Dynamic weight-based rainstorm forecast fusion method and device, electronic equipment and storage medium

Abstract

The invention belongs to the technical field of weather forecast, and particularly relates to a rainstorm forecast fusion method and device based on dynamic weight, electronic equipment and a storage medium. The method comprises the following steps: acquiring radar data and numerical forecasting data for rainstorm forecasting; obtaining a dynamic weight; and performing data fusion based on the dynamic weight, the radar data and the numerical forecasting data to obtain fused rainstorm forecasting data based on the dynamic weight. According to the method, different characteristics of the radar data and the numerical forecasting data are fully considered, the advantages of the numerical forecasting data and the radar observation data can be fully utilized by dynamically adjusting the weight, and the precision of rainstorm forecasting is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of meteorological forecasting, and in particular relates to a dynamic weight-based rainstorm forecast fusion method, device, electronic equipment and storage medium. Background Art

[0002] As an extreme weather phenomenon, rainstorms have a huge impact on human society and the natural environment. Accurate rainstorm forecasts are of great significance for disaster prevention and mitigation.

[0003] Currently, heavy rain forecasts rely primarily on numerical prediction models and radar observation data. Numerical prediction models can provide forecasts over longer timescales, but suffer from insufficient accuracy. Radar observation data can provide precipitation information with high temporal and spatial resolution, but the forecast timeframe is relatively short. Existing heavy rain forecast fusion methods mostly employ fixed weights or simple averaging, which fail to fully account for the variability between models and data under different temporal and spatial conditions. For example, during the initial onset of a heavy rainstorm, some models may be more sensitive to the triggering mechanism, while during the ongoing phase of a heavy rainstorm, other models may more accurately reflect the intensity and duration of precipitation. Furthermore, meteorological conditions and topography vary across regions, affecting the applicability of models. Therefore, fixed-weight fusion methods struggle to adapt to complex heavy rain forecast scenarios, limiting the accuracy and reliability of the fusion results. Summary of the Invention

[0004] The purpose of the present invention is to provide a dynamic weight-based rainstorm forecast fusion method, device, electronic device and storage medium, which can dynamically adjust the weight according to the accuracy of historical data, make full use of the advantages of numerical forecast data and radar observation data, and improve the accuracy and timeliness of rainstorm forecasts.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions: In a first aspect, the present invention provides a rainstorm forecast fusion method based on dynamic weights, comprising: Obtain radar data and numerical forecast data for heavy rain forecasting; Get dynamic weight; Data fusion is performed based on dynamic weights, radar data and numerical forecast data to obtain fused heavy rain forecast data based on dynamic weights.

[0006] A further improvement of the present invention is that before performing data fusion based on the dynamic weight, radar data and numerical forecast data, the method further includes: Preprocessing the radar data and the numerical forecast data to obtain preprocessed radar data and numerical forecast data; Accordingly, the data fusion based on dynamic weights, radar data and numerical forecast data includes: Data fusion is performed based on dynamic weights, preprocessed radar data and numerical forecast data.

[0007] A further improvement of the present invention is that the step of preprocessing the radar data and the numerical forecast data to obtain the preprocessed radar data and the numerical forecast data specifically includes: time alignment, spatial resolution unification, and format unification of the radar data and the numerical forecast data.

[0008] A further improvement of the present invention is that in the step of obtaining the dynamic weight, the dynamic weight is calculated by the following steps: The forecast period of 0 to N hours is divided into several time windows; N is a positive number greater than or equal to 2; For each time window, calculate the historical root mean square error between the radar data and the numerical forecast data in the corresponding time window over the past M days; M is a positive integer greater than or equal to 3; For each time window, dynamic weights are calculated based on the historical root mean square error between radar data and numerical forecast data; The weight of each time window is normalized.

[0009] A further improvement of the present invention is that: in the step of dividing the forecast period of 0 to N hours into several time windows, N is equal to 6; specifically, the step includes: dividing the forecast period of 0 to 6 hours into 4 time windows of 0 to 1 hour, 1 to 2 hours, 2 to 4 hours and 4 to 6 hours.

