon dioxide concentration model is constructed based on the local spatial factor; the step of screening the feature vector representing spatial local heterogeneity from the global spatial feature factor as the local spatial factor based on the preset model of reaction local heterogeneity specifically comprises the following steps: the global spatial feature factor is taken as an independent variable of the preset model, and the preset model is calculated by combining pre-acquired influence factor data; a correction determination coefficient corresponding to each global spatial feature factor is calculated; in the case that the correction determination coefficient obtained at each time is greater than the current maximum correction determination coefficient, the global spatial feature factor corresponding to the correction determination coefficient at the time is retained, and the correction determination coefficient at the time is updated as the current maximum correction determination coefficient; otherwise, the global spatial feature factor corresponding to the correction determination coefficient at the time is removed, wherein the initial value of the current maximum correction determination coefficient is set as zero; the global spatial feature factor retained after the loop calculation is taken as the local spatial factor; the step of taking the global spatial feature factor as the independent variable of the preset model and calculating the preset model by combining the pre-acquired influence factor data specifically comprises the following steps: the global spatial feature factors are randomly sorted and added to a spatial feature vector set; 2. The CO2 concentration spatial filter geographically weighted regression modeling method according to claim 1, characterized in that, the global spatial feature factors in the spatial feature vector set are sequentially added to an independent variable set of the preset model as the independent variables of the preset model to calculate the preset model by combining the pre-acquired influence factor data.

3. The CO2 concentration spatial filter geographically weighted regression modeling method according to claim 1, wherein, The preset model is a geographic weighted regression model, a spatial error model or a spatial lag model. After the step of constructing the carbon dioxide concentration model based on the local spatial factor, the method further comprises the following steps: a grid calculation is performed on the carbon dioxide concentration model to obtain a carbon dioxide concentration distribution map; 4. The CO2 concentration spatial filter geographically weighted regression modeling method according to claim 1, wherein, abnormal values in the carbon dioxide concentration distribution map are processed. The step of constructing the spatial weight matrix based on the geographical positions of the carbon dioxide concentration data monitoring points comprises the following steps:

5. The CO2 concentration spatial filter geographically weighted regression modeling method according to claim 1, wherein, the spatial weight matrix is constructed based on an adjacency relationship, an inverse distance weighting or a composite type of the adjacency relationship and the inverse distance weighting in combination with the geographical positions of the carbon dioxide concentration data monitoring points. Before the step of screening the feature vector representing spatial local heterogeneity from the global spatial feature factor as the local spatial factor based on the preset model of reaction local heterogeneity, the method further comprises the following steps: carbon dioxide concentration data of a preset area are acquired; terrestrial factors, atmospheric factors, human activity factors and radiation factors are taken as influence factors to acquire influence factor data of the preset area; the carbon dioxide concentration data and the influence factor data are preprocessed.

6. The CO2 concentration spatial filter geographically weighted regression modeling method according to claim 5, characterized in that, The preset model based on reaction local heterogeneity further comprises, before the step of screening a feature vector representing spatial local heterogeneity from the global spatial feature factors as a local spatial factor: calculating a variance inflation factor between each impact factor data and the carbon dioxide concentration data, and removing the impact factor corresponding to the current maximum variance inflation factor in the case that the currently calculated variance inflation factor is greater than a preset threshold value; repeating the above steps until the variance inflation factors corresponding to each impact factor after screening are all less than the preset threshold value.

7. A CO2 concentration spatial filter geographically weighted regression modeling device, characterized by, Comprise: a global module configured to construct a spatial weight matrix based on the geographical positions of carbon dioxide concentration data monitoring points, and extract a feature vector representing global position relationship from the spatial weight matrix as a global spatial feature factor, wherein the geographical positions of the carbon dioxide concentration data monitoring points are determined based on the scanning coordinates of active satellites; a local module configured to screen a feature vector representing spatial local heterogeneity from the global spatial feature factors as a local spatial factor based on a preset model based on reaction local heterogeneity; a construction module configured to construct a carbon dioxide concentration model based on the local spatial factor; The local module is specifically configured to use the global spatial feature factors as independent variables of the preset model, and combine the pre-acquired impact factor data to calculate the preset model; calculate the correction determination coefficient corresponding to each global spatial feature factor; in the case that the correction determination coefficient calculated each time is greater than the current maximum correction determination coefficient, retain the global spatial feature factor corresponding to the correction determination coefficient at this time, and update the correction determination coefficient at this time to the current maximum correction determination coefficient, otherwise, remove the global spatial feature factor corresponding to the correction determination coefficient at this time, wherein the initial value of the current maximum correction determination coefficient is set to zero; retain the global spatial feature factors after the loop calculation as the local spatial factors; The local module is specifically configured to add the spatial feature vector set after randomly sorting all the global spatial feature factors; add the global spatial feature factors in the spatial feature vector set to the independent variable set of the preset model in turn as the independent variables of the preset model, combine the pre-acquired impact factor data to calculate the preset model.

8. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor implements the CO2 concentration spatial filter value geographical weighted regression modeling method according to any one of claims 1 to 6 when executing the program.

Images