Rapid air quality prediction method and device for large-scale sparse sampling scene

Abstract

The invention provides an air quality rapid prediction method and device for a large-scale sparse sampling scene, and relates to the field of air quality prediction. The method comprises the steps that air quality data are obtained through an air quality sensor group, and the air quality sensor group comprises a plurality of air quality sensors; according to the position relation of the multiple air quality sensors in the research area, a graph structure corresponding to the air quality sensor group is obtained, and the graph structure comprises the graph node connection relation among the multiple air quality sensors; a data structure conversion assembly is constructed, and the data structure conversion assembly is used for converting the graph structure into a compact grid structure according to the graph node connection relation among the multiple air quality sensors and constructing the mapping relation between the graph structure and the compact grid structure; constructing a lightweight parameter sharing space-time expansion convolutional network component, wherein the lightweight parameter sharing space-time expansion convolutional network component is used for quickly capturing a space-time dependency relationship in the air quality data; according to the mapping relationship and the space-time dependency relationship, constructing an integrated predictor through a data structure conversion component and a lightweight parameter sharing space-time expansion convolutional network component; and according to the air quality data, an air quality prediction result is obtained through an integrated predictor. The problem that an existing model is difficult to balance large-scale coverage and real-time response is solved, and the problem that calculation resources and prediction accuracy are difficult to balance is relieved.

Description

Technical Field

[0001] The present application relates to the field of air quality prediction, and in particular to a method and device for rapid prediction of air quality for large-scale sparse sampling scenarios. Background Art

[0002] With the development of industrialization and urbanization, PM 2.5 Air pollution problems such as sulfur dioxide (SO2) and nitrogen dioxide (NO2) are becoming increasingly serious, posing severe challenges to the sustainable development of human society. Studies have shown that long-term exposure to high-pollution environments may increase the incidence of respiratory diseases, cancer and cardiovascular diseases. In this context, real-time and accurate air quality prediction models have become an important means to reduce human exposure to air pollution. These technologies can predict changes in pollutant concentrations and fluctuations in air quality indexes in advance, providing reference and reference for the public and the government to formulate targeted protective measures.

[0003] The key to air quality prediction is to mine the complex spatiotemporal relationships in historical data, so as to accurately and efficiently infer future air quality data. In recent years, deep learning models have become the mainstream air quality prediction models with their excellent ability to capture complex spatiotemporal relationships. Although existing models have achieved satisfactory results in most air quality prediction scenarios, they still face challenges in large-scale sparse sampling scenarios with high real-time requirements. Specifically, air quality sensors are usually sparsely distributed in large-scale sparse sampling scenarios. The common compact grid structure model is difficult to effectively capture the spatiotemporal dependencies in sparse images, which seriously affects the prediction accuracy of the model. Although the common graph structure model can obtain better prediction accuracy in sparse scenarios, the complexity of the model often increases exponentially with the increase in the number of nodes, which seriously affects the computational efficiency of the model. Therefore, in large-scale sparse sampling scenarios, it is difficult for existing models to strike a balance between large-scale coverage and real-time response, and there is a problem that computing resources and prediction accuracy cannot be taken into account.

[0004] Therefore, there is an urgent need for a method and device for rapid air quality prediction for large-scale sparse sampling scenarios. Summary of the invention

[0005] The present application provides a method for rapid air quality prediction for large-scale sparse sampling scenarios, which solves the problem that existing models are difficult to strike a balance between large-scale coverage and real-time response, and there is a problem that computing resources and prediction accuracy cannot be taken into account at the same time.

[0006] In a first aspect of the present application, a method for rapid air quality prediction for large-scale sparse sampling scenarios is provided, the method comprising: obtaining air quality data through an air quality sensor group, the air quality sensor group comprising multiple air quality sensors; obtaining a graph structure corresponding to the air quality sensor group according to the positional relationship of the multiple air quality sensors in the study area, the graph structure comprising a graph node connection relationship between the multiple air quality sensors; constructing a data structure conversion component, the data structure conversion component being used to convert the graph structure into a compact grid structure according to the graph node connection relationship between the multiple air quality sensors, and constructing a mapping relationship between the graph structure and the compact grid structure; constructing a lightweight parameter-sharing spatiotemporal expansion convolutional network component, the lightweight parameter-sharing spatiotemporal expansion convolutional network component being used to capture the spatiotemporal dependencies in air quality data; constructing an integrated predictor according to the mapping relationship and the spatiotemporal dependencies through the data structure conversion component and the lightweight parameter-sharing spatiotemporal expansion convolutional network component; and obtaining air quality prediction results according to the air quality data through the integrated predictor.

[0007] Optionally, according to the graph node connection relationship between multiple air quality sensors, the graph structure is converted into a compact grid structure, specifically including: obtaining a first starting graph node corresponding to the target air quality sensor in the graph structure, where the target air quality sensor is any one of the multiple air quality sensors; obtaining the central pixel of the compact grid structure, and constructing a first mapping relationship between the first starting graph node and the central pixel; according to the graph node connection relationship, if it is confirmed that the first-order adjacent graph node corresponding to the first starting graph node is obtained, then obtaining the first-order neighboring pixels corresponding to the central pixel, and constructing a second mapping relationship between the first-order adjacent graph node and the first-order neighboring pixels; obtaining the second starting graph node, where the second starting graph node is any one of the first-order adjacent graph nodes corresponding to the first starting graph node, and according to the graph node connection relationship, if it is confirmed that the first-order adjacent graph node corresponding to the second starting graph node is not obtained, then constructing a compact grid structure based on the central pixel and the first-order neighboring pixels.

[0008] Optionally, constructing a mapping relationship between the graph structure and the compact grid structure specifically includes: constructing a mapping relationship between the graph structure and the compact grid structure according to the first mapping relationship and the second mapping relationship.

[0009] Optionally, capturing the spatiotemporal dependencies in the air quality data specifically includes: obtaining air quality data corresponding to multiple time windows respectively; constructing a three-dimensional tensor corresponding to the air quality data through a data structure conversion component; and a lightweight parameter-sharing spatiotemporal expansion convolutional network component capturing the spatiotemporal dependencies in the air quality data through a three-dimensional tensor.