[0010] A further improvement of the present invention is that, in the step of calculating the dynamic weight according to the historical root mean square error of the radar data and the numerical forecast data for each time window, the weight is calculated using the following formula: w _r = 1 / RMSE _r w _n = 1 / RMSE _n Where w_r and w_n represent the weights of radar data and numerical forecast data, respectively; RMSE_r and RMSE_n represent the historical root mean square errors of radar data and numerical forecast data, respectively.

[0011] A further improvement of the present invention is that in the step of normalizing the weight of each time window, the normalization is performed using the following formula: w _r' = w _r / (w _r + w _n ) w _n' = w_n / (w _r + w _n ) Where w_r' and w_n' represent the weights of the normalized radar data and numerical forecast data, respectively.

[0012] A further improvement of the present invention is that, in the step of performing data fusion based on dynamic weights, pre-processed radar data and numerical forecast data to obtain fused heavy rain forecast data based on dynamic weights, for the forecast period of 0 to 2 hours, the following formula is used for fusion: P_f = w_r'×P_r + w_n'×P_n Among them, P_f represents the fused rainstorm forecast data, P_r and P_n represent radar data and numerical forecast data respectively; For the forecast period of 2 to 6 hours, the following formula is used for fusion: P_f(t) = w_r'(t) ×P_r(t) + w_n'(t)×P_n(t) Among them, P_f(t) represents the fused heavy rain forecast data at time t, w_r'(t) and w_n'(t) represent the normalized weights of the radar data and numerical forecast data at time t, respectively, and P_r(t) and P_n(t) represent the radar data and numerical forecast data at time t, respectively.

[0013] In a second aspect, the present invention provides a dynamic weight-based rainstorm forecast fusion device, comprising: An acquisition module is used to obtain radar data and numerical forecast data for heavy rain forecast; Dynamic weight module, used to obtain dynamic weight; The fusion prediction module is used to perform data fusion based on dynamic weights, radar data and numerical forecast data to obtain fused heavy rain forecast data based on dynamic weights.

[0014] In a third aspect, the present invention provides an electronic device comprising a processor and a memory, wherein the processor is configured to execute a computer program stored in the memory to implement the dynamic weight-based rainstorm forecast fusion method.

[0015] In a fourth aspect, the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores at least one instruction, and when the at least one instruction is executed by a processor, the dynamic weight-based rainstorm forecast fusion method is implemented.

[0016] Compared with the prior art, the present invention has the following beneficial effects: The present invention provides a dynamic weight-based rainstorm forecast fusion method, comprising: obtaining radar data and numerical forecast data for rainstorm forecasting; preprocessing the radar data and numerical forecast data to obtain preprocessed radar data and numerical forecast data; obtaining dynamic weights; and fusing the data based on the dynamic weights, the preprocessed radar data, and the numerical forecast data to obtain fused rainstorm forecast data based on the dynamic weights. The present invention fully considers the different characteristics of radar data and numerical forecast data and, by dynamically adjusting weights, can fully utilize the advantages of both numerical forecast data and radar observation data, thereby improving the accuracy of rainstorm forecasts.

[0017] Furthermore, the present invention combines the short-term forecasting advantages of radar data with the long-term forecasting advantages of numerical forecast data to extend the timeliness of heavy rain forecasts.

[0018] Furthermore, the present invention can adjust according to the accuracy of historical data through dynamic weights to adapt to forecasting needs under different weather conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The accompanying drawings, which constitute part of the present invention, are intended to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are intended to explain the present invention and do not constitute an undue limitation of the present invention. In the accompanying drawings: Figure 1 This is a flow chart of a dynamic weight-based rainstorm forecast fusion method according to an embodiment of the present invention; Figure 2 This is a structural block diagram of a dynamic weight-based rainstorm forecast fusion device according to an embodiment of the present invention; Figure 3 This is a structural block diagram of an electronic device of the present invention. DETAILED DESCRIPTION

[0020] The present invention will be described in detail below with reference to the accompanying drawings and in combination with embodiments. It should be noted that, in the absence of conflict, the embodiments and features of the embodiments of the present invention can be combined with each other.

