Multi-modal model-based industry pollution emission contribution pre-judgment method and device

Abstract

The invention relates to the technical field related to pollution, in particular to an industry pollution emission contribution pre-judgment method and device based on a multi-modal model. The method comprises the following steps: collecting original data; wherein the original data comprises related data of PM2.5 (Particulate Matter 2.5); the SwinLSTM model is used, the Swin Transform multi-scale feature extraction capability and the LSTM time sequence processing capability are fused, and the original data are preprocessed to obtain training data; constructing a multi-modal optimization deep learning model composed of a convolutional neural network and a graph attention network; training and optimizing the multi-modal optimization deep learning model based on the training data; obtaining to-be-predicted target industry information; and performing prediction based on the target industry information by using the trained multi-modal optimization deep learning model to obtain pollution emission pre-judgment after the target industry is located in a certain section.

Description

Technical Field

[0001] The present application relates to the field of pollution-related technologies, and specifically to a method and device for predicting industry pollution emission contributions based on a multimodal model. Background Art

[0002] In the field of air pollution control, accurately determining the industry sources of each component of PM2.5 and quantifying their contributions are crucial for effectively controlling air pollution and formulating scientific environmental policies. In previous research and practice, a variety of traditional methods have been used for PM2.5 source analysis, such as the chemical mass balance (CMB) model and positive definite matrix factorization (PMF). These methods can identify the main pollution source categories to a certain extent and estimate their approximate contributions. However, traditional methods have significant limitations. On the one hand, traditional models are often analyzed based on data of a single scale. Whether it is large-scale regional meteorological data or small-scale urban local emission data, it is difficult to fully reflect the complex atmospheric pollution process. For example, at a large scale, it is impossible to accurately capture the differences in pollutant diffusion within the city due to local terrain, building layout and other factors; at a small scale, it is impossible to consider the long-distance transmission of pollutants between regions. On the other hand, traditional methods have limited processing capabilities for massive, complex and dynamically changing data. With the increase in the number of environmental monitoring stations and the increase in monitoring frequency, the amount of data generated is growing exponentially. It is difficult for traditional models to quickly and accurately mine the complex relationship between various industries and PM2.5 components behind the data. Summary of the invention

[0003] In view of this, the embodiments of the present application are directed to providing a method and device for predicting industry pollution emission contributions based on a multimodal model.

[0004] This application provides a method for predicting industry pollution emission contribution based on a multimodal model, including:

[0005] Collecting raw data; wherein the raw data includes data related to PM2.5;

[0006] Using the SwinLSTM model, the multi-scale feature extraction capability of Swin Transformer and the time series processing capability of LSTM are integrated to deeply mine various types of data at different scales, extract key spatial features, and capture the evolution of data over time and its impact on PM2.5 concentration. The original data is preprocessed to obtain training data.

[0007] Construct a multimodal optimization deep learning model consisting of a convolutional neural network and a graph attention network. The convolutional neural network efficiently extracts data spatial features through a combination of convolutional layers and pooling layers, accurately captures the relationship between emission sources of various industries and PM2.5 concentration distribution, and identifies the impact pattern of industry emissions on PM2.5 pollution in different regions. The graph attention network uses the attention mechanism to weight the features of emission source nodes of various industries, deeply analyzes the relationship between emission sources of various industries, and their impact weights on various PM2.5 pollution components. The multimodal optimization deep learning model is used to predict the contribution ratio of different industries to various PM2.5 pollution components in different regions, different time periods, and different meteorological conditions.

[0008] Training and tuning the multimodal optimization deep learning model based on the training data;

[0009] Obtain information on the target industry to be forecasted;

[0010] Using the trained multi-modal optimization deep learning model, predictions are made based on the target industry information to obtain a prediction of pollution emissions when the target industry is located in a certain area.

[0011] In some embodiments, the raw data includes meteorological data, geographic information data, industrial emission data, traffic emission data, energy emission data, and PM2.5 component concentration monitoring data.

[0012] In some embodiments, tuning the multimodal optimization deep learning model based on the training data includes:

[0013] Inputting part of the training data into the multimodal optimization deep learning model to perform prediction to obtain a preliminary prediction result;

[0014] The preliminary prediction results are compared with the true values ​​to obtain the prediction error. The prediction error is fed back to each level of the model layer by layer using the back propagation algorithm. The model weights and parameters are adjusted dynamically. After calculating the loss function, the model parameters are continuously adjusted in the direction of minimizing the loss function according to the gradient descent method, so as to gradually reduce the difference between the prediction results and the true values.

[0015] In some embodiments, the target industry information includes: the industry category, the industry's emission data, the geographic information of the location, and meteorological conditions.

[0016] The present application provides a device for predicting industry pollution emission contribution based on a multimodal model, comprising:

[0017] A collection module, used to collect raw data; wherein the raw data includes PM2.5 related data;

[0018] The preprocessing module is used to use the SwinLSTM model, integrate the Swin Transformer multi-scale feature extraction capability and the LSTM time series processing capability, deeply mine various types of data from different scales, extract key spatial features, and capture the evolution of data over time and its impact on PM2.5 concentration. It preprocesses the original data to obtain training data;

[0019] A construction module is used to construct a multimodal optimized deep learning model composed of a convolutional neural network and a graph attention network. The convolutional neural network efficiently extracts data spatial features through a combination of convolutional layers and pooling layers, accurately captures the relationship between emission sources of various industries and PM2.5 concentration distribution, and identifies the impact pattern of industry emissions on PM2.5 pollution in different regions. The graph attention network uses the attention mechanism to weight the features of emission source nodes of various industries, deeply analyzes the relationship between emission sources of various industries, and their impact weights on various PM2.5 pollution components. The multimodal optimized deep learning model is used to predict the contribution ratio of different industries to various PM2.5 pollution components in different regions, different time periods, and different meteorological conditions.

