Regional lightning prediction method, device, equipment and medium

Abstract

The invention discloses a regional lightning prediction method, device and equipment and a medium, and belongs to the technical field of lightning prediction. The method comprises the following steps: collecting lightning drop point data and radar echo intensity data of a target area in a first time period; preprocessing the lightning drop point data and the radar echo intensity data; based on the preprocessed lightning drop point data, obtaining lattice lightning data of the target area; based on the lattice lightning data of the target area, constructing a lightning threat probability scoring matrix of the target area; and inputting the preprocessed radar echo intensity data and the lightning threat probability scoring matrix of the target area into a pre-trained area lightning prediction model, and obtaining a lightning prediction result of the target area in a second time period. According to the regional lightning prediction method provided by the invention, the lightning threat probability score and the radar echo data are taken as input, and good regional lightning risk prediction is realized through the machine learning model.

Description

Technical Field

[0001] The present invention belongs to the technical field of lightning forecasting, and relates to a method, device, equipment and medium for regional lightning forecasting. Background Art

[0002] Lightning is a strong electrical discharge phenomenon in the atmosphere, usually occurring in severe convective weather. It is caused by the accumulation and release of electric charges between clouds or between clouds and the ground, and its main characteristics include strong currents, extremely high temperatures, and bright light. Lightning not only releases huge amounts of energy in a short period of time, but is also accompanied by a violent sound phenomenon, namely thunder. Lightning has a significant impact on the environment and human activities, and is a serious natural disaster that can cause fires, damage buildings and electronic equipment, and even cause casualties, threatening social and economic development and human safety.

[0003] Previous lightning prediction studies mainly relied on the detected scattered lightning strike point information to assess lightning risk. These traditional methods usually make predictions based on the distribution of lightning strike points. However, due to the complexity and randomness of lightning phenomena, when lightning occurs, its current and energy can span thousands of meters, and its impact range far exceeds a single strike point. Therefore, predicting lightning threats based solely on strike point information often cannot fully reflect its actual risks to surrounding areas. In addition, the data correlation between meteorological elements is extremely complex, and traditional meteorological prediction methods usually rely on linear models or empirical rules, which are difficult to capture the potential nonlinear relationships and complex interaction effects in meteorological data. Summary of the invention

[0004] The purpose of the present invention is to overcome the deficiencies in the prior art and to provide a regional lightning prediction method, device, equipment and medium, which can accurately predict the probability and threat of regional lightning based on regional lightning threat probability scoring and machine learning models.

[0005] To achieve the above object, the present invention is implemented by adopting the following technical solutions:

[0006] In a first aspect, the present invention provides a method for regional lightning prediction, comprising:

[0007] Collect lightning strike point data and radar echo intensity data in the target area in the first time period;

[0008] Preprocessing the lightning strike point data and radar echo intensity data;

[0009] Based on the preprocessed lightning strike point data, the target area is gridded to obtain gridded lightning data of the target area;

[0010] Based on the gridded lightning data of the target area, the lightning threat probability score of each grid point in the target area is obtained, and the lightning threat probability score matrix of the target area is constructed;

[0011] The preprocessed radar echo intensity data and the lightning threat probability matrix of the target area are input into the pre-trained regional lightning forecast model to obtain the lightning prediction results of the target area in the second time period.

[0012] Furthermore, the lightning strike point data includes text data of the lightning strike point latitude and longitude, time, intensity and type.

[0013] Further, the lightning strike point data is preprocessed, including: accumulating the lightning strike point data according to a preset time period;

[0014] The radar echo intensity data is preprocessed, including: processing abnormal values ​​and interpolating missing values ​​in the radar echo intensity data; and converting the processed radar echo intensity data into grayscale image data through a function mapping method.

[0015] Furthermore, the lightning threat probability score of each grid point in the target area is obtained, including:

[0016] ,

[0017] ,

[0018] in, Representing grid points Lightning threat probability score; Indicates The location of the lightning event, Indicates the occurrence location and grid point The distance is less than or equal to The set of locations where lightning events occur; Indicates the lightning threat attenuation parameter, whose value is set according to the lightning impact distance; Indicates the occurrence location and grid point The distance is less than or equal to The total number of lightning events; according to the type of lightning, Representing grid points The cloud flash threat probability score, Representing grid points The probability score of ground-to-ground lightning threat; represents the cloud-flash threat attenuation parameter, Indicates the ground-to-ground lightning threat attenuation parameter; Indicates the occurrence location and grid point The distance is less than or equal to The location set of cloud flash events, Indicates the occurrence location and grid point The distance is less than or equal to The set of locations where lightning ground events occur.