[0021] The following detailed description is an exemplary description and is intended to provide further detailed description of the present invention. Unless otherwise indicated, all technical terms used in the present invention have the same meaning as those generally understood by those skilled in the art to which the present invention belongs. The terms used in the present invention are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present invention.

[0022] An embodiment of the present invention provides a dynamic weight-based rainstorm forecast fusion method, comprising the following steps: S1. Data preprocessing S11. Time alignment: unify the temporal resolution of radar data and numerical forecast data; In a specific embodiment, the time resolution of radar data and numerical forecast data is unified into minute intervals; for example, the time interval is 1 minute, 5 minutes, 10 minutes or 30 minutes; In a specific embodiment, the time resolutions of radar data and numerical forecast data are unified into hourly intervals; for example, the time interval is 1 hour or 2 hours.

[0023] S12. Spatial interpolation: Unify the spatial resolution of radar data and numerical forecast data; In a specific embodiment, the radar data and the numerical forecast data are interpolated to the same grid points.

[0024] S13. Data format conversion: Convert radar data and numerical forecast data into a unified format; In a specific implementation, the radar data and the numerical forecast data are uniformly converted into NetCDF or HDF5.

[0025] S2. Dynamic weight calculation S21. Divide the time window: Divide the 0-6 hour forecast period into multiple time windows; In a specific implementation, the forecast period of 0 to 6 hours is divided into four time windows: 0 to 1 hour, 1 to 2 hours, 2 to 4 hours, and 4 to 6 hours.

[0026] S22. Calculate historical error: For each time window, calculate the historical root mean square error (RMSE) between the radar data and the numerical forecast data within the past three days.

[0027] S23. Calculate dynamic weights: For each time window, calculate dynamic weights based on the historical root mean square error between radar data and numerical forecast data. The smaller the error, the greater the weight.

[0028] In one embodiment, the weight is calculated using the following formula: w _r = 1 / RMSE _r w _n = 1 / RMSE _n Where w_r and w_n represent the weights of radar data and numerical forecast data, respectively; RMSE_r and RMSE_n represent the historical root mean square errors of radar data and numerical forecast data, respectively.

[0029] S24. Weight normalization: Normalize the weights of each time window to ensure that the sum of the weights is 1.

[0030] In one embodiment, the normalization is performed using the following formula: w _r' = w _r / (w _r + w _n ) w _n' = w _n / (w _r + w _n ) Where w_r' and w_n' represent the weights of the normalized radar data and numerical forecast data, respectively.

[0031] S3, Data Fusion S31. For the forecast period of 0 to 2 hours: For the forecast period of 0 to 2 hours, radar data is mainly relied upon because radar data has higher accuracy within this period.

[0032] In one embodiment, the following formula is used for fusion for 0-2 hours: P_f = w_r'×P_r + w_n'×P_n Among them, P_f represents the fused rainstorm forecast data, P_r and P_n represent radar data and numerical forecast data respectively.

[0033] S32. For the 2-6 hour forecast period: As the forecast period increases, the accuracy of radar data decreases, while the accuracy of numerical forecast data increases. Therefore, dynamic weighting is used for fusion, and the weights change over time.

[0034] In one embodiment, the following formula is used for fusion for 2 to 6 hours: P_f(t) = w_r'(t) ×P_r(t) + w_n'(t)×P_n(t) Among them, P_f(t) represents the fused heavy rain forecast data at time t, w_r'(t) and w_n'(t) represent the normalized weights of the radar data and numerical forecast data at time t, respectively, and P_r(t) and P_n(t) represent the radar data and numerical forecast data at time t, respectively.

[0035] S4. Result output The fused forecast data for 0 to 6 hours is output; the forecast data includes information such as precipitation intensity and precipitation range.

[0036] See also Figure 1 As shown, an embodiment of the present invention provides a rainstorm forecast fusion method based on dynamic weights, including: S100, acquiring radar data and numerical forecast data for heavy rain forecast; S200, obtaining dynamic weight; S300: performing data fusion based on dynamic weights, pre-processed radar data, and numerical forecast data to obtain fused heavy rain forecast data based on dynamic weights.