[0010] Optionally, based on the air quality data, air quality prediction results are obtained through an integrated predictor, specifically including: obtaining a current time window from multiple time windows; obtaining multiple historical time windows, multiple historical time windows are located before the current time window; obtaining historical air quality data corresponding to multiple historical time windows respectively, and the integrated predictor obtains air quality prediction results based on the historical air quality data.

[0011] Optionally, the air quality prediction result is obtained according to the following formula:

[0012]

[0013] Among them, M is a compact grid structure, LiEnPred←G is the construction process of the compact grid structure, G is a graph structure, and v i is the i-th air quality sensor in the air quality sensor group, A is the graph node connection relationship between air quality sensors, A∈R N×N , R is a real number, N is the total number of multiple air quality sensors, To construct a three-dimensional tensor based on a compact grid structure, t is the t-th time window, is historical air quality data, and Where P and Q represent the length and width of the compact grid structure, respectively, B is the history-dependent step length, and b is any time window in [1, B]; is the air quality prediction result, and F is the prediction step size, and f is any time window in [1,F].

[0014] Optionally, obtain a dilation factor; and reduce the depth of the neural network in the lightweight parameter sharing spatiotemporal dilated convolutional network component according to the dilation factor and the following formula:

[0015]

[0016] In the above formula,

[0017]

[0018] Among them, STPDCN is the forward propagation process of the shared spatiotemporal dilation convolutional network component. For input of historical air quality data; For predicted air quality data; are the expansion factors of Q, P, and T dimensions in the lth hidden layer respectively; is the convolution kernel shared in the hidden layer, where K T ,K P ,KQ are the convolution kernel sizes of T, P, and Q dimensions respectively; is the hidden state of the air quality data in the lth hidden layer, E is the dimension of the hidden layer, which also corresponds to the number of convolution kernels; is the three-dimensional tensor obtained by the e-th convolution kernel; is a three-dimensional tensor The elements in, p, q, t are three-dimensional tensors The coordinates of P, Q, and T dimensions; is a three-dimensional tensor The elements in, where i, j, and s are three-dimensional tensors The coordinates of the P, Q, and T dimensions in the vector. Conv is a convolution operation used for dimension alignment.

[0019] In a second aspect of the present application, a device for rapid prediction of air quality for large-scale sparse sampling scenarios is provided, the device comprising an acquisition module, an integrated predictor building module, and a prediction module, wherein:

[0020] An acquisition module is used to acquire air quality data through an air quality sensor group, where the air quality sensor group includes multiple air quality sensors; based on the positional relationship of the multiple air quality sensors in the study area, a graph structure corresponding to the air quality sensor group is acquired, where the graph structure includes a graph node connection relationship between the multiple air quality sensors.

[0021] An integrated predictor building module is used to build a data structure conversion component. The data structure conversion component is used to convert the graph structure into a compact grid structure according to the graph node connection relationship between multiple air quality sensors, and to construct a mapping relationship between the graph structure and the compact grid structure; to build a lightweight parameter sharing space-time dilation convolutional network component. The lightweight parameter sharing space-time dilation convolutional network component is used to capture the space-time dependencies in air quality data; and to build an integrated predictor based on the mapping relationship and space-time dependencies through the data structure conversion component and the lightweight parameter sharing space-time dilation convolutional network component.

[0022] The prediction module is used to obtain air quality prediction results through an integrated predictor based on air quality data.

[0023] In the third aspect of the present application, an electronic device is provided, including a processor, a memory, a user interface and a network interface, the memory is used to store instructions, the user interface and the network interface are used to communicate with other devices, and the processor is used to execute the instructions stored in the memory so that the electronic device executes any one of the methods described above.

[0024] In a fourth aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and the computer program is executed by a processor to perform any of the above methods.

[0025] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0026] 1. The present application constructs a graph structure corresponding to a large number of air quality sensors in large-scale sparse scenarios, and constructs a data structure conversion component. The data structure conversion component converts the graph structure into a compact grid structure according to the graph node connection relationship between the multiple air quality sensors, and constructs a lightweight parameter sharing space-time dilation convolutional network component. The lightweight parameter sharing space-time dilation convolutional network component captures the space-time dependency in the air quality data, and thus constructs an integrated predictor based on the mapping relationship and space-time dependency, and through the data structure conversion component and the lightweight parameter sharing space-time dilation convolutional network component, and then by combining the compact grid structure and the graph structure, the prediction accuracy of the integrated predictor in large-scale sparse scenarios is significantly improved.

[0027] 2. This application introduces a dilation factor and shares convolution kernels between different layers, so that the depth of the convolutional network can be greatly reduced without losing prediction accuracy, and ensures that the parameter scale of building an integrated predictor will not increase with the increase in the number of neural network layers.

[0028] 3. This application constructs a compact grid structure transformed from a graph structure, and infers the air quality prediction results in future time windows based on the historical air quality data corresponding to past time windows, thereby making the prediction accuracy more accurate and further improving the prediction accuracy of the integrated predictor in large-scale sparse scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 is a distribution diagram of air quality sensors in a large-scale scene provided by an embodiment of the present application;

[0030] Figure 2 It is a flow chart of a method for rapid prediction of air quality for large-scale sparse sampling scenarios provided in an embodiment of the present application;

[0031] Figure 3 This is a diagram of a graph structure-compact grid structure transformation provided by an embodiment of the present application;

[0032] Figure 4 This is another diagram of the conversion of a graph structure to a compact grid structure provided in an embodiment of the present application;

[0033] Figure 5It is a schematic diagram of a 3D convolutional network operation provided in an embodiment of the present application;

[0034] Figure 6 It is a scatter diagram of calculation efficiency and prediction accuracy provided by an embodiment of the present application;

[0035] Figure 7 It is a scatter diagram of parameter scale and prediction accuracy provided in an embodiment of the present application;

[0036] Figure 8 It is a module schematic diagram of an air quality rapid prediction device for large-scale sparse sampling scenarios provided in an embodiment of the present application;

[0037] Fig. 9 It is a structural schematic diagram of an electronic device provided in an embodiment of the present application.

