Hectometer ozone concentration data estimation method based on LG-Transform model

Abstract

The invention discloses a hectometer ozone concentration data estimation method based on an LG-Transform model. The method comprises the following steps: S1, collecting road network data, ERA5 data and remote sensing data, and correspondingly resampling the data into corresponding grids of a grid map; s2, an LG-Transform model containing a Tannform model and a Light GBM connection layer is constructed, and model training is carried out by using the sample data set; s3, the LightGBM connection layer calculates feature importance of the sequence feature matrix and selects features with high importance as splitting points of the decision trees, the LightGBM connection layer performs weighted summation on prediction results of all the decision trees and generates an output value, and final regression of the output value is completed through a full connection layer; and S4, collecting data of a research area, inputting the data into the trained LG-Transformer model, and correspondingly outputting ozone estimated concentration data according to the grids. According to the method, the relation capture and the model training among the space-time information, the data characteristic information and the ozone concentration are realized, the ozone concentration estimation of the research area can be effectively carried out according to the grid, and the ozone concentration distribution of the whole research area is further obtained.

Description

Technical Field

[0001] The present invention relates to the field of ozone concentration model estimation, and in particular to a method for estimating hundred-meter-level ozone concentration data based on a LG-Transformer model. Background Art

[0002] In the prior art, in order to obtain ozone concentration data, it is necessary to set up an ozone concentration measuring station at a certain location, and use the ozone concentration measuring station to quickly and accurately measure the ozone concentration in the air. However, the ozone concentration measuring station is only suitable for ozone concentration monitoring at a specific location, and the ozone concentration data measured by the ozone concentration measuring station can only represent the ozone concentration at that location. For ozone concentration monitoring in a larger area or a larger area, the monitoring method of the ozone concentration measuring station is often not feasible. At present, the estimation of ozone concentration in a larger area is often carried out by spatial interpolation processing using the data from various ozone concentration measuring stations (the purpose is to convert discrete point data into a continuous data surface), such as the inverse distance weighted interpolation method (the inverse distance weighted interpolation method assumes that points with closer distances have a greater impact on the predicted value than points with farther distances). The prediction estimate is made by assigning weights to known points and calculating the weighted average. The weight is an inverse function of the distance between the known point and the target point. The smaller the distance, the greater the weight. This method is simple and intuitive, with fast calculation speed, and is suitable for situations where data is evenly distributed. However, the inverse distance weighted interpolation method is based on the assumption that spatial changes in all directions are consistent. In practical applications, it performs poorly for situations where data distribution is sparse or complex, and has low estimation accuracy. It is not suitable for ozone concentration estimation over a large area. For example, spatial interpolation processing also includes the Kriging interpolation method (a spatial interpolation method based on geostatistics, assuming that geographic variables have spatial autocorrelation). Kriging interpolation not only considers the values ​​of known points, but also spatial variability (described by the semivariance function); when calculating, it is necessary to first calculate the semivariance function to estimate spatial correlation, and then select a suitable model (such as a spherical model, exponential model, Gaussian model, etc.) for fitting based on the calculated semivariance value, and calculate the weights through the semivariance model with the goal of minimizing the prediction error. Relatively speaking, the Kriging interpolation method can more accurately describe spatial variability, but the Kriging interpolation method is very sensitive to the choice of the semivariance model, and the estimation accuracy is low. It is also not suitable for estimating ozone concentrations over a large area.

[0003] Atmospheric chemistry models can mainly simulate and study the physical transmission and chemical reaction processes of the atmosphere. By comparing and verifying with ground and remote sensing observations, they can simulate the photochemical reaction process of ozone (O3) and its precursors, analyze the sources of pollutants, and estimate the distribution of pollutant concentrations. The core of conventional atmospheric models mainly includes three parts: physical transmission, pollutant emissions and diffusion, and chemical transformation. At the beginning of the model, pollutant emissions are usually input in the form of a list, and then the real-time concentration of atmospheric pollutants is simulated under the drive of the meteorological field and the calculation of the chemical reaction mechanism. According to different input parameters, the concentration of pollutants in a specific time period can be integrated to obtain the average concentration of pollutants in that time period. Machine learning models have developed rapidly in recent years. How to combine ground station monitoring data, multi-source geographic element data, and satellite remote sensing data for deep data fusion learning, iterative optimization, and further use it for near-ground ozone concentration estimation is a technical difficulty that needs to be studied and solved in response to large-scale regional ozone concentration estimation. Summary of the invention