[0037] In a specific embodiment, the present invention provides a method for fusion of rainstorm forecasts based on dynamic weights, which, before performing data fusion based on dynamic weights, radar data, and numerical forecast data, further includes: Preprocessing the radar data and the numerical forecast data to obtain preprocessed radar data and numerical forecast data; Accordingly, the data fusion based on dynamic weights, radar data and numerical forecast data includes: Data fusion is performed based on dynamic weights, preprocessed radar data and numerical forecast data.

[0038] In a specific embodiment, the step of preprocessing the radar data and the numerical forecast data to obtain the preprocessed radar data and the numerical forecast data specifically includes: time alignment, spatial resolution unification, and format unification of the radar data and the numerical forecast data.

[0039] In a specific embodiment, in the step of obtaining the dynamic weight, the dynamic weight is calculated by the following steps: The forecast period of 0 to N hours is divided into several time windows; N is a positive number greater than or equal to 2; For each time window, calculate the historical root mean square error between the radar data and the numerical forecast data in the corresponding time window over the past M days; M is a positive integer greater than or equal to 3; For each time window, dynamic weights are calculated based on the historical root mean square error between radar data and numerical forecast data; The weight of each time window is normalized.

[0040] In a specific embodiment, in the step of dividing the forecast period of 0 to N hours into several time windows, N is equal to 6; specifically, it includes: dividing the forecast period of 0 to 6 hours into 4 time windows of 0 to 1 hour, 1 to 2 hours, 2 to 4 hours and 4 to 6 hours.

[0041] In a specific embodiment, in the step of calculating the dynamic weight according to the historical root mean square error of the radar data and the numerical forecast data for each time window, the weight is calculated using the following formula: w _r = 1 / RMSE _r w_n = 1 / RMSE _n Where w_r and w_n represent the weights of radar data and numerical forecast data, respectively; RMSE_r and RMSE_n represent the historical root mean square errors of radar data and numerical forecast data, respectively.

[0042] In a specific embodiment, in the step of normalizing the weight of each time window, the following formula is used for normalization: w _r' = w _r / (w _r + w _n ) w _n' = w _n / (w _r + w _n ) Where w_r' and w_n' represent the weights of the normalized radar data and numerical forecast data, respectively.

[0043] In a specific embodiment, in the step of performing data fusion based on dynamic weights, pre-processed radar data, and numerical forecast data to obtain fused heavy rain forecast data based on dynamic weights, for a forecast period of 0 to 2 hours, the following formula is used for fusion: P_f = w_r'×P_r + w_n'×P_n Among them, P_f represents the fused rainstorm forecast data, P_r and P_n represent radar data and numerical forecast data respectively; For the forecast period of 2 to 6 hours, the following formula is used for fusion: P_f(t) = w_r'(t) ×P_r(t) + w_n'(t)×P_n(t) Among them, P_f(t) represents the fused heavy rain forecast data at time t, w_r'(t) and w_n'(t) represent the normalized weights of the radar data and numerical forecast data at time t, respectively, and P_r(t) and P_n(t) represent the radar data and numerical forecast data at time t, respectively.

[0044] A meteorological bureau is responsible for regional rainstorm forecasting and currently uses numerical forecasting models and radar observation systems. Numerical forecasting can provide precipitation forecasts for the next six hours, but its accuracy is insufficient. Radar can provide high-precision precipitation forecasts for the next two hours, but its timeliness is limited. To improve the accuracy and timeliness of rainstorm forecasts, a dynamic weighted rainstorm forecast fusion method based on the present invention is employed. The specific steps include: Data Preparation 1. Numerical forecast data: The precipitation forecast data for the next 6 hours is obtained from the numerical forecast system of the Meteorological Bureau, with a spatial resolution of 1 km × 1 km and a temporal resolution of 1 hour.

[0045] 2. Radar observation data: Precipitation forecast data for the next two hours is obtained from the radar system with a spatial resolution of 1 km × 1 km and a temporal resolution of 10 minutes.