[0038] Explanation of the reference numerals: 81, acquisition module; 82, integrated predictor building module; 83, prediction module; 901, processor; 902, communication bus; 903, user interface; 904, network interface; 905, memory. DETAILED DESCRIPTION

[0039] In order to enable technicians in this field to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments.

[0040] The terms used in the following embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to be used as limitations to the present application. As used in the specification of the present application, the singular expressions "one", "a kind of", "said", "above", "the" and "this" are intended to also include plural expressions, unless there is a clear indication to the contrary in the context. It should also be understood that the term "and / or" used in the present application refers to and includes any or all possible combinations of one or more listed items.

[0041] In the following, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as suggesting or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features, and in the description of the embodiments of the present application, unless otherwise specified, "plurality" means two or more.

[0042] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings.

[0043] Research has found that existing models face challenges in large-scale sparse sampling scenarios with high real-time requirements. Figure 1 , which shows a distribution map of air quality sensors in a large-scale scene provided by an embodiment of the present application. Air quality sensors are usually sparsely distributed in large-scale scenes. The grid model commonly used for large-scale predictions is difficult to effectively capture the spatiotemporal dependencies in sparse images, which seriously affects the prediction accuracy of the model. Although graph-based models can achieve better prediction accuracy in sparse scenes, the complexity of the model often grows exponentially with the increase in the number of nodes, which seriously affects the computational efficiency of the model. In actual scenarios, real-time large-scale sparse sampling predictions are crucial to public health protection and environmental sustainability. However, existing models find it difficult to strike a balance between large-scale coverage and real-time response, and face challenges in computing resources and accuracy.

[0044] Therefore, this application proposes a method and device for rapid prediction of air quality for large-scale sparse sampling scenarios. It should be noted that in the embodiments of this application, by collecting air quality data from more than 2,000 stations in China in the past three years, and taking PM2.5 concentration as an example, the prediction performance of the method in this application in large-scale sparse distribution scenarios is verified.

[0045] Please refer to Figure 2 , which shows a flow chart of a method for rapid prediction of air quality for large-scale sparse sampling scenarios provided in an embodiment of the present application. The method is applied to a server, and the flow chart mainly includes the following steps: S201 to S206.

[0046] Step S201, obtaining air quality data through an air quality sensor group, where the air quality sensor group includes a plurality of air quality sensors.

[0047] Specifically, the air quality sensor group is carried out according to the specific needs of the monitoring area, and usually involves the deployment of multiple air quality sensors. In this embodiment, the air quality sensor group is distributed in different geographical locations across the country, that is, the air quality sensors in more than 2,000 sites in China in the past three years. The types of air quality sensors include but are not limited to monitoring equipment for various air quality indicators such as PM2.5, PM10, nitrogen dioxide (NO2), carbon monoxide (CO), and ozone (O3). Each air quality sensor will continuously collect air quality data according to the preset sampling frequency and data transmission strategy. In this embodiment, the PM2.5 concentration is mainly used as an example for explanation. The above air quality data is usually stored in the form of a time series, and the air quality data collected by each air quality sensor is accompanied by a clear timestamp, so that the data collected by different air quality sensors can be accurately aligned in the subsequent analysis process.

[0048] Step S202: acquiring a graph structure corresponding to the air quality sensor group according to the position relationship of the plurality of air quality sensors in the study area, wherein the graph structure includes a graph node connection relationship between the plurality of air quality sensors.

[0049] Specifically, all air quality sensors in the study area form a graph structure where v i ∈V represents the i-th air quality sensor under the graph structure, A∈R N×N represents the graph node connection relationship or dependency relationship between air quality sensors, and N represents the total number of all air quality sensors.

[0050] Step S203: construct a data structure conversion component, which is used to convert the graph structure into a compact grid structure according to the graph node connection relationship between multiple air quality sensors, and construct a mapping relationship between the graph structure and the compact grid structure.

[0051] Specifically, please refer to Figure 3 , which shows a diagram of a graph structure-compact grid structure transformation provided by an embodiment of the present application, wherein a data structure transformation component (DST) generates a compact grid structure N∈R by using a graph structure between air quality sensors. P×Q , where m pq >0 means the pixel in the pth row and qth column corresponds to an air quality sensor (m pq equal to the index of the air quality sensor in the graph structure), m pq =0 indicates that the pixel in the p-th row and q-th column in the compact grid structure has no corresponding air quality sensor, P and Q represent the total length and width of the compact graph, respectively, and during the conversion process, the data structure conversion component establishes a mapping relationship between the graph structure and the compact grid structure.

[0052] In a possible implementation, step 203 also includes: obtaining a first starting graph node corresponding to the target air quality sensor in the graph structure, where the target air quality sensor is any one of a plurality of air quality sensors; obtaining a central pixel of a compact grid structure, and constructing a first mapping relationship between the first starting graph node and the central pixel; according to the graph node connection relationship, if it is confirmed that the first-order adjacent graph node corresponding to the first starting graph node is obtained, then obtaining the first-order neighboring pixels corresponding to the central pixel, and constructing a second mapping relationship between the first-order adjacent graph node and the first-order neighboring pixels; obtaining a second starting graph node, where the second starting graph node is any one of the first-order adjacent graph nodes corresponding to the first starting graph node, and according to the graph node connection relationship, if it is confirmed that the first-order adjacent graph node corresponding to the second starting graph node is not obtained, then constructing a compact grid structure based on the central pixel and the first-order neighboring pixels.