[0004] The purpose of the present invention is to provide a method for estimating ozone concentration data at the hundred-meter level based on the LG-Transformer model. The LG-Transformer model is constructed by deeply fusing the Tansformer model and the LightGBM connection layer. The data is resampled according to the grid of the grid map, and the spatial information, time information, and feature information of the data in the grid are fused and encoded. The relationship between spatiotemporal information, data feature information and ozone concentration is captured and model training is achieved, and the ozone concentration in the study area can be effectively estimated according to the grid, thereby obtaining the ozone concentration distribution in the entire study area.

[0005] The purpose of the present invention is achieved through the following technical solutions:

[0006] A method for estimating hundred-meter-level ozone concentration data based on the LG-Transformer model, the method comprising:

[0007] S1, construct a grid map, collect road network data, ERA5 data and remote sensing data and resample them to the corresponding grids of the grid map;

[0008] S2. Construct an LG-Transformer model including the Tansformer model and the LightGBM connection layer and use the sample data set to train the model. The sample data set is based on the sample data and ozone concentration labels corresponding to the grid association. The sample data includes road network data, ERA5 data and remote sensing data. The Tansformer model of the LG-Transformer model integrates and encodes the spatial information, temporal information and feature information of the data in the grid. The Tansformer model sets the time step and uses a multi-head attention mechanism to encode the correlation between each time step, followed by residual connection and layer normalization, and then uses the ReLu function for nonlinear transformation. The Tansformer model uses an interactive attention mechanism for decoding and outputs a sequence feature matrix.

[0009] S3, LightGBM connection layer calculates the feature importance of the sequence feature matrix and selects the features with high importance as the splitting points of the decision tree, and calculates the residual between the model prediction value and the true value as the model constraint; LightGBM connection layer weights the prediction results of all decision trees and generates an output value, and completes the final regression through the fully connected layer;

[0010] S4. Collect the road network data, ERA5 data and remote sensing data of the study area and resample them to the corresponding grids of the grid map. Then input the trained LG-Transformer model and output the ozone estimation concentration data according to the grid correspondence.

[0011] In order to better implement the present invention, in step S1, the grid in the grid map is 100m*100m pixel size, the ERA5 data comes from two data sets, Era5-land and Era5 single level, the ERA5 data includes 2m dew point temperature, 2m temperature, surface latent heat flux, surface solar radiation, surface thermal radiation, boundary layer height, 100m wind speed, and the ERA5 data is resampled to the grid of the grid map using bilinear interpolation; the remote sensing data includes normalized vegetation index data, aerosol optical thickness data, nighttime radiation data, tropospheric formaldehyde data, tropospheric nitrogen oxides data and atmospheric ozone column concentration data, and the remote sensing data is resampled to the grid of the grid map using bilinear interpolation.

[0012] Preferably, in step S2, the Tansformer model uses a spherical embedding layer to extract spatial information from the data in the grid, as follows:

[0013] The spatial information of the data in the grid is converted into two-dimensional longitude and latitude information, and the two-dimensional longitude and latitude information is mapped to a three-dimensional sphere. The expression of the spherical coordinate vector (x, y, z) is as follows:

[0014] x=cos(φ)×cos(φ), y=cos(φ)×sin(λ), z=sin(φ), where φ and λ represent the latitude and longitude in the two-dimensional longitude and latitude information respectively;

[0015] The feature fusion of spatial correlation is extracted through neural network, and the fusion expression is as follows:

[0016] fusion(x, y, z) = ReLU((x, y, z)×W1+b1)×W2+b2, fusion(x, y, z) represents the result of feature fusion, W1 and W2 represent weight matrices, b1 and b2 are bias terms.

[0017] Preferably, in step S2, the Tansformer model extracts the time information of the data in the grid and performs forward and backward position encoding in sequence according to the set time step. The position encoding expression is as follows:

[0018]

[0019] Where P t，2i Indicates the coded data of even number of bits, P t，2i+1 represents the odd-numbered coded data, t represents the time step index number, i represents the data item in the grid data, d model Represents the feature dimension.