[0046] 3. Observation data: The precipitation observation data for the past three days are obtained from the ground weather station and used to calculate the RMSE.

[0047] Data preprocessing 1. Time alignment: The time resolution of radar data is interpolated from 10 minutes to 1 hour to unify with the numerical forecast data.

[0048] 2. Spatial interpolation: The spatial resolution of radar data and numerical forecast data is uniformly interpolated to the same grid points.

[0049] 3. Data format conversion: Convert radar data and numerical forecast data into NetCDF format for subsequent processing.

[0050] Dynamic weight calculation 1. Divide the time window: Divide the forecast period of 0 to 6 hours into four time windows: 0 to 1 hour, 1 to 2 hours, 2 to 4 hours, and 4 to 6 hours.

[0051] 2. Calculate historical error: For each time window, calculate the RMSE of the radar data and numerical forecast data over the past three days. For the 0-1 hour time window, the RMSE for the radar data is 2.5 mm, and the RMSE for the numerical forecast data is 3.0 mm.

[0052] 3. Calculate dynamic weights: Calculate dynamic weights based on RMSE. For the 0-1 hour time window, the radar data weight is w_r = 1 / 2.5 = 0.4, and the numerical forecast data weight is w_n = 1 / 3.0 ≈ 0.33.

[0053] 4. Weight normalization: Normalize the weights so that their sum is 1. For the 0-1 hour time window, the normalized radar data weight w_r' = 0.4 / (0.4 + 0.33) ≈ 0.55, and the numerical forecast data weight w_n' = 0.33 / (0.4 + 0.33) ≈ 0.45.

[0054] Data fusion 1. 0-2 hours: Mainly relies on radar data, fused using the formula. Within the 0-2 hour time window, the fused forecast data P_f = 0.55 × P_r + 0.45 × P_n.

[0055] 2. 2-6 hours: As the forecast time increases, the accuracy of radar data gradually decreases, while the accuracy of numerical forecast data gradually increases. Dynamic weights are used for fusion, and the weights change over time. For 2-6 hours, the following formula is used for fusion: P_f(t) = w_r'(t) × P_r(t) + w_n'(t) × P_n(t).

[0056] Result Output The fused forecast data for hours 0-6 is output, including information on precipitation intensity and range. The output forecast indicates that heavy rain will occur in the region within the next six hours, with an intensity between 20 and 50 mm / h and a range covering 50% of the area.

[0057] Compared with the existing fixed weight technology, this embodiment has the following significant improvements: 1. Improved forecast accuracy: By comparing with actual observation data, the fused rainstorm forecast accuracy has increased by 2%.

[0058] 2. Extended forecast time: The short-term forecast advantages of radar data are successfully combined with the long-term forecast advantages of numerical forecast data, extending the rainstorm forecast time from the original 2 hours to 6 hours.

[0059] 3. Strong adaptability: Dynamic weights can be adjusted based on the accuracy of historical data to meet forecast requirements under different weather conditions.

[0060] See also Figure 2 As shown, an embodiment of the present invention provides a rainstorm forecast fusion device based on dynamic weights, comprising: An acquisition module is used to obtain radar data and numerical forecast data for heavy rain forecast; Dynamic weight module, used to obtain dynamic weight; The fusion prediction module is used to perform data fusion based on dynamic weights, pre-processed radar data and numerical forecast data to obtain fused heavy rain forecast data based on dynamic weights.

[0061] In a specific embodiment, an embodiment of the present invention provides a rainstorm forecast fusion device based on dynamic weights, further comprising: A preprocessing module is used to preprocess the radar data and the numerical forecast data to obtain the preprocessed radar data and the numerical forecast data; The fusion prediction module is specifically configured to perform data fusion based on dynamic weights, pre-processed radar data and numerical forecast data.

[0062] In a specific embodiment, the step of preprocessing the radar data and the numerical forecast data to obtain the preprocessed radar data and the numerical forecast data specifically includes: time alignment, spatial resolution unification, and format unification of the radar data and the numerical forecast data.