[0019] Furthermore, the processed radar echo intensity data is converted into grayscale image data through a function mapping method, including:

[0020] ,

[0021] in, is the gray value of the pixel after conversion, is the radar echo intensity value, is the maximum value in the radar echo intensity data; It is the minimum value in the radar echo intensity data.

[0022] Furthermore, the preprocessed radar echo intensity data and the lightning threat probability score matrix of the target area are input into the pre-trained regional lightning forecast model to obtain the lightning forecast result of the target area in the second time period, including:

[0023] The preprocessed radar echo intensity data is input into a pre-trained first-region lightning prediction model to obtain a first feature matrix;

[0024] Inputting the lightning threat probability score matrix of the target area into a pre-trained second area lightning forecast model to obtain a second feature matrix;

[0025] Concatenate the first feature matrix and the second feature matrix to obtain feature fusion data;

[0026] Inputting the feature fusion data into a linear convolution layer, and outputting a lightning prediction result of the target area in the second time period;

[0027] The lightning prediction result of the target area in the second time period includes the lightning threat probability of each grid point in the target area in the second time period.

[0028] Furthermore, the first area lightning forecast model and the second area lightning forecast model include 3D-Unet machine learning models.

[0029] In a second aspect, the present invention further provides a regional lightning prediction device, the device comprising:

[0030] A data acquisition module, used to collect lightning strike point data and radar echo intensity data in the target area in the first time period;

[0031] A data preprocessing module, used for preprocessing the lightning strike point data and radar echo intensity data;

[0032] A grid lightning data acquisition module is used to grid the target area based on the preprocessed lightning point data to acquire the grid lightning data of the target area;

[0033] A lightning threat probability score matrix construction module is used to obtain the lightning threat probability score of each grid point in the target area based on the grid lightning data of the target area, and to construct the lightning threat probability score matrix of the target area;

[0034] The lightning prediction result acquisition module is used to input the preprocessed radar echo intensity data and the lightning threat probability score matrix of the target area into the pre-trained regional lightning prediction model to obtain the lightning prediction result of the target area in the second time period.

[0035] In a third aspect, the present invention further provides a computer device, comprising:

[0036] Memory for storing computer programs;

[0037] A processor is used to execute the computer program to implement the steps of the above-mentioned regional lightning prediction method.

[0038] In a fourth aspect, the present invention further provides a computer-readable storage medium having a computer program stored thereon, the program being executed by a processor to implement the steps of the above-mentioned regional lightning prediction method.

[0039] Compared with the prior art, the present invention has the following beneficial effects:

[0040] The regional lightning prediction method provided by the present invention converts lightning strike data into lightning threat probability scoring data through a mathematical model, which can comprehensively evaluate the actual threat of lightning to a specific area and has practical value in meteorological warning and disaster prevention and mitigation; by comprehensively analyzing factors such as the spatial distribution of lightning events, energy release, and weather conditions, the lightning threat probability in a specific area can be effectively calculated. By comprehensively analyzing lightning strike data and radar echo data, it is possible to dynamically monitor changes in severe convective weather, and by combining advanced machine learning models and using computer-automated real-time updates of lightning forecasts, the accuracy and timeliness of lightning forecasts are improved. The regional lightning forecast model can output the lightning threat probability prediction results for each grid point in the monitoring area, clarify the threat range of lightning, and facilitate users to intuitively understand the specific distribution of lightning risks, and then take targeted preventive and emergency measures. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 A schematic diagram of a flow chart of a regional lightning prediction method provided by an embodiment of the present invention;

[0042] Figure 2A schematic diagram of gridded lightning data of a target area in an embodiment of the present invention;

[0043] Figure 3 Schematic diagram of the principle of lightning threat probability score calculation in an embodiment of the present invention;

[0044] Figure 4 Schematic diagram of comparison between gridded lightning data of a target area and lightning threat probability score data of the target area in an embodiment of the present invention, wherein (a) is a schematic diagram of gridded lightning data of a target area, and (b) is a schematic diagram of lightning threat probability score data of a target area;