[0020] Preferably, the multi-head attention mechanism of the Tansformer model is divided into query Q, key K and value V operations when performing correlation operations between each time step, and the expression is as follows:

[0021] Where Q = X input ·W Q , K = X input ·W K , V = X input ·W v , W Q , W k They represent the projection matrix, K T represents the transpose of key K, d k represents the dimension of key K, X input Represents the input of the correlation operation; the multi-head attention mechanism performs the correlation operation independently on each head and iterates the correlation operation h times.

[0022] Preferably, the residual connection and layer normalization processing method of the Tansformer model is as follows: Input the correlation operation into X input The output of the iterative correlation operation h times of the multi-head attention mechanism is residually connected, and then the layer is normalized; the nonlinear transformation expression of the ReLu function is as follows:

[0023] X ffn =ReLu(X attn ·W1+b1)·W2+b2, where X attn Represents the output of the normalized layer, X ffn Represents the output after nonlinear transformation.

[0024] Preferably, in step S3, the final regression expression is as follows:

[0025] y final =F final (x)·W out +b out , where y final represents the estimated ozone concentration, F final (x) indicates that the LightGBM connection layer generates an output value by weighted summing up the prediction results of all decision trees, W out represents the weight layer, b out Represents the bias layer.

[0026] Preferably, the sequence feature matrix is ​​flattened according to all time step information before entering the LightGBM connection layer; the LightGBM connection layer divides the continuous features into discrete buckets through an optimized histogram algorithm.

[0027] Preferably, in step S1, the road network data is road network vector data, and the road network data is subjected to kernel density analysis to output road network density raster data with a pixel size of 100m*100m.

[0028] Preferably, in step S2, the ozone concentration label in the sample data set is derived from the real ozone concentration actually monitored by the measuring station.

[0029] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0030] (1) The present invention obtains the LG-Transformer model by deeply fusing the Tansformer model and the LightGBM connection layer. The data is resampled according to the grids of the grid map, and the spatial information, time information, and feature information of the data in the grid are fused and encoded, realizing the capture and model training of the relationship between spatiotemporal information, data feature information and ozone concentration, and being able to estimate the ozone concentration at a resolution of 100 meters. The final estimation result reaches a spatial resolution of 100m×100m, which significantly improves the precision and practical application value of the estimation result. The present invention can effectively estimate the ozone concentration in the study area according to the grid, and then obtain the ozone concentration distribution in the entire study area.

[0031] (2) The present invention uses the Transformer model to obtain the time information of the data in the grid and encodes the position in sequence, which can effectively process the correlation between the previous and next sequence data and enhance the time correlation in the ozone concentration estimation; the LightGBM connection layer of the present invention improves the processing capabilities of high-dimensional discrete and continuous features and improves the estimation accuracy; the present invention can capture slight changes in ozone concentration, making the positioning of pollution sources more accurate and the spatial distribution of ozone concentration more refined; it significantly improves the spatial precision of the ozone concentration distribution results, and can provide more targeted support for urban micro-scale atmospheric environmental governance, health risk assessment and policy decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a flowchart of the method of the present invention;

[0033] Figure 2 This is a fitting relationship diagram of the real value and the predicted value of a city selected as the study area in the embodiment;

[0034] Figure 3 This is a schematic diagram of the ozone concentration distribution at the hundred-meter level in a certain city selected as the study area in the embodiment;

[0035] Figure 4 This is a graph showing the ozone concentration results for a certain area of ​​the study area in the example. DETAILED DESCRIPTION

[0036] The present invention is further described in detail below in conjunction with embodiments:

[0037] Example

[0038] like Figure 1 As shown, a method for estimating ozone concentration data at the hundred-meter level based on the LG-Transformer model includes:

[0039] S1. Build a grid map, collect road network data, ERA5 data and remote sensing data and resample them to the corresponding grids of the grid map.