[0063] In a specific embodiment, in the step of obtaining the dynamic weight, the dynamic weight is calculated by the following steps: The forecast period of 0 to N hours is divided into several time windows; N is a positive number greater than or equal to 2; For each time window, calculate the historical root mean square error between the radar data and the numerical forecast data in the corresponding time window over the past M days; M is a positive integer greater than or equal to 3; For each time window, dynamic weights are calculated based on the historical root mean square error between radar data and numerical forecast data; The weight of each time window is normalized.

[0064] In a specific embodiment, in the step of dividing the forecast period of 0 to N hours into several time windows, N is equal to 6; specifically, it includes: dividing the forecast period of 0 to 6 hours into 4 time windows of 0 to 1 hour, 1 to 2 hours, 2 to 4 hours and 4 to 6 hours.

[0065] In a specific embodiment, in the step of calculating the dynamic weight according to the historical root mean square error of the radar data and the numerical forecast data for each time window, the weight is calculated using the following formula: w _r = 1 / RMSE _r w _n = 1 / RMSE _n Where w_r and w_n represent the weights of radar data and numerical forecast data, respectively; RMSE_r and RMSE_n represent the historical root mean square errors of radar data and numerical forecast data, respectively.

[0066] In a specific embodiment, in the step of normalizing the weight of each time window, the following formula is used for normalization: w _r' = w _r / (w _r + w _n ) w _n' = w _n / (w _r + w _n ) Where w_r' and w_n' represent the weights of the normalized radar data and numerical forecast data, respectively.

[0067] In a specific embodiment, in the step of performing data fusion based on dynamic weights, pre-processed radar data, and numerical forecast data to obtain fused heavy rain forecast data based on dynamic weights, for a forecast period of 0 to 2 hours, the following formula is used for fusion: P_f = w_r'×P_r + w_n'×P_n Among them, P_f represents the fused rainstorm forecast data, P_r and P_n represent radar data and numerical forecast data respectively; For the forecast period of 2 to 6 hours, the following formula is used for fusion: P_f(t) = w_r'(t) ×P_r(t) + w_n'(t)×P_n(t) Among them, P_f(t) represents the fused heavy rain forecast data at time t, w_r'(t) and w_n'(t) represent the normalized weights of the radar data and numerical forecast data at time t, respectively, and P_r(t) and P_n(t) represent the radar data and numerical forecast data at time t, respectively.

[0068] See also Figure 3 As shown, an embodiment of the present invention provides an electronic device 100 for implementing a dynamic weight-based rainstorm forecast fusion method; the electronic device 100 includes a memory 101, at least one processor 102, a computer program 103 stored in the memory 101 and executable on the at least one processor 102, and at least one communication bus 104.

[0069] The memory 101 can be used to store the computer program 103. The processor 102 implements the steps of the dynamic weight-based rainstorm forecast fusion method described in the embodiment by running or executing the computer program stored in the memory 101 and calling the data stored in the memory 101. The memory 101 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.); the data storage area can store data (such as audio data) created according to the use of the electronic device 100. In addition, the memory 101 can include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other non-volatile solid-state storage devices.

[0070] The at least one processor 102 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The processor 102 may be a microprocessor or any conventional processor, etc. The processor 102 is the control center of the electronic device 100 and connects various parts of the entire electronic device 100 using various interfaces and lines.

[0071] The memory 101 in the electronic device 100 stores a plurality of instructions to implement a heavy rain forecast fusion method based on dynamic weights, and the processor 102 can execute the plurality of instructions to implement: Obtain radar data and numerical forecast data for heavy rain forecasting; Preprocessing the radar data and the numerical forecast data to obtain preprocessed radar data and numerical forecast data; Get dynamic weight; Data fusion is performed based on dynamic weights, preprocessed radar data and numerical forecast data to obtain fused heavy rain forecast data based on dynamic weights.

[0072] If the module / unit integrated in the electronic device 100 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 storage medium. Based on this understanding, the present invention implements all or part of the process in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned method embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device that can carry the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory and read-only memory (ROM, Read-Only Memory).