[0020] A tuning module, used for training and tuning the multimodal optimization deep learning model based on the training data;

[0021] An acquisition module is used to obtain target industry information to be predicted;

[0022] The prediction module is used to use the trained multi-modal optimization deep learning model to make predictions based on the target industry information to obtain a prediction of pollution emissions after the target industry is located in a certain area.

[0023] In some embodiments, the raw data includes meteorological data, geographic information data, industrial emission data, traffic emission data, energy emission data, and PM2.5 component concentration monitoring data.

[0024] In some embodiments, tuning the multimodal optimization deep learning model based on the training data includes:

[0025] Inputting part of the training data into the multimodal optimization deep learning model to perform prediction to obtain a preliminary prediction result;

[0026] The preliminary prediction results are compared with the true values ​​to obtain the prediction error. The prediction error is fed back to each level of the model layer by layer using the back propagation algorithm. The model weights and parameters are adjusted dynamically. After calculating the loss function, the model parameters are continuously adjusted in the direction of minimizing the loss function according to the gradient descent method, so as to gradually reduce the difference between the prediction results and the true values.

[0027] In some embodiments, the target industry information includes: the industry category, the industry's emission data, the geographic information of the location, and meteorological conditions.

[0028] The present application provides an electronic device, including:

[0029] A processor, and a memory for storing a program executable by the processor;

[0030] The processor is used to implement the above-mentioned method for predicting industry pollution emission contribution based on the multimodal model by running the program in the memory.

[0031] The present application provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the processor executes the above-mentioned method for predicting industry pollution emission contribution based on a multimodal model.

[0032] The present application provides a method for predicting the contribution of industrial pollution emissions based on a multimodal model. First, the original data is collected; wherein the original data includes data related to PM2.5; the SwinLSTM model is used to integrate the multi-scale feature extraction capability of SwinTransformer and the time series processing capability of LSTM, and various types of data are deeply mined from different scales to extract key spatial features, while capturing the evolution of data over time and its impact on PM2.5 concentration, pre-processing the original data to obtain training data; a multimodal optimized deep learning model composed of a convolutional neural network and a graph attention network is constructed; wherein the convolutional neural network efficiently extracts data spatial features through a combination of convolutional layers and pooling layers, and accurately captures the relationship between emission sources of various industries and PM2.5 The correlation between concentration distributions is used to identify the impact pattern of industry emissions on PM2.5 pollution in different regions; the graph attention network uses the attention mechanism to weight the characteristics of the emission source nodes of each industry, deeply analyze the relationship between the emission sources of each industry, and their impact weights on each PM2.5 pollution component. The multimodal optimization deep learning model is used to predict the contribution ratio of different industries to each PM2.5 pollution component in different regions, different time periods, and different meteorological conditions; the multimodal optimization deep learning model is trained and tuned based on the training data; the target industry information to be predicted is obtained; the trained multimodal optimization deep learning model is used to predict based on the target industry information to obtain the pollution emission prediction after the target industry is located in a certain area. In this way, through multi-scale model coupling and AI technology, accurate prediction of each PM2.5 component is achieved, the contribution of each industry to PM2.5 is accurately traced, and its spatial distribution law is deeply analyzed. With the help of the SwinLSTM model, the massive and complex data such as meteorology, geography, and emissions of various industries are deeply mined to accurately extract key spatial features, capture the law of time evolution, and greatly improve the efficiency and quality of data processing. The innovative multi-modal optimization deep learning model (CNN+GAT) can not only accurately present the changing trend of fine particulate matter composition and concentration at a single site, but also build an accurate regional concentration model. On this basis, the contribution ratio of each industry to each component of PM2.5 in different regions, time periods and meteorological conditions is accurately calculated, and the responsibility of each industry in air pollution is clearly defined, providing a clear direction for the formulation of targeted governance strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0034] Figure 1 It is a flow chart of a method for predicting industry pollution emission contribution based on a multimodal model provided in one embodiment of the present application.

[0035] Figure 2 It is a partial flow chart of a method provided by an embodiment of the present application.

[0036] Figure 3 It is a structural schematic diagram of an industry pollution emission contribution prediction device based on a multimodal model provided in an embodiment of the present application.