[0053] Specifically, please refer to Figure 4 , which shows another diagram of the conversion of graph structure-compact grid structure provided by an embodiment of the present application. The data structure conversion component establishes a mapping relationship between graph nodes and grid pixels. Inspired by the breadth-first search algorithm, the data structure conversion component first determines the first starting graph node and the coordinates of the first starting graph node in the compact grid structure, that is, the first starting graph node corresponding to the target air quality sensor in the graph structure. The target air quality sensor is any one of the multiple air quality sensors. In this embodiment, for ease of explanation, we can use the node at the center of the graph structure and the corresponding air quality sensor as the first starting graph node and the target air quality sensor, and use the first starting graph node as the center pixel for constructing a compact grid structure. Afterwards, according to the connection relationship between the graph nodes, the center pixel is expanded outward layer by layer, and the coordinates of the remaining graph nodes in the compact grid are determined in turn. As Figure 4 As shown in a, first, we map the first starting graph node v4 to the central pixel of the compact grid structure, and determine whether the first starting graph node has a first-order adjacent graph node. The first-order adjacent graph node is a graph node directly connected to the first starting graph node, that is, Figure 4 v2, v5, v6, and v8 in the figure correspond to the first-order neighboring graph nodes of v4 as the first-order neighboring pixels of the center pixel in a clockwise manner to form the following Figure 4 The compact grid structure shown in a in the figure, at this time, the first mapping relationship between the mapped first-order neighbor pixels and the first-order neighboring graph nodes is constructed. Then, we regard each first-order neighboring graph node of v4 as the central node in turn, that is, the first-order neighboring graph node corresponding to any first starting graph node can be used as the second starting graph node, and the second-order neighboring graph nodes of v4 are determined. The second-order neighboring graph nodes are graph nodes directly connected to the first-order neighboring graph nodes. Taking v2 as the second starting graph node as an example, the second-order neighboring graph nodes are v1 and v3 (for v2, v1 and v3 are its first-order neighboring graph nodes), and the first-order neighboring graph nodes of v2 are corresponded to the first-order neighboring pixels of the center pixel (the center pixel is v2 at this time) in clockwise order to form the following: Figure 4 The compact grid structure shown in b in Figure 1 constructs the second mapping relationship between the mapped second-order neighboring pixels and the second-order adjacent graph nodes. When the eight neighboring pixels of the target node in the compact grid structure cannot be used to place the graph node, we will expand outward from the target node until we find a pixel that can be used to place the graph node, as shown in Figure 1. Figure 4As shown in c and d. Repeat the above operation until all graph nodes are mapped to the compact grid structure. In addition, the data structure conversion component constructs a mapping relationship between 1-n order adjacent graph nodes in the compact grid structure, that is, a first mapping relationship, a second mapping relationship, ..., an nth mapping relationship, and uses all first mapping relationships, second mapping relationships, ..., and nth mapping relationships as mapping relationships between the graph structure and the compact grid structure. In an embodiment of the present application, all mapping relationships of adjacent graph nodes in the compact grid structure, that is, 1-nth mapping relationships, are represented by the first mapping relationship and the second mapping relationship.

[0054] Step S204, constructing a lightweight parameter-sharing spatiotemporal dilation convolutional network component, which is used to capture the spatiotemporal dependencies in the air quality data.

[0055] Specifically, through the data structure conversion component, the air quality data of T time windows will form a three-dimensional tensor In the air quality prediction scenario with high real-time requirements, it is necessary to quickly and accurately Mining the spatiotemporal dependencies in air quality data is crucial. However, most existing prediction models have difficulty balancing prediction accuracy and computational efficiency. Models with high prediction accuracy often face problems such as complex model design process and low computational efficiency, making it difficult to meet the needs of real-time prediction. In view of this, we propose a lightweight parameter-shared spatio-temporal dilation convolution network component (STP-DCN) to address the above challenges.

[0056] In a possible implementation, step 204 also includes: obtaining air quality data corresponding to multiple time windows respectively; constructing a three-dimensional tensor corresponding to the air quality data through a data structure conversion component; and a lightweight parameter sharing spatiotemporal expansion convolutional network component capturing the spatiotemporal dependencies in the air quality data through a three-dimensional tensor.

[0057] Specifically, in the compact grid image, air pollutants (such as PM 2.5 concentration, NO2 concentration, CO concentration, etc.) to form a three-dimensional tensor The lightweight parameter-sharing spatiotemporal dilated convolutional network component captures the spatiotemporal dependencies in air quality data via a three-dimensional tensor.

[0058] Step S205, constructing an integrated predictor through a data structure conversion component and a lightweight parameter sharing spatiotemporal expansion convolutional network component according to the mapping relationship and the spatiotemporal dependency.

[0059] Specifically, the integrated predictor consists of a data structure conversion component and a lightweight parameter-sharing spatiotemporal dilation convolutional network component. The data structure conversion component is used to establish the function mapping relationship between graph nodes and compact grid pixels; the lightweight parameter-sharing spatiotemporal dilation convolutional network component is used to mine the spatiotemporal dependencies in compact grid pixels.

[0060] Step S206, obtaining an air quality prediction result through an integrated predictor according to the air quality data.

[0061] Specifically, the air pollutants collected in T time windows represent the air quality data corresponding to the multiple time windows, where X t ∈R P×Q represents the air pollutants collected by the air quality sensor group in the tth time window, x pqt ∈R 1×1 Indicates air quality sensor m pq Air pollutants collected in the tth time window. For any t∈[1,T], it is necessary to use the historical three-dimensional tensor based on the compact grid structure M. Quickly and efficiently infer future air quality The details are shown in the following formula.

[0062]

[0063] Where M is a compact grid structure, LiEnPred←G is the construction process of the compact grid structure, G is a graph structure, and v i is the i-th air quality sensor in the air quality sensor group, A is the graph node connection relationship between air quality sensors, A∈R N×N , R is a real number, N is the total number of multiple air quality sensors, To construct a three-dimensional tensor based on a compact grid structure, t is the t-th time window, is historical air quality data, and Where P and Q represent the length and width of the compact grid structure, respectively, B is the history-dependent step length, and b is any time window in [1, B]; is the air quality prediction result, and F is the prediction step size, and f is any time window in [1, F]. Through the above method, the air quality prediction results in the future time window are inferred based on the historical air quality data corresponding to the past time window, so that the prediction accuracy is more accurate, and the prediction accuracy of the integrated predictor in large-scale sparse scenes is further improved.