[0073] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, 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 magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0074] 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 flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, 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 flowcharts and / or block diagrams. 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.

[0075] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work 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 The function specified in one or more boxes.

[0076] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0077] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the scope of protection of the claims of the present invention.

Claims

1. A rainstorm forecast fusion method based on dynamic weights, characterized in that: include: Obtain radar data and numerical forecast data for heavy rain forecasting; Get dynamic weight; Data fusion is performed based on dynamic weights, radar data and numerical forecast data to obtain fused heavy rain forecast data based on dynamic weights.

2. The method for fusion of rainstorm forecast based on dynamic weight according to claim 1, characterized in that: Before performing data fusion based on the dynamic weights, radar data, and numerical forecast data, the method further includes: Preprocessing the radar data and the numerical forecast data to obtain preprocessed radar data and numerical forecast data; Accordingly, the data fusion based on dynamic weights, radar data and numerical forecast data includes: Data fusion is performed based on dynamic weights, preprocessed radar data and numerical forecast data.

3. The method for fusion of rainstorm forecast based on dynamic weight according to claim 2, characterized in that: The step of preprocessing the radar data and the numerical forecast data to obtain the preprocessed radar data and the numerical forecast data specifically includes: time alignment, spatial resolution unification and format unification of the radar data and the numerical forecast data.

4. The method for fusion of rainstorm forecast based on dynamic weight according to claim 1, characterized in that: In the step of obtaining the dynamic weight, the dynamic weight is calculated by the following steps: The forecast period of 0 to N hours is divided into several time windows; N is a positive number greater than or equal to 2; For each time window, calculate the historical root mean square error between the radar data and the numerical forecast data in the corresponding time window over the past M days; M is a positive integer greater than or equal to 3; For each time window, dynamic weights are calculated based on the historical root mean square error between radar data and numerical forecast data; The weight of each time window is normalized.

5. The method for fusion of rainstorm forecast based on dynamic weight according to claim 4, characterized in that: In the step of dividing the forecast period of 0 to N hours into several time windows, N is equal to 6; specifically, the step includes dividing the forecast period of 0 to 6 hours into 4 time windows of 0 to 1 hour, 1 to 2 hours, 2 to 4 hours and 4 to 6 hours.

6. A dynamic weight-based rainstorm forecast fusion method according to claim 4, characterized in that: In the step of calculating the dynamic weight according to the historical root mean square error of radar data and numerical forecast data for each time window, the weight is calculated using the following formula: w _r = 1 / RMSE _r w _n = 1 / RMSE _n Where w_r and w_n represent the weights of radar data and numerical forecast data, respectively; RMSE_r and RMSE_n represent the historical root mean square errors of radar data and numerical forecast data, respectively.

7. The method for fusion of rainstorm forecast based on dynamic weight according to claim 6, characterized in that: In the step of normalizing the weight of each time window, the following formula is used for normalization: In _r' = in _r / (In _r + in _n ) In _n' = in _n / (In _r + in _n ) Where w_r' and w_n' represent the weights of the normalized radar data and numerical forecast data, respectively.

8. The method for fusion of rainstorm forecast based on dynamic weight according to claim 7, characterized in that: In the step of performing data fusion based on dynamic weights, pre-processed radar data, and numerical forecast data to obtain fused heavy rain forecast data based on dynamic weights, for the forecast period of 0 to 2 hours, the following formula is used for fusion: P_f = w_r'×P_r + w_n'×P_n Among them, P_f represents the fused rainstorm forecast data, P_r and P_n represent radar data and numerical forecast data respectively; For the forecast period of 2 to 6 hours, the following formula is used for fusion: P_f(t) = w_r'(t) ×P_r(t) + w_n'(t)×P_n(t) Among them, P_f(t) represents the fused heavy rain forecast data at time t, w_r'(t) and w_n'(t) represent the normalized weights of the radar data and numerical forecast data at time t, respectively, and P_r(t) and P_n(t) represent the radar data and numerical forecast data at time t, respectively.