[0037] Figure 4 It is a schematic diagram of the structure of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0038] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0039] Application Overview

[0040] In the field of air pollution control, accurately determining the industry sources of each component of PM2.5 and quantifying their contributions are crucial for effectively controlling air pollution and formulating scientific environmental policies. In previous research and practice, a variety of traditional methods have been used for PM2.5 source analysis, such as the chemical mass balance (CMB) model and positive definite matrix factorization (PMF). These methods can identify the main pollution source categories to a certain extent and estimate their approximate contributions. However, traditional methods have significant limitations. On the one hand, traditional models are often analyzed based on data of a single scale. Whether it is large-scale regional meteorological data or small-scale urban local emission data, it is difficult to fully reflect the complex atmospheric pollution process. For example, at a large scale, it is impossible to accurately capture the differences in pollutant diffusion within the city due to local terrain, building layout and other factors; at a small scale, it is impossible to consider the long-distance transmission of pollutants between regions. On the other hand, traditional methods have limited processing capabilities for massive, complex and dynamically changing data. With the increase in the number of environmental monitoring stations and the increase in monitoring frequency, the amount of data generated is growing exponentially. It is difficult for traditional models to quickly and accurately mine the complex relationship between various industries and PM2.5 components behind the data. With the rapid development of big data and artificial intelligence technologies, as well as the increasing requirements for refined air pollution control, there is an urgent need for an innovative method that can integrate multi-scale data and, with the help of artificial intelligence's powerful data processing and analysis capabilities, accurately trace the industry contribution of each component of PM2.5, thereby providing solid technical support for more efficient and targeted air pollution control strategies.

[0041] In the survey before the study was carried out, it was found that some researchers used a single-scale atmospheric diffusion model, such as the Gaussian plume model, to try to analyze the industry contribution of each component of PM2.5. When using this model, it is first necessary to collect the pollutant emission inventory of each industry in a specific area, and clarify the location, emission amount, and emission height of the emission sources of different industries. At the same time, obtain the meteorological data of the area, such as average wind speed, dominant wind direction, etc. Based on these data, the formula of the Gaussian plume model is used to calculate the contribution of pollutants emitted by various industries to the concentration of each component of PM2.5 at the monitoring point according to the relative position of the emission source and the monitoring point, and the influence of meteorological conditions on the diffusion of pollutants. For example, for a chemical industry emission source, the model calculates the proportion of sulfate formed by its sulfur dioxide emissions at the surrounding monitoring points in PM2.5. However, this single-scale model has many limitations. Only local small-scale meteorological conditions and emission source information are considered, and it is impossible to reflect the impact of large-scale regional meteorological changes on pollutant transmission, such as long-distance transport of pollutants caused by cross-regional atmospheric circulation. The model has limited processing capabilities for complex terrain and urban underlying surfaces, and cannot accurately simulate the diffusion path and concentration distribution of pollutants in mountainous areas and urban areas with tall buildings. Due to the relatively simple assumptions of the model, it is difficult to accurately estimate the complex chemical reactions between pollutants emitted by various industries and the secondary PM2.5 components. This results in the analysis results of the industry contribution of each component of PM2.5 being incomplete and inaccurate, and unable to meet the current demand for refined governance of air pollution.

[0042] Some researchers also use traditional factor analysis methods to explore the sources and industry contributions of various components of PM2.5. They collect long-term series of PM2.5 chemical component data from multiple monitoring sites, as well as relevant potential pollution source information. After standardizing the data, the factor analysis algorithm is used to reduce many related variables to a few unrelated comprehensive factors. During the analysis process, the degree of correlation between each factor and each variable is determined based on the factor loading matrix, and then the main pollution source categories are identified, such as industrial sources, traffic sources, dust sources, etc. Then, based on the scores of each factor, the relative contribution of different pollution source categories to each component of PM2.5 is estimated. For example, if a factor is highly correlated with characteristic pollutants emitted by industry and the factor score is high, it is inferred that industrial sources contribute more to certain components in PM2.5.

[0043] However, this traditional factor analysis method has obvious shortcomings. It mainly relies on the statistical characteristics of the data and has extremely high requirements for data quality. If there is noise or outliers in the data, it will seriously affect the accuracy of the analysis results. Moreover, this method cannot dynamically update and adapt to the ever-changing pollution source situation. It is difficult to timely and accurately incorporate new industries or emission sources into the analysis system. In addition, it can only give the approximate contribution of each pollution source category to the PM2.5 component, which is difficult to be accurate to specific industries and cannot meet the needs of accurate tracing and formulating targeted control measures.

[0044] This invention aims to develop a precise method for tracing the contribution of each industry to PM2.5 components by using multi-scale model coupling and AI technology. At present, air pollution control faces severe challenges, and there are many technical problems that need to be overcome. On the one hand, the formation and distribution of PM2.5 are affected by multiple factors such as meteorology, geography, and emissions from various industries, making it difficult to accurately determine the specific contribution ratio of each industry to each PM2.5 pollution component in different regions, time periods, and meteorological conditions.

[0045] Traditional analytical methods are unable to cope with such complex interactions. On the other hand, faced with massive, multi-source and heterogeneous data, existing technologies lack the ability to efficiently integrate and deeply mine, making it difficult to fully extract key information, which seriously affects the accuracy of tracing the sources of PM2.5 components. Furthermore, even if some tracing information is obtained, there is no systematic solution for how to effectively transform it into a scientific and effective pollution prevention and control strategy. Based on this,

[0046] The present invention constructs a multimodal model system integrating the SwinLSTM model, convolutional neural network (CNN) and graph attention network (GAT), which accurately analyzes the contribution ratio of various industries to each PM2.5 pollution component. At the same time, based on the precise calculation results, the emission data of a certain industry, the geographical information of the location, the meteorological conditions, etc. are input into the model to simulate its impact on the PM2.5 components in the surrounding environment, and realize the prediction of pollution emissions after the industry is located in a certain area, providing core data support for air pollution control, and promoting air pollution control to move towards refinement and efficiency.