[0064] In a possible implementation, step S204 also includes: obtaining a current time window among multiple time windows; obtaining multiple historical time windows, wherein the multiple historical time windows are located before the current time window; obtaining historical air quality data corresponding to the multiple historical time windows respectively, and the integrated predictor obtains air quality prediction results based on the historical air quality data.

[0065] Specifically, please refer to Figure 5 , which shows a 3D convolutional network operation schematic diagram provided by an embodiment of the present application, wherein a lightweight parameter sharing space-time dilation convolutional network component is extended from a classic 3D convolutional network. Compared with the classic 3D convolutional network, the lightweight parameter sharing space-time dilation convolutional network component constrains the convolution operation to meet the principle of front-back dependency in the time dimension, ensuring that the convolution kernel can only operate the time window before the current moment, that is, multiple historical time windows are located before the current time window. Secondly, we introduced an expansion factor based on the classic 3D convolutional network. According to the expansion factor, the depth of the neural network in the lightweight parameter sharing space-time dilation convolutional network component is reduced, so that the depth of the neural network is greatly reduced without losing the prediction accuracy. Finally, the lightweight parameter sharing space-time dilation convolutional network component includes multiple neurons, and multiple neurons share the same set of convolutional kernels, which ensures that the parameter scale of the model will not increase with the increase of the number of neural network layers. Compared with the classic 3D convolutional network, the lightweight parameter sharing space-time dilation convolutional network component module has three main advantages. First, the lightweight parameter sharing spatiotemporal dilation convolutional network component satisfies the basic constraints of time order in the time dimension, making the convolution operation in the lightweight parameter sharing spatiotemporal dilation convolutional network component more reasonable than the convolution operation in the classic 3D convolutional network. Secondly, the lightweight parameter sharing spatiotemporal dilation convolutional network component module controls the depth of the neural network through the dilation factor, making the computational efficiency of the lightweight parameter sharing spatiotemporal dilation convolutional network component higher than that of the classic 3D convolutional network. Finally, the lightweight parameter sharing spatiotemporal dilation convolutional network component shares the convolution kernels between different layers, making the parameter scale of the lightweight parameter sharing spatiotemporal dilation convolutional network component smaller than the parameter scale of the classic 3D convolutional network. With a three-dimensional tensor As an example, the following formula shows the forward propagation process of the lightweight parameter sharing spatiotemporal expansion convolutional network component:

[0066]

[0067] Among them, STPDCN is the forward propagation process of the shared spatiotemporal dilation convolutional network component. For input of historical air quality data; For predicted air quality data; are the expansion factors of Q, P, and T dimensions in the lth hidden layer respectively; is the convolution kernel shared in the hidden layer, where K T ,K P ,K Q are the convolution kernel sizes of T, P, and Q dimensions respectively; is the hidden state of the air quality data in the lth hidden layer, E is the dimension of the hidden layer, which also corresponds to the number of convolution kernels; is the three-dimensional tensor obtained by the e-th convolution kernel; is a three-dimensional tensor The elements in, p, q, t are three-dimensional tensors The coordinates of P, Q, and T dimensions; is a three-dimensional tensor The elements in, where i, j, and s are three-dimensional tensors The coordinates of the P, Q, and T dimensions in . Conv is a convolution operation used for dimension alignment.

[0068] In addition, in the forward propagation process of the lightweight parameter sharing spatiotemporal expansion convolutional network component in step S204, the ensemble predictor (Lightweight Ensemble Predictor, LiEnPred) uses the air quality data of the first B time windows Predict the air quality data for F time windows in the future In general, directly optimizing the true value and predicted values The trained prediction model can be obtained by calculating the square loss between . However, considering that the compact grid structure matrix M may not be completely filled with graph nodes, we only include the loss of some positions into the loss function, as shown in the following formula:

[0069]

[0070] in, Indicates air quality sensor m pq The observation value in the tb-th time window (ground truth); Indicates air quality sensor m pq The predicted value in the tb-th time window; Indicates that only m pq >0, we will calculate the loss represents the learnable weights in the ensemble predictor.

[0071] It should be noted that in this application, we use real large-scale sparse sampling data to evaluate the prediction performance of the LiEnPred method. The air quality data comes from 2014 monitoring sensors in China, including seven indicators such as PM2.5, CO, NO2, PM10, SO2, O3, and AQI. Among them, the sampling interval of the seven indicators is 60 minutes, and the time span is from January 1, 2021 to December 31, 2023. Considering that the integrated predictor is a universal air quality prediction model, we only used PM2.5 concentration data to verify the prediction performance of the integrated predictor.

[0072] The benchmark models in this application can be roughly divided into two categories. The former is a spatiotemporal prediction model based on a compact grid structure, including the ConvLSTM model, the ST-3DNet model, and the ST-3DMDDN model; the latter is a spatiotemporal prediction model based on a graph structure, including the DSTAGNN model and the STGDN model.

[0073] In order to fairly compare the prediction accuracy of the compact grid structure model and the graph structure model, this application deals with the correspondence between pixels and sensors in sparse grids. Specifically, we use a 200×200 grid to divide the study area and ensure that one pixel corresponds to one sensor (if one pixel corresponds to multiple sensors, we correspond the sensor to the adjacent pixels). The prediction accuracy results of the integrated predictor and the benchmark model are shown in Table 1.

[0074] [Table 1]

[0075] Evaluation metrics(RMSE / MAPE) of LiEnPred and baselines.