[0045] Figure 5 Schematic diagram of the structure of a regional lightning forecast model in an embodiment of the present invention;

[0046] Figure 6 It is a schematic diagram of lightning prediction results of a regional lightning prediction method at different classification thresholds and different prediction time periods when the hit rate POD is used as an evaluation indicator in an embodiment of the present invention;

[0047] Figure 7 It is a schematic diagram of lightning prediction results of a regional lightning prediction method at different classification thresholds and different prediction time periods when the false alarm rate FAR is used as an evaluation indicator in an embodiment of the present invention;

[0048] Figure 8 A schematic diagram of the structure of a regional lightning prediction device provided by an embodiment of the present invention;

[0049] Fig. 9 An internal structure diagram of a computer device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0050] The technical solution of the present invention is described in detail below through the accompanying drawings and specific embodiments. The same reference numerals in the accompanying drawings indicate the same or similar components or parts. It should be understood by those skilled in the art that these drawings are not necessarily drawn to scale. The embodiments of the present application and the specific features in the embodiments are detailed descriptions of the technical solution of the present application, rather than limitations on the technical solution of the present application. In the absence of conflict, the embodiments of the present application and the technical features in the embodiments can be combined with each other.

[0051] The term "and / or" in this article is only a description of the association relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0052] Embodiment 1:

[0053] like Figures 1 to 7 As shown, an embodiment of the present invention provides a method for regional lightning prediction. Figure 1 The flowchart is a flow chart of the regional lightning prediction method. This flowchart only shows the logical sequence of the method described in this embodiment. Under the premise of no conflict, in other possible embodiments of the present invention, different methods may be used. Figure 1 The steps shown or described are accomplished in the order shown.

[0054] The regional lightning prediction method provided in this embodiment can be applied to a terminal, and can be executed by a regional lightning prediction device, which can be implemented by software and / or hardware, and can be integrated in a terminal.

[0055] See also Figure 1 The method of the embodiment of the present invention specifically includes the following steps 1 to 4, wherein:

[0056] Step 1: Collect lightning strike point data and radar echo intensity data in the target area in the first time period.

[0057] Among them, the lightning landing point data can be obtained through a lightning detection station, and the data includes text data such as the longitude and latitude, time, intensity and type of the lightning landing point; the radar echo intensity data can be obtained through meteorological radar detection, and the data is in matrix form.

[0058] Meteorological radar is one of the most effective tools for monitoring severe convective weather systems. Weather systems with strong vertical convective movement and intense activity often appear as echo areas with high reflectivity on radar echo maps. Lightning is closely related to severe convective movement. The positions with higher echo intensity in radar echoes can correspond to locations with high probability of lightning occurrence, thereby obtaining the probability distribution of lightning disasters in severe convective weather in various locations in the region.

[0059] In order to realize lightning prediction, the first time period and the second time period of the present invention are usually a historical time period and a future time period based on the current time.

[0060] Step 2: Preprocess the lightning strike point data and radar echo intensity data.

[0061] Among them, the lightning strike point data is preprocessed, including: accumulating the lightning strike point data according to a preset time period, that is, accumulating lightning strike point data for a longer period of time in a single data frame, while the time difference between adjacent data frames is shorter, and the processed lightning strike point data can reflect the overall change trend of regional lightning.

[0062] The calculation formula for accumulating lightning strike point data by time period is:

[0063] ,

[0064] in, represents the total data set after processing; Indicates the number of time periods; Represents the cumulative results for a single time period; Represents raw lightning strike data.

[0065] The radar echo intensity data is preprocessed, including: using a linear interpolation method to process outliers and interpolate missing values ​​in the radar echo intensity data; and converting the processed radar echo intensity data into grayscale image data through a function mapping method to obtain radar grayscale image data.

[0066] Among them, the calculation formula for converting the processed radar echo intensity data into grayscale image value through the function mapping method is:

[0067] ,

[0068] in, is the gray value of the pixel after conversion, is the radar echo intensity value (in dBZ), is the maximum value in the radar echo intensity data; It is the minimum value in the radar echo intensity data.

[0069] Step 3: Based on the preprocessed lightning strike data, the target area is gridded to obtain the gridded lightning data of the target area.