[0040] In some embodiments, the grid in the grid map is 100m*100m pixel size, ERA5 data comes from two data sets, Era5-land and Era5 single level, ERA5 data includes 2m dew point temperature (abbreviated as d2m), 2m temperature (abbreviated as t2m), surface latent heat flux (abbreviated as slhf), surface solar radiation (abbreviated as ssr), surface thermal radiation (abbreviated as str), boundary layer height (abbreviated as b1h), 100m wind speed, in this embodiment, 100m wind speed includes u100 and v100, u100 refers to the component of wind speed at 100 meters in the east-west direction (i.e., longitude), and v100 refers to the component of wind speed at 100 meters in the north-south direction (i.e., latitudinal). ERA5 data are resampled to the grid of the grid map using bilinear interpolation, and bilinear interpolation is used to resample the ERA5 data from the initial resolution to the grid of the grid map with a resolution of 100m*100m. Remote sensing data include normalized vegetation index data, aerosol optical thickness data, nighttime radiation data, tropospheric formaldehyde data, tropospheric nitrogen oxide data and atmospheric ozone column concentration data. Remote sensing data can be derived from multi-source data, such as normalized vegetation index data (abbreviated as NDVI) and aerosol optical thickness data (abbreviated as AOD) derived from MODIS (English full name Moderate Resolution Imaging Spectroradiometer, i.e. remote sensing data in medium resolution imaging spectrometer) products, nighttime radiation data (abbreviated as NR) derived from VIIRS (English full name Visible Infrared Imaging Radiometer Suite), tropospheric formaldehyde data, tropospheric nitrogen oxide data and atmospheric ozone column concentration data derived from TropOMI (English full name Tropospheric Monitoring Instrument, i.e. tropospheric monitoring instrument). Remote sensing data are resampled to the grid of the grid map by bilinear interpolation (resampled to the grid of 100m*100m resolution of the grid map by initial resolution interpolation, and the remote sensing data of this embodiment are collected according to the monthly scale). The road network data is road network vector data (in this embodiment, it can be derived from 0pen Street Map, including all road network vector data including expressways, trunk roads, main roads and secondary roads). The road network data is subjected to kernel density analysis to output road network density raster data with a pixel size of 100m*100m.

[0041] In this embodiment, a certain city is selected as the study area as an example, and the study area is divided into 100m*100m grids on the grid map. The road network data, ERA5 data and remote sensing data are collected and matched to the corresponding grids of the grid map according to the correspondence in time and space (the grid contains the road network data, ERA5 data and remote sensing data, and also contains time information and space information, and the space information is the longitude and latitude information).

[0042] S2. Construct an LG-Transformer model including a Tansformer model and a LightGBM connection layer and use a sample data set for model training. The LG-Transformer model of the present invention is formed by deep fusion of the Tansformer model and the LightGBM connection layer. The sample data set is associated with the corresponding sample data and ozone concentration labels according to the grid (the ozone concentration labels in the sample data set are derived from the actual ozone concentration actually monitored by the measuring station), and the sample data includes road network data, ERA5 data, and remote sensing data; the Tansformer model of the LG-Transformer model fuses and encodes the spatial information, temporal information, and feature information of the data in the grid. The Tansformer model sets the time step and uses a multi-head attention mechanism to encode the correlation between each time step, followed by residual connection and layer normalization, and then uses the ReLu function for nonlinear transformation.

[0043] In some embodiments, the Tansformer model uses a spherical embedding layer to extract spatial information from data in a grid as follows:

[0044] The spatial information of the data in the grid is converted into two-dimensional longitude and latitude information, and the two-dimensional longitude and latitude information is mapped to a three-dimensional sphere (the two-dimensional longitude and latitude information is mapped to a three-dimensional sphere and nonlinearly fused to learn complex spatial relationships). The expression of the spherical coordinate vector (x, y, z) is as follows:

[0045] x=cos(φ)×cos(φ), y=cos(φ)×sin(λ), z=sin(φ), where φ and λ represent the latitude and longitude in the two-dimensional longitude and latitude information respectively;

[0046] The feature fusion of spatial correlation is extracted through neural network, and the fusion expression is as follows:

[0047] fusion(x, y, z) = ReLU((x, y, z)×W1+b1)×W2+b2, fusion(x, y, z) represents the result of feature fusion (the output geographic embedding vector), W1 and W2 represent weight matrices, and b1 and b2 are bias terms.