[0076]

[0077] The results show that the prediction accuracy of the second type of model is better than that of the first type of model, that is, the prediction effect of the graph structure model is better than that of the compact grid structure model. This is because the graph structure model explicitly defines the spatial dependency between sensors, thereby improving the prediction ability. However, the compact grid structure model is difficult to capture this spatial dependency in a sparse grid, which reduces the prediction effect. In addition, although the ensemble predictor is a compact grid structure model, its prediction accuracy is better than that of the first type of model, and it is close to or exceeds the accuracy of the second type of model in multi-step prediction. This shows that without considering the computational efficiency and parameter scale, the ensemble predictor has achieved a prediction accuracy that is better than or comparable to the baseline model. The computational efficiency of the ensemble predictor is further analyzed, and the results are shown in Table 2. Among them, the one-time inference time is mainly used to reflect the online prediction speed of the ensemble predictor, and the one-time optimization time is mainly used to reflect the offline training speed of the ensemble predictor. The results show that the computational efficiency of the ensemble predictor is not only better than that of the graph structure model, but also better than that of the classic compact grid structure model.

[0078] [Table 2]

[0079] Running time(milliseconds)of LiEnPred and baselines with batch size being 8.

[0080]

[0081] In addition, please refer to Figure 6 , which shows a scatter diagram of calculation efficiency and prediction accuracy provided by an embodiment of the present application. Figure 6 In the example, the closer the scattered points are to the center of the coordinates, the better the computational efficiency and prediction accuracy of the model. The results show that the ensemble predictor takes into account the prediction accuracy and computational efficiency of the model to a certain extent, especially in large-scale sparse scenarios. Similar to computational efficiency, we analyzed the parameter scale of the learnable weights in the ensemble predictor, and the results are shown in Table 3. The results show that the ensemble predictor significantly outperforms the ConvLSTM model, ST-3DNet model, ST-3DMDDN model, and DSTAGNN model in terms of parameter scale, and is close to the lightweight STGDN model.

[0082] [Table 3]

[0083] Parameter scale(kilobyte)of learnable weights for LiEnPred andbaselines.

[0084]

[0085]

[0086] In addition, please refer to Figure 7 , which shows a scatter diagram of parameter scale and prediction accuracy provided by an embodiment of the present application. Figure 7 In the above figure, the closer the scatter points are to the center of the coordinate, the better the prediction accuracy of the model and the lower the parameter scale. The results show that the integrated predictor achieves acceptable prediction accuracy without significantly increasing the parameter scale of the model.

[0087] The present application adopts the above method, constructs a graph structure corresponding to a large number of air quality sensors in a large-scale sparse scene, and constructs a data structure conversion component. The data structure conversion component converts the graph structure into a compact grid structure according to the graph node connection relationship between the multiple air quality sensors, and constructs a lightweight parameter sharing space-time expansion convolutional network component at the same time. The lightweight parameter sharing space-time expansion convolutional network component captures the space-time dependency in the air quality data compared to the classic prediction model based on the compact grid structure, thereby constructing an integrated predictor based on the mapping relationship and space-time dependency through the data structure conversion component and the lightweight parameter sharing space-time expansion convolutional network component, and then by combining the compact grid structure and the graph structure, the prediction accuracy of the integrated predictor in large-scale sparse scenes is significantly improved.

[0088] Please refer to Figure 8 , which shows a module schematic diagram of an air quality rapid prediction device for large-scale sparse sampling scenarios provided by an embodiment of the present application, the device includes an acquisition module 81, an integrated predictor construction module 82, and a prediction module 83, wherein:

[0089] The acquisition module 81 is used to acquire air quality data through an air quality sensor group, where the air quality sensor group includes multiple air quality sensors; and acquire a graph structure corresponding to the air quality sensor group, where the graph structure includes a graph node connection relationship between the multiple air quality sensors.

[0090] The integrated predictor construction module 82 is used to obtain a data structure conversion component, which is used to convert the graph structure into a compact grid structure according to the graph node connection relationship between multiple air quality sensors, and construct a mapping relationship between the graph structure and the compact grid structure; obtain a lightweight parameter sharing space-time expansion convolutional network component, which is used to capture the space-time dependencies in air quality data; and construct an integrated predictor based on the mapping relationship and space-time dependencies through the data structure conversion component and the lightweight parameter sharing space-time expansion convolutional network component.

[0091] The prediction module 83 is used to obtain air quality prediction results through an integrated predictor based on the air quality data.

[0092] In one possible implementation, the integrated predictor construction module 82 is used to convert the graph structure into a compact grid structure based on the graph node connection relationship between multiple air quality sensors, specifically including: obtaining a first starting graph node corresponding to the target air quality sensor in the graph structure, where the target air quality sensor is any one of the multiple air quality sensors; obtaining the central pixel of the compact grid structure, and constructing a first mapping relationship between the first starting graph node and the central pixel; according to the graph node connection relationship, if it is confirmed that the first-order adjacent graph node corresponding to the first starting graph node is obtained, then obtaining the first-order neighboring pixels corresponding to the central pixel, and constructing a second mapping relationship between the first-order adjacent graph node and the first-order neighboring pixels; obtaining the second starting graph node, where the second starting graph node is any one of the first-order adjacent graph nodes corresponding to the first starting graph node, and according to the graph node connection relationship, if it is confirmed that the first-order adjacent graph node corresponding to the second starting graph node is not obtained, then constructing a compact grid structure based on the central pixel and the first-order neighboring pixels.

[0093] In a possible implementation, the integrated predictor construction module 82 is used to construct a mapping relationship between the graph structure and the compact grid structure, specifically including: constructing a mapping relationship between the graph structure and the compact grid structure according to the first mapping relationship and the second mapping relationship.

[0094] In one possible implementation, the integrated predictor construction module 82 is used to capture the spatiotemporal dependencies in air quality data, specifically including: obtaining air quality data corresponding to multiple time windows; constructing a three-dimensional tensor corresponding to the air quality data through a data structure conversion component; and a lightweight parameter-sharing spatiotemporal expansion convolutional network component capturing the spatiotemporal dependencies in the air quality data through a three-dimensional tensor.

[0095] In a possible implementation, the prediction module 83 is used to obtain air quality prediction results based on air quality data through an integrated predictor, specifically including: obtaining a current time window from multiple time windows; obtaining multiple historical time windows, multiple historical time windows are located before the current time window; obtaining historical air quality data corresponding to multiple historical time windows respectively, and the integrated predictor obtains air quality prediction results based on the historical air quality data.