[0070] Gridding refers to dividing the space into multiple small areas (grid points), each grid point represents a specific geographical location. Figure 2 As shown, based on the preprocessed lightning strike point data, the target area is divided into grids, and the grid points are assigned values ​​according to the energy intensity of the lightning detected at the grid point position, and normalized to obtain the gridded lightning data of the target area for subsequent lightning threat probability score calculation.

[0071] In the embodiment of the present invention, the area to be predicted is equivalent to the target area, therefore, gridding the target area is equivalent to gridding the area to be predicted. In other embodiments, in order to consider the comprehensiveness of the lightning threat probability score at the boundary area of ​​the area to be predicted, the target area can be expanded so that the area to be predicted is at the center of the target area. In this case, the target area and the area to be predicted are gridded respectively, and the gridded lightning data of the target area are used to perform lightning prediction in the area to be predicted.

[0072] Step 4: Based on the gridded lightning data of the target area, obtain the lightning threat probability score of each grid point in the target area and construct the lightning threat probability score matrix of the target area.

[0073] The calculation formula for the lightning threat probability score of each grid point in the target area is:

[0074] ,

[0075] ,

[0076] in, Representing grid points Lightning threat probability score; Indicates The location of the lightning event, Indicates the occurrence location and grid point The distance is less than or equal to The set of locations where lightning events occur; Indicates the lightning threat attenuation parameter, whose value is set according to the lightning impact distance; Indicates the occurrence location and grid point The distance is less than or equal to The total number of lightning events. According to the location of lightning, lightning is divided into cloud-to-ground lightning and ground-to-cloud lightning. Representing grid points The cloud flash threat probability score, Representing grid points The probability score of ground-to-ground lightning threat; represents the cloud-flash threat attenuation parameter, Indicates the ground-to-ground lightning threat attenuation parameter; Indicates the occurrence location and grid point The distance is less than or equal to The location set of cloud flash events, Indicates the occurrence location and grid point The distance is less than or equal to The set of locations where lightning ground events occur.

[0077] like Figure 3 As shown, the present invention designs a calculation formula for the lightning threat probability score of each grid point based on the idea that the closer the lightning distance is, the greater the threat level is. The threat level of each lightning event near this grid point to this grid point is calculated in turn, and the accumulated results are used as the lightning threat probability score of this grid point.

[0078] According to the lightning threat probability score of each grid point in the target area, a lightning threat probability score matrix of the target area is constructed.

[0079] Figure 4A schematic diagram comparing the gridded lightning data of the target area and the lightning threat probability score data of the target area is shown. It can be seen that a few scattered lightnings correspond to low threat probability values, and lightning-dense areas correspond to high threat probability values. It can be seen that the lightning threat probability score calculation method designed by the present invention is reasonable and effective.

[0080] Step 5: Input the preprocessed radar echo intensity data and the lightning threat probability score matrix of the target area into the pre-trained regional lightning forecast model to obtain the lightning prediction results of the target area in the second time period.

[0081] Step 5 specifically includes:

[0082] The preprocessed radar echo intensity data, namely, the radar grayscale image data, is input into a pre-trained first-region lightning prediction model to obtain a first feature matrix;

[0083] Inputting lightning threat probability score matrix data of the target area into the pre-trained second area lightning forecast model to obtain a second feature matrix;

[0084] The first feature matrix and the second feature matrix are concatenated to obtain feature fusion data;

[0085] The feature fusion data is input into the linear convolution layer, and the lightning prediction result of the target area in the second time period is output;

[0086] The lightning prediction result of the target area in the second time period includes the lightning threat probability of each grid point in the target area in the second time period.

[0087] In an embodiment of the present invention, the first area lightning forecast model and the second area lightning forecast model both use the 3D-Unet machine learning model. The network structure of 3D-Unet is extended on the basis of traditional U-Net to process three-dimensional data, and mainly includes an encoder, a bottleneck layer, a decoder and an output layer. The encoder part is composed of multiple 3D convolution blocks and a maximum pooling layer, which is used to extract spatial features and gradually reduce the spatial size of the feature map; the bottleneck layer is located between the encoder and the decoder, and further extracts features through deep 3D convolution; the decoder part uses 3D transposed convolution for upsampling, and splices the feature map in the encoder with the feature map in the decoder through jump connections to restore detail information; the output layer maps the feature map of the decoder to the target space to generate a prediction result.

