A multi-source field fusion method and system for meteorological gridded data
By performing two-dimensional Fourier transform and decomposition on the meteorological gridded data, and combining low-wavenumber large-scale and high-wavenumber small-scale data processing, the problem of small-scale weather processes being flattened in ensemble forecasts was solved, thus improving forecast accuracy.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHINA ELECTRIC POWER RESEARCH INSTITUTE CO LTD
- Filing Date
- 2021-03-08
- Publication Date
- 2026-06-09
AI Technical Summary
In existing meteorological gridded data weather forecasts, high-frequency, small-scale severe weather processes exhibit significant differences in performance across different forecast fields. This causes ensemble forecasting methods to flatten these processes, affecting the accuracy of wind and solar energy resource forecasts.
Two-dimensional Fourier transform was used to divide meteorological element data into low wavenumber large-scale data and high wavenumber small-scale data. The data were then averaged and overlaid separately, and multi-source gridded data were fused to retain high-frequency small-scale information.
It improves the small-scale forecasting effect of weather forecasts, increases forecast accuracy, and maintains the error reduction effect of traditional ensemble forecasts.
Smart Images
Figure CN113094638B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of meteorological monitoring, and in particular relates to a multi-source field fusion method and system for meteorological gridded data. Background Technology
[0002] Meteorological gridded data is primarily used to observe / forecast meteorological conditions across the entire space, serving as crucial data support for weather analysis and forecasting. Gridding refers to the spatially uniform, discrete distribution of grid points, with each grid point containing a time series of observations / forecasts over a given period. The duration of the observation data is flexible; weather forecasts typically cover the short to medium term (1-10 days), while climate predictions generally target the long term (more than one month, or even more than one year).
[0003] Currently, the most commonly used meteorological / climate gridded data in the world include the US GFS (Global Forecasting System) weather forecast data, the European ECMWF (European Centre for Medium-Range Weather Forecasts) weather forecast data, the Canadian GEM (Global Environmental Multiscale model) weather forecast data, the Japanese GSM (Global Spectrum Model) weather forecast data, the US CFS (Climate Forecasting System) climate prediction data, and the CMIP (Coupled Model Intercomparison Project) long-term climate prediction data from the Global Climate Model Intercomparison Project. In addition, many researchers and institutions worldwide have built refined forecast models based on these forecast fields, producing refined forecast fields tailored to their specific needs.
[0004] Weather / climate forecasts inevitably contain errors and cannot be completely accurate. To reduce observation and forecast errors and improve accuracy, many processing methods for multiple weather fields have been introduced. One of the most important methods is ensemble forecasting (although it focuses on forecasts rather than observations, the processing methods of ensemble forecasting are equally applicable to observation fields). Ensemble forecasting averages multiple forecast fields, typically by simply averaging gridded data from different fields at the same location and time, or by using a weighted average according to certain rules. Compared to a single forecast field, the uncertainty of the averaged forecast field is reduced, thus improving forecast accuracy. However, averaging suppresses high-frequency, small-scale severe weather processes, such as wind speed ramp-ups, rapid irradiance fluctuations, and strong convection. This is mainly because the manifestation of high-frequency, small-scale severe weather processes varies greatly across different forecast fields. Therefore, averaging flattens out these high-frequency, small-scale processes, resulting in poor small-scale forecast performance. This is particularly detrimental to resource and power forecasts for wind and solar energy. Summary of the Invention
[0005] To overcome the shortcomings of the existing technology, this invention proposes a multi-source field fusion method for meteorological gridded data, comprising:
[0006] Collect gridded meteorological element data from multiple data sources within the same time and spatial range;
[0007] Two-dimensional Fourier transforms were performed on the gridded meteorological element data from each data source to obtain the complex plane wave data corresponding to each gridded meteorological element data.
[0008] According to the preset plan, the complex plane wave data corresponding to each gridded meteorological element data are divided into low wavenumber large-scale data and high wavenumber small-scale data.
[0009] The low wavenumber large-scale data is averaged and superimposed with any one of the high wavenumber small-scale data. Then, a two-dimensional inverse Fourier transform is performed to obtain the fused multi-source gridded new meteorological element data.