[0096] In a possible implementation, the prediction module 83 is used to obtain the air quality prediction result according to the following formula:

[0097]

[0098] Among them, M is a compact grid structure, LiEnPred←G is the construction process of the compact grid structure, G is a graph structure, and v iis the i-th air quality sensor in the air quality sensor group, A is the graph node connection relationship between air quality sensors, A∈R N×N , R is a real number, N is the total number of multiple air quality sensors, To construct a three-dimensional tensor based on a compact grid structure, t is the t-th time window, is historical air quality data, and Where P and Q represent the length and width of the compact grid structure, respectively, B is the history-dependent step length, and b is any time window in [1, B]; is the air quality prediction result, and F is the prediction step size, and f is any time window in [1,F].

[0099] In one possible implementation, the prediction module 83 is used to obtain a dilation factor; based on the dilation factor, the depth of the neural network in the lightweight parameter sharing spatiotemporal dilation convolutional network component is reduced according to the following formula:

[0100]

[0101] In the above formula,

[0102]

[0103] Among them, STPDCN is the forward propagation process of the shared spatiotemporal dilation convolutional network component. For input of historical air quality data; For predicted air quality data; are the expansion factors of Q, P, and T dimensions in the lth hidden layer respectively; is the convolution kernel shared in the hidden layer, where K T ,K P ,K Q are the convolution kernel sizes of T, P, and Q dimensions respectively; is the hidden state of the air quality data in the lth hidden layer, E is the dimension of the hidden layer, which also corresponds to the number of convolution kernels; is the three-dimensional tensor obtained by the e-th convolution kernel; is a three-dimensional tensor The elements in, p, q, t are three-dimensional tensors The coordinates of P, Q, and T dimensions; is a three-dimensional tensor The elements in, where i, j, and s are three-dimensional tensors The coordinates of the P, Q, and T dimensions in . Conv is a convolution operation used for dimension alignment.

[0104] It should be noted that: when the device provided in the above embodiment realizes its function, only the division of the above functional modules is used as an example. In actual application, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the device and method embodiments provided in the above embodiment belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0105] The present application also provides an electronic device. Fig. 9 , Fig. 9 9 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device may include: at least one processor 901, at least one communication bus 902, a user interface 903, at least one network interface 904, and a memory 905.

[0106] The communication bus 902 is used to realize the connection and communication between these components.

[0107] The user interface 903 may include a display screen (Display) and a camera (Camera), and the optional user interface 903 may also include a standard wired interface and a wireless interface.

[0108] The network interface 904 may optionally include a standard wired interface or a wireless interface (such as a WI-FI interface).

[0109] Among them, the processor 901 may include one or more processing cores. The processor 901 uses various interfaces and lines to connect various parts in the entire server, and executes various functions of the server and processes data by running or executing instructions, programs, code sets or instruction sets stored in the memory 905, and calling data stored in the memory 905. Optionally, the processor 901 can be implemented in at least one hardware form of digital signal processing (Digital Signal Processing, DSP), field programmable gate array (Field-Programmable Gate Array, FPGA), and programmable logic array (Programmable Logic Array, PLA). The processor 901 can integrate one or more combinations of a central processing unit (Central Processing Unit, CPU), a graphics processing unit (Graphics Processing Unit, GPU) and a modem. Among them, the CPU mainly processes the operating system, user interface and application programs; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; the modem is used to process wireless communications. It can be understood that the above-mentioned modem may not be integrated into the processor 901, and it can be implemented by a single chip.

[0110] Among them, the memory 905 may include a random access memory (Random Access Memory, RAM) and may also include a read-only memory (Read-Only Memory). Optionally, the memory 905 includes a non-transitory computer-readable storage medium. The memory 905 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 905 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area may store data involved in the above-mentioned various method embodiments, etc. The memory 905 may also be optionally at least one storage device located away from the aforementioned processor 901. Refer to Fig. 9 , the memory 905 as a computer storage medium may include an operating system, a network communication module, a user interface module, and an air quality rapid prediction application for large-scale sparse sampling scenarios.

[0111] exist Fig. 9In the electronic device shown, the user interface 903 is mainly used to provide an input interface for the user and obtain the data input by the user; and the processor 901 can be used to call the air quality rapid prediction application for large-scale sparse sampling scenarios stored in the memory 905. When executed by one or more processors 901, the electronic device executes one or more of the methods described in the above embodiments. It should be noted that for the aforementioned method embodiments, for the sake of simple description, they are all expressed as a series of action combinations, but technicians in this field should be aware that the present application is not limited to the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, technicians in this field should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required for this application.

[0112] The present application also provides a computer-readable storage medium, which stores instructions. When executed by one or more processors, the electronic device executes one or more of the methods described in the above embodiments.

[0113] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0114] In the several embodiments provided in this application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0115] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0116] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0117] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a memory and includes several instructions for a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the various embodiments of the present application. The aforementioned memory includes: various media that can store program codes, such as USB flash drives, mobile hard drives, magnetic disks or optical disks.

[0118] The above are only exemplary embodiments disclosed in this application, and the scope of the disclosure of this application cannot be limited by this. That is, any equivalent changes and modifications made according to the teachings disclosed in this application are still within the scope of the disclosure of this application. After considering the disclosure of the specification and the truth of practice, technicians in this field will easily think of other implementation schemes disclosed in this application.

[0119] This application is intended to cover any modifications, uses or adaptive changes disclosed in this application, which follow the general principles disclosed in this application and include common knowledge or customary technical means in the technical field not recorded in this application.

Claims

1. A rapid air quality prediction method for large-scale sparse sampling scenarios, characterized in that: The method comprises: Acquiring air quality data through an air quality sensor group, wherein the air quality sensor group includes a plurality of air quality sensors; According to the positional relationship of the plurality of air quality sensors in the study area, a graph structure corresponding to the air quality sensor group is acquired, wherein the graph structure includes a graph node connection relationship between the plurality of air quality sensors; Constructing a data structure conversion component, wherein the data structure conversion component is used to convert the graph structure into a compact grid structure according to the graph node connection relationship between the plurality of air quality sensors, and construct a mapping relationship between the graph structure and the compact grid structure; Constructing a lightweight parameter-sharing spatiotemporal dilation convolutional network component, wherein the lightweight parameter-sharing spatiotemporal dilation convolutional network component is used to capture the spatiotemporal dependencies in the air quality data; According to the mapping relationship and the spatiotemporal dependency, an integrated predictor is constructed through the data structure conversion component and the lightweight parameter sharing spatiotemporal expansion convolutional network component; According to the air quality data, an air quality prediction result is obtained through the integrated predictor.