[0048] In some embodiments, the Tansformer model extracts the time information of the data in the grid and performs forward and backward position encoding in sequence according to the set time step (this embodiment selects the month, that is, one month as the time step). The position encoding expression is as follows:

[0049]

[0050] Where P t，2i Indicates the coded data of even number of bits, P t，2i+1 represents the odd-numbered coded data, t represents the time step index number, i represents the data item in the grid data, d model Represents the feature dimension; to ensure that the encodings of different positions are not too similar and provide good time step distinction, 10000 is selected as the base; position encoding introduces the uniqueness of position in high-dimensional space. The input feature d is transformed into input (The input feature is the feature dimension of each time step, and the feature dimension is the dimension of the data item in the grid data) mapped to the feature dimension d of the Tansformer model model .

[0051] In some embodiments, the multi-head attention mechanism of the Tansformer model is divided into query Q, key K and value V operations when performing correlation operations between each time step, and the expression is as follows:

[0052] Where Q = X input ·W Q , K = X input ·W K , V = X input ·W v , W Q , W k They represent the projection matrix, K T represents the transpose of key K, d k represents the dimension of key K, X input Represents the input of the correlation operation; the multi-head attention mechanism performs the correlation operation independently for each head and iterates it h times. For example, if there are k heads, the data output expression of the multi-head attention mechanism is as follows:

[0053] MultiHead(X)=Concat(head1,...,head k )·W0, head k represents the output of the k-th head correlation operation, W0 represents the output projection matrix, Concat represents the CONCAT function, and MultiHead(X) represents the output of the multi-head attention mechanism iterative correlation operation h times.

[0054] In some embodiments, the residual connection and layer normalization processing method of the Tansformer model is as follows: the correlation operation is input to X input The output MultiHHead(X) of the multi-head attention mechanism iterative correlation operation h times is residually connected, and then the layer is normalized; the nonlinear transformation expression of the ReLu function is as follows:

[0055] X ffn =ReLu(X attn ·W1+b1)·W2+b2, where X attn Represents the output of the normalized layer, X ffn Preferably, the output after the layer normalization process is connected to the output after the nonlinear transformation by normalization.

[0056] The Tansformer model uses an interactive attention mechanism to decode and output a sequence feature matrix. The interactive attention mechanism is used to capture the relationship between the input sequence and the target sequence. The expression is as follows:

[0057] Where Q u represents the input sequence, K u 、V u Represent the key and value respectively (usually the hidden state of the last layer in the encoding phase), Indicates key K u The dimension, K u T Indicates key K u The transpose of .

[0058] S3, LightGBM connection layer calculates the feature importance of the sequence feature matrix and selects the features with high importance as the splitting point of the decision tree, and calculates the residual between the model prediction value and the true value as the model constraint; LightGBM connection layer weights and sums the prediction results of all decision trees and generates an output value, and the output value is passed through the fully connected layer to complete the final regression. Preferably, the sequence feature matrix is ​​flattened according to all time step information before entering the LightGBM connection layer; LightGBM connection layer divides continuous features into discrete buckets (bins) through an optimized histogram algorithm, which can speed up calculations and reduce memory usage.

[0059] In some embodiments, the residual expression between the predicted value of the calculation model and the true value is as follows:

[0060] where y j represents the true value at the jth position, F t-1 (x j) represents the model’s predicted value for the jth sample after the t-1th iteration.

[0061] The loss function uses mean square error, which is expressed as follows: in Represents the sum of the differences between the true value and the predicted value, y j , Represent the true value and predicted value at the jth position, and n represents the total number of samples j. i Build a new decision tree; first, for each feature, try multiple split points, calculate the residual mean square error (MSE) of the left and right child nodes after the split, and select the split point that can maximize the error reduction; then start from the root node and gradually split the features until the maximum depth or number of leaf nodes is reached. Then update the model and add the output of the new tree to the current model:

[0062] F t (x) = F t-1 (x)+η·T(x), where η is the learning rate and T(x) is the prediction result of the new tree; in each round of training, the LightGBM connection layer calculates the residual (i.e., negative gradient) as the new target variable according to the loss function and fits it; finally, the prediction results of all decision trees are weighted and summed to generate the output value F final (x).