[0088] The 3D-Unet structure can effectively capture the spatial and temporal features in three-dimensional data. Figure 5As shown, the first area lightning forecast model and the second area lightning forecast model of the embodiment of the present invention use the network structure model after removing the last two convolutional layers of the 3D-Unet model, which is named as the 3D-Unet Core module. The radar grayscale image and the lightning threat probability score matrix are input into the 3D-Unet Core module, and the output is a plurality of feature maps of the intermediate stage. The feature maps of the two channels are spliced ​​together in the sequence dimension, and the splicing method is similar to the feature jump splicing method in Unet.

[0089] A 1×1×1 3D convolution kernel is used to linearly combine the feature fusion data to achieve information interaction, and a second feature extraction is performed. Finally, the Relu linear activation function is used to output the feature map to achieve dimensionality reduction and obtain the final two-dimensional predicted image.

[0090] In an embodiment of the present invention, the first area lightning forecast model and the second area lightning forecast model are pre-trained, verified and evaluated through the historical lightning threat probability score data and historical radar echo image data of the target area in the past two years. The historical lightning threat probability score data and the historical radar echo image data are divided into a training set and a test set in a ratio of 4:1. It is found that the proportion of lightning data in the collected historical data is relatively small, with lightning events accounting for about 4% and non-lightning events accounting for about 96%. In order to solve the problem of extremely unbalanced data sets, data in similar time periods when lightning events occur are selected during data sampling to increase the proportion of lightning events in the overall data.

[0091] like Figure 6 and Figure 7 As shown, the hit rate POD and the false alarm rate FAR are used as evaluation indicators to verify the effect of the regional lightning prediction method of the present invention.

[0092] The second time period is set to 30 minutes, 1 hour, 1 hour 30 minutes and 2 hours in the future based on the current time to perform lightning prediction. It can be seen from the lightning prediction result evaluation diagram that as the prediction time becomes longer, the method performance does not show a significant decrease. In addition, lowering the classification threshold can improve the prediction hit rate, and increasing the classification threshold can reduce the false alarm rate. The present invention can obtain corresponding model performance according to different requirements by adjusting the threshold.

[0093] Embodiment 2:

[0094] Based on the same inventive concept as in Example 1, the embodiment of the present invention further provides a regional lightning prediction device for implementing the above regional lightning prediction method. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in the embodiment of the regional lightning prediction device provided below can refer to the limitations of the regional lightning prediction method above, and will not be repeated here.

[0095] like Figure 8 As shown, an embodiment of the present invention provides a regional lightning prediction device, comprising:

[0096] A data acquisition module, used to collect lightning strike point data and radar echo intensity data in the target area in the first time period;

[0097] A data preprocessing module, used for preprocessing the lightning strike point data and radar echo intensity data;

[0098] A grid lightning data acquisition module is used to grid the target area based on the preprocessed lightning point data to acquire the grid lightning data of the target area;

[0099] A lightning threat probability score matrix construction module is used to obtain the lightning threat probability score of each grid point in the target area based on the grid lightning data of the target area, and to construct the lightning threat probability score matrix of the target area;

[0100] The lightning prediction result acquisition module is used to input the preprocessed radar echo intensity data and the lightning threat probability score matrix of the target area into the pre-trained regional lightning prediction model to obtain the lightning prediction result of the target area in the second time period.

[0101] Embodiment 3:

[0102] The embodiment of the present invention further provides a computer device, which may be a server, and its internal structure diagram may be as shown in FIG. Fig. 9 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the regional lightning prediction method in the aforementioned embodiment is implemented.

[0103] Those skilled in the art will understand that Fig. 9The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0104] Embodiment 4:

[0105] The embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the steps of the following method are implemented:

[0106] Collect lightning strike point data and radar echo intensity data in the target area in the first time period;

[0107] Preprocessing the lightning strike point data and radar echo intensity data;

[0108] Based on the preprocessed lightning strike point data, the target area is gridded to obtain gridded lightning data of the target area;

[0109] Based on the gridded lightning data of the target area, the lightning threat probability score of each grid point in the target area is obtained, and the lightning threat probability score matrix of the target area is constructed;

[0110] The preprocessed radar echo intensity data and the lightning threat probability matrix of the target area are input into the pre-trained regional lightning forecast model to obtain the lightning prediction results of the target area in the second time period.

[0111] It should be understood by those skilled in the art that the embodiments of the present invention may be provided as methods, systems, or computer program products, and therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0112] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0113] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0114] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0115] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the enlightenment of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the purpose of the present invention and the claims, which all fall within the protection of the present invention.