[0047] After introducing the basic principles of the present application, various non-limiting embodiments of the present application will be described in detail with reference to the accompanying drawings.

[0048] Exemplary Methods

[0049] like Figure 1 , Figure 2 As shown, the method for predicting industry pollution emission contribution based on a multimodal model provided in this application includes the following contents.

[0050] Step S110, collecting raw data; wherein the raw data includes data related to PM2.5;

[0051] Specifically, the original data includes meteorological data, geographic information data, industrial emission data, traffic emission data, energy emission data and PM2.5 component concentration monitoring data.

[0052] Collecting raw data is the basis of the entire invention, and these data will be used for subsequent model training and analysis. The collected raw data include:

[0053] Meteorological data: such as temperature, humidity, wind speed, wind direction, etc. These data can help understand the movement and diffusion characteristics of the atmosphere and have an important impact on the transmission and diffusion of pollutants.

[0054] Geographic information data: including information such as topography, landforms, and land use types. These data help analyze the impact of the geographical environment on the diffusion of pollutants, such as the blocking and guiding effects of mountains, rivers, and urban building layout on airflow.

[0055] Industrial emission data: covers information such as pollutant emissions, emission source locations, and emission heights of various industrial enterprises. Since industrial emissions are one of the important sources of PM2.5, the emission characteristics and impact ranges of different industries vary.

[0056] Traffic emission data: such as motor vehicle exhaust emissions, traffic flow, road distribution, etc. Traffic emissions have a significant impact on urban air quality, especially in traffic congested areas.

[0057] Energy emission data: involves pollutant emissions during energy production and use, such as exhaust emissions from thermal power plants. Energy-related emissions have an important impact on regional air quality.

[0058] Monitoring data on the concentration of PM2.5 components: The concentration data of different chemical components in PM2.5 (such as sulfate, nitrate, organic carbon, etc.) obtained through environmental monitoring stations. This is the basic data for evaluating the contribution of various industries and directly reflects the current status of PM2.5 pollution in the atmosphere.

[0059] Step S120, using the SwinLSTM model, integrating the Swin Transformer multi-scale feature extraction capability and the LSTM time series processing capability, deeply mining various types of data from different scales, extracting key spatial features, and capturing the evolution of data over time and its impact on PM2.5 concentration, preprocessing the original data to obtain training data;

[0060] Specifically, after collecting the raw data, the SwinLSTM model needs to be used to preprocess the data to obtain training data suitable for model training. The specific process is as follows:

[0061] Combining Swin Transformer's multi-scale feature extraction capabilities with LSTM's time series processing capabilities: Swin Transformer can extract features from data at different scales and capture spatial features in the data, while LSTM is good at processing time series data and can capture the changing patterns of data over time. Combining the two can fully mine key information in various types of data, including features in spatial and temporal dimensions, and the impact of these features on PM2.5 concentrations.

[0062] Deeply mine various types of data and extract key spatial features: Through multi-scale analysis of the original data, key spatial features related to PM2.5 concentration are extracted, such as the distribution characteristics of emission sources in different regions, the impact of geographical obstacles on pollutant diffusion, etc. These features help to understand the spatial distribution of pollutants.

[0063] Capture the evolution of data over time and its impact on PM2.5 concentration: Analyze how changes in meteorological conditions, emission source activities, etc. over time affect PM2.5 concentration, such as the impact of seasonal changes, day-night changes, etc. on pollutant generation and diffusion, so as to grasp the temporal evolution trend of PM2.5 concentration.

[0064] Preprocess the raw data to obtain training data: After extracting key features and capturing evolution patterns, screen the data to remove invalid or abnormal data points, and then perform standardization and normalization to form an ordered data set as input data for subsequent model training to ensure that the model can learn effective patterns from high-quality data.

[0065] Step S130, constructing a multimodal optimized deep learning model composed of a convolutional neural network and a graph attention network; wherein the convolutional neural network efficiently extracts data spatial features through a combination of convolutional layers and pooling layers, accurately captures the correlation between emission sources of various industries and PM2.5 concentration distribution, and identifies the impact pattern of industry emissions on PM2.5 pollution in different regions; the graph attention network uses the attention mechanism to weight the features of emission source nodes of various industries, deeply analyzes the relationship between emission sources of various industries, and their impact weights on various PM2.5 pollution components, and the multimodal optimized deep learning model is used to predict the contribution ratio of different industries to various PM2.5 pollution components in different regions, different time periods, and different meteorological conditions;

[0066] Specifically, a multimodal optimization deep learning model consisting of a convolutional neural network (CNN) and a graph attention network (GAT) is constructed to predict the contribution ratio of each component of PM2.5 in different industries under different conditions. The specific construction process is as follows:

[0067] Convolutional Neural Network (CNN) section: Through the combination of convolutional layers and pooling layers, data spatial features are efficiently extracted: the convolutional layer can automatically learn local features in the data, and the pooling layer is used to reduce the data dimension and calculation amount while retaining important features. This combination can effectively extract spatial features related to PM2.5 concentration distribution from preprocessed data, such as the intensity and distribution pattern of emission sources in different regions. Accurately capture the association between emission sources of various industries and PM2.5 concentration distribution: By learning spatial features in the data, CNN can identify the correlation between the location, intensity and other characteristics of emission sources in different industries and the spatial distribution of PM2.5 concentration, such as the corresponding relationship between certain high-emission industry areas and high-concentration PM2.5 areas. Identify the impact pattern of industry emissions in different regions on PM2.5 pollution: Further analyze the comprehensive impact of emissions from various industries in different regions, and identify the spatial distribution pattern of industry emission combinations in different regions on PM2.5 pollution, such as the pollution characteristics of emissions from different functional areas such as urban industrial areas and traffic-intensive areas on the surrounding environment.