9. A rainstorm forecast fusion device based on dynamic weights, characterized in that: include: An acquisition module is used to obtain radar data and numerical forecast data for heavy rain forecast; Dynamic weight module, used to obtain dynamic weight; The fusion prediction module is used to perform data fusion based on dynamic weights, radar data and numerical forecast data to obtain fused heavy rain forecast data based on dynamic weights.

10. The dynamic weight-based rainstorm forecast fusion device according to claim 9, characterized in that: Also includes: A preprocessing module is used to preprocess the radar data and the numerical forecast data to obtain the preprocessed radar data and the numerical forecast data; The fusion prediction module is specifically configured to perform data fusion based on dynamic weights, pre-processed radar data and numerical forecast data.

11. The dynamic weight-based rainstorm forecast fusion device according to claim 10, characterized in that: The step of preprocessing the radar data and the numerical forecast data to obtain the preprocessed radar data and the numerical forecast data specifically includes: time alignment, spatial resolution unification and format unification of the radar data and the numerical forecast data.

12. The dynamic weight-based rainstorm forecast fusion device according to claim 9, characterized in that: In the step of obtaining the dynamic weight, the dynamic weight is calculated by the following steps: The forecast period of 0 to N hours is divided into several time windows; N is a positive number greater than or equal to 2; For each time window, calculate the historical root mean square error between the radar data and the numerical forecast data in the corresponding time window over the past M days; M is a positive integer greater than or equal to 3; For each time window, dynamic weights are calculated based on the historical root mean square error between radar data and numerical forecast data; The weight of each time window is normalized.

13. The dynamic weight-based rainstorm forecast fusion device according to claim 12, characterized in that: In the step of dividing the forecast period of 0 to N hours into several time windows, N is equal to 6; specifically, the step includes dividing the forecast period of 0 to 6 hours into 4 time windows of 0 to 1 hour, 1 to 2 hours, 2 to 4 hours and 4 to 6 hours.

14. The dynamic weight-based rainstorm forecast fusion device according to claim 12, characterized in that: In the step of calculating the dynamic weight according to the historical root mean square error of radar data and numerical forecast data for each time window, the weight is calculated using the following formula: w _r = 1 / RMSE _r w _n = 1 / RMSE _n Where w_r and w_n represent the weights of radar data and numerical forecast data, respectively; RMSE_r and RMSE_n represent the historical root mean square errors of radar data and numerical forecast data, respectively.

15. The dynamic weight-based rainstorm forecast fusion device according to claim 14, characterized in that: In the step of normalizing the weight of each time window, the following formula is used for normalization: In _r' = in _r / (In _r + in _n ) In _n' = in _n / (In _r + in _n ) Where w_r' and w_n' represent the weights of the normalized radar data and numerical forecast data, respectively.

16. The dynamic weight-based rainstorm forecast fusion device according to claim 15, characterized in that: In the step of performing data fusion based on dynamic weights, pre-processed radar data, and numerical forecast data to obtain fused heavy rain forecast data based on dynamic weights, for the forecast period of 0 to 2 hours, the following formula is used for fusion: P_f = w_r'×P_r + w_n'×P_n Among them, P_f represents the fused rainstorm forecast data, P_r and P_n represent radar data and numerical forecast data respectively; For the forecast period of 2 to 6 hours, the following formula is used for fusion: P_f(t) = w_r'(t) ×P_r(t) + w_n'(t)×P_n(t) Among them, P_f(t) represents the fused heavy rain forecast data at time t, w_r'(t) and w_n'(t) represent the normalized weights of the radar data and numerical forecast data at time t, respectively, and P_r(t) and P_n(t) represent the radar data and numerical forecast data at time t, respectively.

17. An electronic device, characterized in that: The invention comprises a processor and a memory, wherein the processor is used to execute a computer program stored in the memory to implement a heavy rain forecast fusion method based on dynamic weights as described in any one of claims 1 to 8.

18. A computer-readable storage medium, characterized in that The computer-readable storage medium stores at least one instruction, and when the at least one instruction is executed by a processor, the method for fusion of heavy rain forecast based on dynamic weights according to any one of claims 1 to 8 is implemented.

Images