Claims

1. A regional lightning prediction method, characterized in that: include: Collect lightning strike point data and radar echo intensity data in the target area in the first time period; Preprocessing the lightning strike point data and radar echo intensity data; Based on the preprocessed lightning strike point data, the target area is gridded to obtain gridded lightning data of the target area; Based on the gridded lightning data of the target area, the lightning threat probability score of each grid point in the target area is obtained, and the lightning threat probability score matrix of the target area is constructed; The preprocessed radar echo intensity data and the lightning threat probability score matrix of the target area are input into the pre-trained regional lightning forecast model to obtain the lightning prediction results of the target area in the second time period.

2. The regional lightning prediction method according to claim 1, characterized in that: The lightning strike point data includes text data of the lightning strike point latitude and longitude, time, intensity and type.

3. The regional lightning prediction method according to claim 2, characterized in that: Preprocessing the lightning strike point data includes: accumulating the lightning strike point data according to a preset time period; The radar echo intensity data is preprocessed, including: processing abnormal values ​​and interpolating missing values ​​in the radar echo intensity data; and converting the processed radar echo intensity data into grayscale image data through a function mapping method.

4. The regional lightning prediction method according to claim 3, characterized in that: Get the lightning threat probability score for each grid point in the target area, including: ， ， in, Representing grid points Lightning threat probability score; Indicates The location of the lightning event, Indicates the occurrence location and grid point The distance is less than or equal to The set of locations where lightning events occur; Indicates the lightning threat attenuation parameter, whose value is set according to the lightning impact distance; Indicates the occurrence location and grid point The distance is less than or equal to The total number of lightning events; according to the type of lightning, Representing grid points The cloud flash threat probability score, Representing grid points The probability score of ground-to-ground lightning threat; represents the cloud-flash threat attenuation parameter, Indicates the ground-to-ground lightning threat attenuation parameter; Indicates the occurrence location and grid point The distance is less than or equal to The location set of cloud flash events, Indicates the occurrence location and grid point The distance is less than or equal to The set of locations where lightning ground events occur.

5. The regional lightning prediction method according to claim 3, characterized in that: The processed radar echo intensity data is converted into grayscale image data through function mapping method, including: ， in, is the gray value of the pixel after conversion, is the radar echo intensity value, is the maximum value in the radar echo intensity data; It is the minimum value in the radar echo intensity data.

6. The regional lightning prediction method according to claim 5, characterized in that: The preprocessed radar echo intensity data and the lightning threat probability score matrix of the target area are input into the pre-trained regional lightning forecast model to obtain the lightning forecast results of the target area in the second time period, including: The preprocessed radar echo intensity data is input into a pre-trained first-region lightning prediction model to obtain a first feature matrix; Inputting the lightning threat probability score matrix of the target area into a pre-trained second area lightning forecast model to obtain a second feature matrix; Concatenate the first feature matrix and the second feature matrix to obtain feature fusion data; Inputting the feature fusion data into a linear convolution layer, and outputting a lightning prediction result of the target area in the second time period; The lightning prediction result of the target area in the second time period includes the lightning threat probability of each grid point in the target area in the second time period.

7. The regional lightning prediction method according to claim 6, characterized in that: The first area lightning forecast model and the second area lightning forecast model include 3D-Unet machine learning models.

8. A regional lightning prediction device, characterized in that: include: A data acquisition module, used to collect lightning strike point data and radar echo intensity data in the target area in the first time period; A data preprocessing module, used for preprocessing the lightning strike point data and radar echo intensity data; A grid lightning data acquisition module is used to grid the target area based on the preprocessed lightning point data to acquire the grid lightning data of the target area; A lightning threat probability score matrix construction module is used to obtain the lightning threat probability score of each grid point in the target area based on the grid lightning data of the target area, and to construct the lightning threat probability score matrix of the target area; The lightning prediction result acquisition module is used to input the preprocessed radar echo intensity data and the lightning threat probability score matrix of the target area into the pre-trained regional lightning prediction model to obtain the lightning prediction result of the target area in the second time period.

9. A computer device, characterized in that: include: Memory for storing computer programs; A processor, configured to execute the computer program to implement the steps of the regional lightning prediction method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the regional lightning prediction method according to any one of claims 1 to 7 are implemented.

Images