[0068] Graph Attention Network (GAT) section: Use the attention mechanism to weight the characteristics of the emission source nodes of each industry: Treat the emission sources of each industry as nodes in the graph, and use the attention mechanism to weight the characteristics of emission sources of different industries according to the relationship and importance between the nodes, so that the model can pay more attention to the emission sources of industries that contribute more to PM2.5. In-depth analysis of the relationship between emission sources of various industries: Consider the possible synergy or antagonism between emission sources of different industries. For example, the precursors emitted by some industries react chemically in the atmosphere to generate secondary pollutants, affecting the pollution contribution of other industries. By analyzing these relationships, the comprehensive impact of various industries can be more accurately evaluated. Analyze the impact weight of emission sources of various industries on each pollution component of PM2.5: Determine the contribution weight of emission sources of different industries to different chemical components in PM2.5. For example, some industries may mainly contribute sulfate components, while others have a greater impact on organic carbon components. This is crucial for accurately quantifying the pollution contribution of various industries.

[0069] The overall function of the multimodal optimization deep learning model: The spatial features extracted by CNN and the industry emission source relationships and weight information analyzed by GAT are combined to form a multimodal model architecture that can comprehensively consider the characteristics of both spatial and industry relationships, thereby more accurately predicting the contribution ratio of different industries to each component of PM2.5 in different regions, different time periods, and different meteorological conditions.

[0070] Step S140, training and tuning the multimodal optimization deep learning model based on the training data;

[0071] Step S150, obtaining target industry information to be predicted;

[0072] Step S160, using the trained multi-modal optimization deep learning model, based on the target industry information, makes a prediction to obtain a prediction of pollution emissions when the target industry is located in a certain area.

[0073] With this setting, through multi-scale model coupling and AI technology, accurate prediction of each component of PM2.5 can be achieved, the contribution of each industry to PM2.5 can be accurately traced, and its spatial distribution pattern can be deeply analyzed. With the help of the SwinLSTM model, the massive and complex data such as meteorology, geography, and emissions from various industries are deeply mined to accurately extract key spatial features, capture the laws of temporal evolution, and greatly improve the efficiency and quality of data processing. The innovative multimodal optimization deep learning model (CNN+GAT), its construction and subsequent optimization can not only accurately present the changing trends of the fine particulate matter composition and concentration at a single site, but also build an accurate regional concentration model. On this basis, the contribution ratio of each industry to each component of PM2.5 under different regions, time periods, and meteorological conditions is accurately calculated, and the responsibility of each industry in air pollution is clearly defined, providing a clear direction for the formulation of targeted governance strategies.

[0074] Specifically, tuning the multimodal optimization deep learning model based on the training data includes: inputting part of the training data into the multimodal optimization deep learning model to perform prediction to obtain preliminary prediction results; comparing the preliminary prediction results with the true values ​​to obtain prediction errors, using a back propagation algorithm to feed back the prediction errors to each level of the model layer by layer, dynamically adjusting the model weights and parameters, and after calculating the loss function, continuously adjusting the model parameters in the direction of minimizing the loss function according to the gradient descent method to gradually reduce the difference between the prediction results and the true values.

[0075] The following are the specific steps to tune the multimodal optimization deep learning model based on training data:

[0076] Model training and preliminary prediction:

[0077] Data preparation and division: The training data obtained in step S120 is divided into two parts, one part is used as the actual training set for the initial training of the model, and the other part is used as the validation set for subsequent model evaluation and tuning. The division of the training set and the validation set usually adopts a certain ratio, such as the common 70% of the data as the training set and 30% as the validation set, or adopts methods such as cross-validation to ensure the stability and generalization ability of the model on different data subsets.

[0078] Model input and preliminary prediction: The training set data is input into the multimodal optimization deep learning model. The model calculates and analyzes the input data based on the previously constructed architecture and initialized parameters, and outputs preliminary prediction results. These preliminary prediction results include the contribution ratio of each component of PM2.5 in different industries in different regions, time periods and meteorological conditions corresponding to the training set.

[0079] Prediction error calculation and feedback:

[0080] Comparison and error calculation: Compare the initial prediction results output by the model with the corresponding true values ​​in the training set (actual observations or known data on the contribution ratio of each component of PM2.5), and calculate the difference between the two, i.e. the prediction error. Commonly used error calculation methods include mean square error (MSE), mean absolute error (MAE), etc. These error indicators can quantify the accuracy of model predictions and provide direction for subsequent model tuning.

[0081] Back propagation and error feedback: Using the back propagation algorithm, the calculated prediction error is fed back to each layer of the model layer by layer. During the feedback process, the weights and parameters of the model are dynamically adjusted according to the error. Specifically, starting from the output layer, the gradient of each layer is calculated forward in sequence according to the error gradient until the input layer. This process enables the model to understand the degree of influence of each parameter on the final prediction result, thereby providing a basis for the optimization and adjustment of the parameters.