[0063] In some embodiments, the final regression expression is as follows:

[0064] y final =F final (x)·W out +b out , where y final represents the estimated ozone concentration, F final (x) indicates that the LightGBM connection layer generates an output value by weighted summing up the prediction results of all decision trees, W out represents the weight layer, b out Represents the bias layer.

[0065] S4. Collect the road network data, ERA5 data and remote sensing data of the study area and resample them to the corresponding grids of the grid map. Then input the trained LG-Transformer model and output the estimated ozone concentration data according to the grid. In this embodiment, a certain city is selected as the study area. The study area is divided into 100m*100m grids on the grid map. Finally, the road network data, ERA5 data and remote sensing data of the study area are collected and input into the trained LG-Transformer model to obtain the estimated ozone concentration of each grid in the study area in the grid map. Then, according to the following example, Figure 3As shown in FIG. 1 , a schematic diagram of the estimated ozone concentration distribution at the 100-meter level can be obtained. The model in this embodiment is verified using a five-fold cross-validation indicator, such as Figure 2 As shown in the figure, the LG-Transformer model of the present invention has good centralization of ozone concentration estimation, and the deviation from the true value is small. The estimation result has a good fitting effect with the true value, and the ozone concentration estimation result has high accuracy. Its determination coefficient R 2 The value reached 0.936 (close to 1, indicating that the estimation is relatively accurate), and the RMSE reached 7.461 μg / m 3 , MAE is 5.7000 μg / m 3 , the Pearson correlation coefficient r is 0.9759 (close to 1, indicating good correlation), and the MAPE reaches 6.53% (small deviation). The present invention greatly improves the accuracy of estimating near-ground ozone concentration data by integrating ozone-related multi-source feature data and innovating the model structure. This embodiment selects a certain city as the study area, divides the study area into 100m*100m grids on the grid map, and obtains Figure 3 The schematic diagram of the estimated ozone concentration distribution at the hundred-meter level is shown in FIG. Figure 4 As shown, the 100m*100m resolution near-ground ozone concentration full coverage data implemented in the present invention can provide more detailed ozone concentration information within the city, especially in small areas such as blocks and road intersections; this high-resolution monitoring can capture small changes in ozone concentration, making the location of pollution sources more accurate and the spatial distribution of ozone concentration more refined. The present invention significantly improves the spatial precision of ozone concentration distribution results, and can provide more targeted support for urban micro-scale atmospheric environmental governance, health risk assessment, and policy decision-making.

[0066] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for estimating ozone concentration data at the hundred-meter level based on the LG-Transformer model, characterized by: The methods include: S1, construct a grid map, collect road network data, ERA5 data and remote sensing data and resample them to the corresponding grids of the grid map; S2. Construct an LG-Transformer model including the Tansformer model and the LightGBM connection layer and use the sample data set to train the model. The sample data set is based on the sample data and ozone concentration labels corresponding to the grid association. The sample data includes road network data, ERA5 data and remote sensing data. The Tansformer model of the LG-Transformer model integrates and encodes the spatial information, temporal information and feature information of the data in the grid. The Tansformer model sets the time step and uses a multi-head attention mechanism to encode the correlation between each time step, followed by residual connection and layer normalization, and then uses the ReLu function for nonlinear transformation. The Tansformer model uses an interactive attention mechanism for decoding and outputs a sequence feature matrix. S3, LightGBM connection layer calculates the feature importance of the sequence feature matrix and selects the features with high importance as the splitting points of the decision tree, and calculates the residual between the model prediction value and the true value as the model constraint; LightGBM connection layer weights the prediction results of all decision trees and generates an output value, and completes the final regression through the fully connected layer; S4. Collect the road network data, ERA5 data and remote sensing data of the study area and resample them to the corresponding grids of the grid map. Then input the trained LG-Transformer model and output the ozone estimation concentration data according to the grid correspondence.