2. The method according to claim 1, characterized in that The converting the graph structure into a compact grid structure according to the graph node connection relationship between the plurality of air quality sensors specifically includes: Obtaining a first starting graph node corresponding to a target air quality sensor in the graph structure, wherein the target air quality sensor is any one of the multiple air quality sensors; Acquire a central pixel of the compact grid structure, and construct a first mapping relationship between the first starting graph node and the central pixel; According to the graph node connection relationship, if it is confirmed that the first-order adjacent graph node corresponding to the first starting graph node is obtained, the first-order neighboring pixels corresponding to the central pixel are obtained, and a second mapping relationship between the first-order adjacent graph node and the first-order neighboring pixels is constructed; Obtain a second starting graph node, where the second starting graph node is any one of the first-order adjacent graph nodes corresponding to the first starting graph node, and based on the graph node connection relationship, if it is confirmed that the first-order adjacent graph node corresponding to the second starting graph node has not been obtained, construct the compact grid structure based on the central pixel and the first-order neighborhood pixels.

3. The method according to claim 2, characterized in that The constructing of a mapping relationship between the graph structure and the compact grid structure specifically includes: The mapping relationship between the graph structure and the compact grid structure is constructed according to the first mapping relationship and the second mapping relationship.

4. The method according to claim 1, characterized in that: The capturing of the spatiotemporal dependency in the air quality data specifically includes: Acquire the air quality data corresponding to each of the multiple time windows; Constructing a three-dimensional tensor corresponding to the air quality data through the data structure conversion component; The lightweight parameter-sharing spatiotemporal dilated convolutional network component captures the spatiotemporal dependencies in the air quality data through the three-dimensional tensor.

5. The method according to claim 4, characterized in that The step of obtaining an air quality prediction result by the integrated predictor according to the air quality data specifically includes: Obtaining a current time window from among the multiple time windows; Acquire multiple historical time windows, where the multiple historical time windows are located before the current time window; The historical air quality data corresponding to the plurality of historical time windows are obtained, and the integrated predictor obtains the air quality prediction result according to the historical air quality data.

6. The method according to claim 5, characterized in that The air quality prediction result is obtained according to the following formula: Where M is the compact grid structure, LiEnPred←G is the construction process of the compact grid structure, G is the graph structure, and v i is the i-th air quality sensor in the air quality sensor group, A is the graph node connection relationship between the air quality sensors, A∈R N×N , R is a real number, N is the total number of the plurality of air quality sensors, To construct the three-dimensional tensor based on the compact grid structure, t is the t-th time window, is the historical air quality data, and Where P and Q represent the length and width of the compact grid structure, respectively, B is the history-dependent step length, and b is any time window in [1, B]; is the air quality prediction result, and F is the prediction step size, and f is any time window in [1,F].

7. The method according to claim 6, characterized in that After obtaining the lightweight parameter sharing spatiotemporal dilation convolutional network component, the method further includes: Get the expansion factor; According to the dilation factor, the depth of the neural network in the lightweight parameter sharing spatiotemporal dilated convolutional network component is reduced according to the following formula: In the above formula, Among them, STPDCN is the forward propagation process of the shared spatiotemporal dilation convolutional network component, The historical air quality data for; The air quality prediction result; are the sizes of the dilation factors of the Q dimension, P dimension, and T dimension in the lth hidden layer respectively; is the convolution kernel shared in the hidden layer, where K T ,K P ,K Q are the convolution kernel sizes of T, P, and Q dimensions respectively; is the hidden state of the air quality data in the lth hidden layer, E is the dimension of the hidden layer, which also corresponds to the number of convolution kernels; is the three-dimensional tensor obtained by the e-th convolution kernel; is a three-dimensional tensor The elements in, p, q, t are the three-dimensional tensors The coordinates of P, Q, and T dimensions; is a three-dimensional tensor The elements in, where i, j, s are the three-dimensional tensors The coordinates of the P, Q, and T dimensions in the vector. Conv is a convolution operation used for dimension alignment.

8. A rapid air quality prediction device for large-scale sparse sampling scenarios, characterized in that: The device comprises an acquisition module, an integrated predictor construction module, and a prediction module, wherein: The acquisition module is used to acquire air quality data through an air quality sensor group, wherein the air quality sensor group includes a plurality of air quality sensors; and acquire a graph structure corresponding to the air quality sensor group according to a positional relationship of the plurality of air quality sensors in a study area, wherein the graph structure includes a graph node connection relationship between the plurality of air quality sensors; The integrated predictor construction module is used to construct a data structure conversion component, the data structure conversion component is used to convert the graph structure into a compact grid structure according to the graph node connection relationship between the multiple air quality sensors, and construct a mapping relationship between the graph structure and the compact grid structure; construct a lightweight parameter sharing space-time expansion convolutional network component, the lightweight parameter sharing space-time expansion convolutional network component is used to capture the space-time dependency relationship in the air quality data; according to the mapping relationship and the space-time dependency relationship, the integrated predictor is constructed through the data structure conversion component and the lightweight parameter sharing space-time expansion convolutional network component; The prediction module is used to obtain the air quality prediction result through the integrated predictor according to the air quality data.

9. An electronic device, characterized in that: It includes a processor, a communication bus, a user interface, a network interface and a memory, wherein the memory is used to store instructions, the user interface and the network interface are used to communicate with other devices, and the processor is used to execute the instructions stored in the memory so that the electronic device executes the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and when the instructions are executed, the method according to any one of claims 1 to 7 is performed.

Images