[0082] Model parameter adjustment and optimization:

[0083] Loss function calculation and gradient descent: While feeding back the error, the loss function of the model is calculated. The loss function is usually defined based on the prediction error, such as the mean square error loss function. Then, according to the gradient descent method, the model parameters are continuously adjusted in the direction of minimizing the loss function according to the gradient direction of the loss function to the model parameters. The gradient descent method iteratively updates the parameters and gradually reduces the value of the loss function, thereby improving the prediction performance of the model.

[0084] Learning rate and adjustment step size: In the process of parameter adjustment, the learning rate is an important hyperparameter, which determines the step size of each parameter update. A learning rate that is too large may cause the model to oscillate or diverge during the optimization process, while a learning rate that is too small will slow down the training process. Therefore, it is necessary to reasonably set the learning rate according to the training situation of the model, or adopt an adaptive learning rate adjustment strategy to ensure that the model can quickly and stably converge to the optimal or near-optimal parameter state.

[0085] Model evaluation and validation:

[0086] Validation set evaluation: After completing a round of parameter adjustment, use the validation set to evaluate the model, calculate the prediction error and loss function value of the model on the validation set, and verify the generalization ability and performance of the model on data that has not participated in training. If the model performs poorly on the validation set, such as large errors or overfitting (that is, it performs well on the training set but poorly on the validation set), it is necessary to further adjust the structure or parameters of the model, such as adding regularization terms, adjusting the number of network layers or the number of neurons, etc., to improve the generalization performance of the model.

[0087] Model optimization iteration: The above process (model training, error calculation, parameter adjustment, and validation set evaluation) is iterated multiple times, and each iteration is retrained and evaluated based on the new parameter state until the model's performance on the validation set reaches a satisfactory level, or the loss function value tends to be stable and no longer decreases significantly. At this time, the model is considered to have been fully tuned and the training process can be stopped. The purpose of the entire tuning process is to enable the model to learn effective patterns and rules from the training data, so that when faced with new and unseen data, it can accurately predict the contribution ratio of different industries to each component of PM2.5, and provide reliable decision-making support for air pollution control.

[0088] Specifically, the target industry information to be predicted is obtained; the target industry information includes: the industry category, the industry emission data, the geographical information of the location, and the meteorological conditions. The trained multi-modal optimization deep learning model is used to make predictions based on the target industry information to obtain the pollution emission prediction of the target industry after it is located in a certain location.

[0089] The following are the specific steps to obtain target industry information and use the trained multi-modal optimization deep learning model to make predictions to obtain the target industry pollution emission predictions:

[0090] Get target industry information:

[0091] Determine the industry category: clarify the category of the target industry to be predicted, such as chemical, steel, cement, transportation and other different industry types. Because different industries have significant differences in production processes, raw material usage and pollutant emission characteristics, their contributions to the various components of PM2.5 are also different.

[0092] Collect industry emission data: Obtain detailed emission data of the target industry, including the emission of major pollutants (such as sulfur dioxide, nitrogen oxides, particulate matter, etc.), the location of the emission source (latitude and longitude coordinates, etc.), the emission height (such as chimney height), and the emission time pattern (continuous emission or intermittent emission, etc.). These data can be obtained from the company's environmental impact assessment report, environmental monitoring data, government environmental statistics departments, or through field monitoring.

[0093] Obtain geographic information of the location: Collect geographic information of the location of the target industry, such as topography (such as whether it is located in a valley, plain, coastal area, etc.), surrounding land use type (such as whether it is close to residential areas, commercial areas, other industrial areas, etc.), distance from major transportation arteries and water systems and relative position relationship, etc. This geographic information can be obtained through geographic information system (GIS) data, satellite remote sensing images or field surveys.

[0094] Collect meteorological data: Organize meteorological data for the target industry, including the prevailing wind direction, wind speed, temperature change, humidity change, precipitation distribution, etc. Meteorological data can come from local meteorological observation stations, climate data released by meteorological departments, or professional meteorological data service platforms. These data are helpful for analyzing the transmission and diffusion of pollutants in the atmosphere.

[0095] Prediction process:

[0096] Data preprocessing and format conversion: The collected target industry information is preprocessed and format converted according to the model input requirements. For example, the emission data is standardized or normalized to make it consistent with the data range and format during model training; the geographic information and meteorological condition data are converted into numerical forms or feature vectors that the model can recognize and process, such as encoding the geographic location into longitude and latitude coordinates, and converting the terrain type into the corresponding numerical code.

[0097] Model input and prediction calculation: The pre-processed target industry information is input into the trained multi-modal optimization deep learning model. The model will predict the pollution emissions of the target industry after it is located in the area based on the patterns and rules learned from previous training, taking into account the emission characteristics, geographical location, and meteorological conditions of the target industry.

[0098] Output prediction results and interpretation of predictions: The model outputs the prediction results of the pollution emissions of the target industry on each component of PM2.5 in the surrounding environment, including detailed information such as the concentration contribution value and impact range of different components. Based on these prediction results, the potential impact of the pollution emissions of the target industry on the local air quality can be further analyzed to determine whether corresponding pollution prevention and control measures need to be taken, or to provide a scientific basis for the environmental management department to formulate environmental access policies and regulatory measures for the industry.