[0010] Preferably, the step of dividing the complex plane wave data corresponding to each gridded meteorological element data into low wavenumber large-scale data and high wavenumber small-scale data according to a preset scheme includes:
[0011] If the wavenumber of the complex plane wave data is within a preset threshold range, it is classified as low wavenumber large-scale data; otherwise, it is classified as high wavenumber small-scale data.
[0012] Preferably, the preset threshold range is 0 to 5.
[0013] Preferably, the step of averaging the low-wavenumber large-scale data, superimposing any one of the high-wavenumber small-scale data, and then performing a two-dimensional inverse Fourier transform to obtain the fused multi-source gridded new meteorological element data includes:
[0014] Based on the number of data sources, the low wavenumber large-scale data corresponding to each wavenumber in all low wavenumber large-scale data are averaged to obtain the averaged data of low wavenumber large-scale data for each wavenumber.
[0015] After the averaged data of all wavenumbers corresponding to the large-scale low wavenumber data are superimposed once, they are superimposed a second time with any of the high wavenumber small-scale data, and then the fused multi-source gridded new meteorological element data are obtained by two-dimensional inverse Fourier transform.
[0016] Preferably, the formula for calculating the averaged data is as follows:
[0017]
[0018] Among them, fft2[W average (k l [)] represents the averaged data of large-scale low-wavenumber data with wavenumber l, k l This represents large-scale data with a low wavenumber of l, where l ranges from 0 to 5, and fft2[W i (k l ] represents the i-th low-wavenumber large-scale data with wavenumber l in the two-dimensional discrete Fourier transform, and N represents the number of data sources.
[0019] Preferably, the average includes: cumulative average or weighted average.
[0020] Preferably, the calculation formula for the fused multi-source gridded new meteorological element data is as follows:
[0021] W new (x,y)=ifft2{fft2[W new (k)]}
[0022] Among them, W new (x,y) represents the new meteorological data after fusion at x and y in the two-dimensional spatial coordinate system of the gridded meteorological element data, fft2[W new [(k)] represents the new stacked full-wavenumber scale meteorological data after averaging the low-wavenumber large-scale data and superimposing any high-wavenumber small-scale data after the two-dimensional discrete Fourier transform, and ifft2(·) represents the two-dimensional discrete Fourier inverse transform function.
[0023] Based on the same inventive concept, this invention also provides a multi-source field fusion system for meteorological gridded data, including: an acquisition module, a calculation module, a partitioning module, and a fusion module;
[0024] The acquisition module is used to acquire gridded meteorological element data from multiple data sources within the same time and spatial range;
[0025] The calculation module is used to perform two-dimensional Fourier transform on the gridded meteorological element data of each data source to obtain the complex plane wave data corresponding to each gridded meteorological element data.
[0026] The partitioning module is used to divide the complex plane wave data corresponding to each gridded meteorological element data into low wavenumber large-scale data and high wavenumber small-scale data according to a preset scheme.
[0027] The fusion module is used to average the low wavenumber large-scale data, and after superimposing any one of the high wavenumber small-scale data, it performs a two-dimensional inverse Fourier transform to obtain the fused multi-source gridded new meteorological element data.
[0028] Preferably, the partitioning module includes: a partitioning judgment submodule;
[0029] The division and judgment submodule is used to classify complex plane wave data as low wavenumber large-scale data if the wavenumber is within a preset threshold range, and otherwise classify it as high wavenumber small-scale data.
[0030] Preferably, the preset threshold range is 0 to 5.