2. The method for estimating ozone concentration data at the hundred-meter level based on the LG-Transformer model according to claim 1, characterized in that: In step S1, the grid in the grid map has a pixel size of 100m*100m, the ERA5 data comes from two data sets, Era5-land and Era5 s ingle 1evel, the ERA5 data includes 2m dew point temperature, 2m temperature, surface latent heat flux, surface solar radiation, surface thermal radiation, boundary layer height, 100m wind speed, and the ERA5 data is resampled to the grid of the grid map using bilinear interpolation; the remote sensing data includes normalized vegetation index data, aerosol optical thickness data, nighttime radiation data, tropospheric formaldehyde data, tropospheric nitrogen oxides data and atmospheric ozone column concentration data, and the remote sensing data is resampled to the grid of the grid map using bilinear interpolation.

3. The method for estimating ozone concentration data at the hundred-meter level based on the LG-Transformer model according to claim 1, characterized in that: In step S2, the Tansformer model uses the spherical embedding layer to extract spatial information from the data in the grid as follows: The spatial information of the data in the grid is converted into two-dimensional longitude and latitude information, and the two-dimensional longitude and latitude information is mapped to a three-dimensional sphere. The expression of the spherical coordinate vector (x, y, z) is as follows: x=cos(φ)×cos(φ), y=cos(φ)×sin(λ), z=sin(φ), where φ, λ 分 Indicates the latitude and longitude in the two-dimensional longitude and latitude information; The feature fusion of spatial correlation is extracted through neural network, and the fusion expression is as follows: fusion(x, y, z) = ReLU((x, y, z)×W1+b1)×W2+b2, fusion(x, y, z) represents the result of feature fusion, W1 and W2 represent weight matrices, b1 and b2 are bias terms.

4. The method for estimating ozone concentration data at the hundred-meter level based on the LG-Transformer model according to claim 1, characterized in that: In step S2, the Tansformer model extracts the time information of the data in the grid and performs forward and backward position encoding in sequence according to the set time step. The position encoding expression is as follows: Among them, Pt t，2i Indicates the coded data of even number of bits, P t，2i+1 represents the odd-numbered coded data, t represents the time step index number, i represents the data item in the grid data, d model Represents the feature dimension.

5. The method for estimating ozone concentration data at the hundred-meter level based on the LG-Transformer model according to claim 1, characterized in that: The multi-head attention mechanism of the Tansformer model is divided into query Q, key K and value V operations for the correlation operations between each time step. The expression is as follows: Where Q = X input ·W Q , K = X input ·W K , V = X input ·w v , W Q , W k They represent the projection matrix, K T represents the transpose of key K, d k represents the dimension of key K, X input Represents the input of the correlation operation; the multi-head attention mechanism performs the correlation operation independently on each head and iterates the correlation operation h times.

6. The method for estimating ozone concentration data at the hundred-meter level based on the LG-Transformer model according to claim 5, characterized in that: The residual connection and layer normalization processing method of the Tansformer model is as follows: Input the correlation operation into X input The output of the iterative correlation operation h times of the multi-head attention mechanism is residually connected, and then the layer is normalized; the nonlinear transformation expression of the ReLu function is as follows: X ffn =ReLu(X attn ·W1+b1)·W2+b2, where X attn Represents the output of the normalized layer, X ffn Represents the output after nonlinear transformation.

7. The method for estimating ozone concentration data at the hundred-meter level based on the LG-Transformer model according to claim 1, characterized in that: In step S3, the final regression expression is as follows: y final =F final (x)·W out +b out , where y final represents the estimated ozone concentration, F final (x) indicates that the LightGBM connection layer generates an output value by weighted summing up the prediction results of all decision trees, W out represents the weight layer, b out Represents the bias layer.

8. The method for estimating ozone concentration data at the hundred-meter level based on the LG-Transformer model according to claim 1, characterized in that: The sequence feature matrix is ​​flattened according to all time step information before entering the LightGBM connection layer; the LightGBM connection layer divides the continuous features into discrete buckets through an optimized histogram algorithm.

9. The method for estimating ozone concentration data at the hundred-meter level based on the LG-Transformer model according to claim 1, characterized in that: In step S1, the road network data is road network vector data, and the road network data is subjected to kernel density analysis to output road network density raster data with a pixel size of 100m*100m.

10. The method for estimating ozone concentration data at the hundred-meter level based on the LG-Transformer model according to claim 1, characterized in that: In step S2, the ozone concentration label in the sample data set is derived from the real ozone concentration actually monitored by the measuring station.

Images