[0099] Exemplary Devices

[0100] The device embodiments of the present application can be used to execute the method embodiments of the present application. For details not disclosed in the device embodiments of the present application, please refer to the method embodiments of the present application.

[0101] Figure 3FIG. 1 is a block diagram of an industry pollution emission contribution prediction device based on a multimodal model provided by an embodiment of the present application. Figure 3 As shown, the device comprises:

[0102] A collection module 31 is used to collect raw data; wherein the raw data includes PM2.5 related data;

[0103] The preprocessing module 32 is used to use the SwinLSTM model, integrate the Swin Transformer multi-scale feature extraction capability and the LSTM time series processing capability, deeply mine various types of data from different scales, extract key spatial features, and capture the evolution of data over time and its impact on PM2.5 concentration, preprocess the original data, and obtain training data;

[0104] Building module 33, for building a multimodal optimization deep learning model composed of a convolutional neural network and a graph attention network; wherein the convolutional neural network efficiently extracts data spatial features through a combination of convolutional layers and pooling layers, accurately captures the correlation between emission sources of various industries and PM2.5 concentration distribution, and identifies the impact pattern of industry emissions on PM2.5 pollution in different regions; the graph attention network uses the attention mechanism to weight the features of emission source nodes of various industries, deeply analyzes the relationship between emission sources of various industries, and their impact weights on various PM2.5 pollution components, and the multimodal optimization deep learning model is used to predict the contribution ratio of different industries to various PM2.5 pollution components in different regions, different time periods, and different meteorological conditions;

[0105] A tuning module 34, configured to train and tune the multimodal optimization deep learning model based on the training data;

[0106] An acquisition module 35 is used to acquire target industry information to be predicted;

[0107] The prediction module 36 is used to use the trained multi-modal optimization deep learning model to make predictions based on the target industry information to obtain a prediction of pollution emissions when the target industry is located in a certain area.

[0108] In some embodiments, the raw data includes meteorological data, geographic information data, industrial emission data, traffic emission data, energy emission data, and PM2.5 component concentration monitoring data.

[0109] In some embodiments, tuning the multimodal optimization deep learning model based on the training data includes:

[0110] Inputting part of the training data into the multimodal optimization deep learning model to perform prediction to obtain a preliminary prediction result;

[0111] The preliminary prediction results are compared with the true values ​​to obtain the prediction error. The prediction error is fed back to each level of the model layer by layer using the back propagation algorithm. The model weights and parameters are adjusted dynamically. After calculating the loss function, the model parameters are continuously adjusted in the direction of minimizing the loss function according to the gradient descent method, so as to gradually reduce the difference between the prediction results and the true values.

[0112] In some embodiments, the target industry information includes: the industry category, the industry's emission data, the geographic information of the location, and meteorological conditions.

[0113] Exemplary Electronic Devices

[0114] Below, reference Figure 4 To describe an electronic device according to an embodiment of the present application. Figure 4 A block diagram of an electronic device according to an embodiment of the present application is illustrated.

[0115] like Figure 4 As shown, electronic device 400 includes one or more processors 410 and memory 420 .

[0116] The processor 410 may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device 400 to perform desired functions.

[0117] The memory 420 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory (cache), etc. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 410 may run the program instructions to implement the industry pollution emission contribution prediction method based on a multimodal model and / or other desired functions of the various embodiments of the present application described above. Various contents such as category correspondences may also be stored in the computer-readable storage medium.

[0118] In one example, the electronic device 400 may further include: an input device 430 and an output device 440 , and these components are interconnected via a bus system and / or other forms of connection mechanisms (not shown).

[0119] In addition, the input device 430 may also include, for example, a keyboard, a mouse, an interface, etc. The output device 440 may output various information to the outside, including analysis results, etc. The output device 440 may include, for example, a display, a speaker, a printer, a communication network and a remote output device connected thereto, etc.

[0120] Of course, to simplify, Figure 4 Only some of the components in the electronic device related to the present application are shown, and components such as a bus, an input / output interface, etc. are omitted. In addition, the electronic device may further include any other appropriate components according to specific application conditions.

[0121] Exemplary computer program products and computer-readable storage media

[0122] In addition to the above-mentioned methods and devices, an embodiment of the present application may also be a computer program product, which includes computer program instructions, which, when executed by a processor, enable the processor to execute the steps of the method for predicting industry pollution emission contributions based on a multimodal model according to various embodiments of the present application described in the above "Exemplary Method" section of this specification.

[0123] The computer program product may be written in any combination of one or more programming languages ​​to write program codes for performing the operations of the embodiments of the present application, including object-oriented programming languages, such as Java, C++, etc., and conventional procedural programming languages, such as "C" language or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as an independent software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0124] In addition, an embodiment of the present application may also be a computer-readable storage medium having computer program instructions stored thereon, which, when executed by a processor, causes the processor to execute the steps of the method for predicting industry pollution emission contributions based on a multimodal model according to various embodiments of the present application described in the above “Exemplary Method” section of this specification.

[0125] The computer readable storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can include, for example, but is not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0126] The above description has been given for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present application to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.

Claims

1. A method for predicting industry pollution emission contribution based on a multimodal model, characterized in that: include: Collecting raw data; wherein the raw data includes data related to PM2.5; Using the SwinLSTM model, the multi-scale feature extraction capability of Swin Transformer and the time series processing capability of LSTM are integrated to deeply mine various types of data at different scales, extract key spatial features, and capture the evolution of data over time and its impact on PM2.5 concentration. The original data is preprocessed to obtain training data. Construct a multimodal optimization deep learning model consisting of a convolutional neural network and a graph attention network. The convolutional neural network efficiently extracts data spatial features through a combination of convolutional layers and pooling layers, accurately captures the relationship between emission sources of various industries and PM2.5 concentration distribution, and identifies the impact pattern of industry emissions on PM2.5 pollution in different regions. The graph attention network uses the attention mechanism to weight the features of emission source nodes of various industries, deeply analyzes the relationship between emission sources of various industries, and their impact weights on various PM2.5 pollution components. The multimodal optimization deep learning model is used to predict the contribution ratio of different industries to various PM2.5 pollution components in different regions, different time periods, and different meteorological conditions. Training and tuning the multimodal optimization deep learning model based on the training data; Obtain information on the target industry to be forecasted; Using the trained multi-modal optimization deep learning model, predictions are made based on the target industry information to obtain a prediction of pollution emissions when the target industry is located in a certain area.

2. The method for predicting industry pollution emission contribution based on a multimodal model according to claim 1 is characterized in that: The original data includes meteorological data, geographic information data, industrial emission data, traffic emission data, energy emission data and PM2.5 component concentration monitoring data.

3. The method for predicting industry pollution emission contribution based on a multimodal model according to claim 2 is characterized in that: Tuning the multimodal optimization deep learning model based on the training data includes: Inputting part of the training data into the multimodal optimization deep learning model to perform prediction to obtain a preliminary prediction result; The preliminary prediction results are compared with the true values ​​to obtain the prediction error. The prediction error is fed back to each level of the model layer by layer using the back propagation algorithm. The model weights and parameters are adjusted dynamically. After calculating the loss function, the model parameters are continuously adjusted in the direction of minimizing the loss function according to the gradient descent method, so as to gradually reduce the difference between the prediction results and the true values.

4. The method for predicting industry pollution emission contribution based on a multimodal model according to claim 2 is characterized in that: The target industry information includes: the industry category, the industry's emission data, the location's geographical information, and meteorological conditions.

5. A device for predicting industry pollution emission contribution based on a multimodal model, characterized in that: include: A collection module, used to collect raw data; wherein the raw data includes PM2.5 related data; The preprocessing module is used to use the SwinLSTM model, integrate the Swin Transformer multi-scale feature extraction capability and the LSTM time series processing capability, deeply mine various types of data from different scales, extract key spatial features, and capture the evolution of data over time and its impact on PM2.5 concentration. It preprocesses the original data to obtain training data; A construction module is used to construct a multimodal optimized deep learning model composed of a convolutional neural network and a graph attention network. The convolutional neural network efficiently extracts data spatial features through a combination of convolutional layers and pooling layers, accurately captures the relationship between emission sources of various industries and PM2.5 concentration distribution, and identifies the impact pattern of industry emissions on PM2.5 pollution in different regions. The graph attention network uses the attention mechanism to weight the features of emission source nodes of various industries, deeply analyzes the relationship between emission sources of various industries, and their impact weights on various PM2.5 pollution components. The multimodal optimized deep learning model is used to predict the contribution ratio of different industries to various PM2.5 pollution components in different regions, different time periods, and different meteorological conditions. A tuning module, used for training and tuning the multimodal optimization deep learning model based on the training data; An acquisition module is used to obtain target industry information to be predicted; The prediction module is used to use the trained multi-modal optimization deep learning model to make predictions based on the target industry information to obtain a prediction of pollution emissions after the target industry is located in a certain area.

6. The device for predicting industry pollution emission contribution based on a multimodal model according to claim 5 is characterized in that: The original data includes meteorological data, geographic information data, industrial emission data, traffic emission data, energy emission data and PM2.5 component concentration monitoring data.

7. The device for predicting industry pollution emission contribution based on a multimodal model according to claim 6 is characterized in that: Tuning the multimodal optimization deep learning model based on the training data includes: Inputting part of the training data into the multimodal optimization deep learning model to perform prediction to obtain a preliminary prediction result; The preliminary prediction results are compared with the true values ​​to obtain the prediction error. The prediction error is fed back to each level of the model layer by layer using the back propagation algorithm. The model weights and parameters are adjusted dynamically. After calculating the loss function, the model parameters are continuously adjusted in the direction of minimizing the loss function according to the gradient descent method, so as to gradually reduce the difference between the prediction results and the true values.

8. The device for predicting industry pollution emission contribution based on a multimodal model according to claim 7 is characterized in that: The target industry information includes: the industry category, the industry's emission data, the location's geographical information, and meteorological conditions.

9. An electronic device, characterized in that: include: A processor, and a memory for storing a program executable by the processor; The processor is used to implement the industry pollution emission contribution prediction method based on a multimodal model as described in any one of claims 1 to 7 by running the program in the memory.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which, when executed by a processor, enables the processor to execute the method for predicting industry pollution emission contribution based on a multimodal model as described in any one of claims 1 to 7.

Images