[0031] Compared with the closest existing technology, the present invention has the following beneficial effects:
[0032] This invention implements a multi-source field fusion method and system for meteorological gridded data, comprising: collecting gridded meteorological element data from multiple data sources within the same time and spatial range; performing two-dimensional Fourier transform on the gridded meteorological element data from each data source to obtain complex plane wave data corresponding to each gridded meteorological element data; dividing the complex plane wave data corresponding to each gridded meteorological element data into low wavenumber large-scale data and high wavenumber small-scale data according to a preset scheme; averaging the low wavenumber large-scale data, superimposing any one of the high wavenumber small-scale data, and then performing a two-dimensional inverse Fourier transform to obtain fused multi-source gridded new meteorological element data; this invention improves the small-scale forecasting effect of weather forecasts by averaging and smoothing high-frequency small-scale processes, while also demonstrating the advantages of traditional ensemble forecasting by reducing errors and improving forecast accuracy through averaging multiple forecast fields. Attached Figure Description
[0033] Figure 1This is a schematic diagram of a multi-source field fusion method for meteorological gridded data provided by the present invention;
[0034] Figure 2 This is an example diagram of two-dimensional Fourier decomposition and fusion of multiple meteorological fields in an embodiment of the present invention;
[0035] Figure 3 This is a schematic diagram of the technical process in an embodiment of the present invention;
[0036] Figure 4 This invention provides a schematic diagram of a multi-source field fusion system for meteorological gridded data. Detailed Implementation
[0037] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.
[0038] Example 1:
[0039] The application principle of the present invention will be further described below with reference to the accompanying drawings and specific embodiments.
[0040] This invention provides a multi-source field fusion method for meteorological gridded data, specifically, as follows: Figure 1 As shown, it includes:
[0041] Step 1: Collect gridded meteorological element data from multiple data sources within the same time and spatial range;
[0042] Determine the meteorological element field W(x,y) at a specific time and altitude that needs to be decomposed (decomposition is performed separately for each time point), such as the gridded wind speed field at an altitude of 70m at a specific time, where x and y are two-dimensional spatial coordinates. The meteorological element field corresponds to the gridded meteorological element data.
[0043] Step 2: Perform two-dimensional Fourier transform on the gridded meteorological element data of each data source to obtain the complex plane wave data corresponding to each gridded meteorological element data;
[0044] Performing a two-dimensional Fourier transform on W(x,y) (which can be implemented using Python's np.fft.fft2 function) yields the decomposed wavenumber space meteorological field fft2[W(k l The meteorological field in wavenumber space is the complex plane wave corresponding to each gridded meteorological element field.
[0045] The two-dimensional discrete Fourier transform of an image (M*N) can transform the image from the spatial domain to the frequency domain. In the spatial domain, x and y represent spatial coordinates, and in the frequency domain, u and v represent frequencies. Here, f(x,y) represents the original image at two-dimensional spatial coordinates x and y (i.e., the gridded meteorological element data image M*N), F(u,v) represents the result after the Fourier transform at frequencies u and v (i.e., the complex plane wave data corresponding to each gridded meteorological element), M is the number of vertical data points in the image, and N is the number of horizontal data points in the image. The formula for the two-dimensional discrete Fourier transform is as follows:
[0046]
[0047] In Python, the NumPy library's `fft` module provides a pre-implemented 2D Discrete Fourier Transform function, `fft2`. It takes a grayscale image as input and outputs the result after a 2D Discrete Fourier Transform. However, the implementation doesn't directly use the formula mentioned above; instead, it uses the Fast Fourier Transform. The result needs to be visualized by taking the absolute value using `abs`, but the visual effect is not ideal because the Fourier spectrum range is very large. Therefore, a logarithmic transform is used to improve the visual effect.
[0048] When using the log function, it should be written as log(1+x) instead of directly using log(x) to avoid performing logarithmic operations on zero. Additionally, the origin of the image transformation needs to be moved to the center of the frequency domain rectangle, so the fftshift function should be used on the result of fft2. Finally, log can also be used to improve visualization.
[0049] Step 3: According to the preset scheme, divide the complex plane wave data corresponding to each gridded meteorological element data into low wavenumber large-scale data and high wavenumber small-scale data;
[0050] Step 3 includes:
[0051] 3-1 If the wavenumber of the complex plane wave data is within the preset threshold range, it is classified as low wavenumber large-scale data; otherwise, it is classified as high wavenumber small-scale data.
[0052] Extracting the low-wavenumber large-scale meteorological field fft2[W(k) after two-dimensional Fourier decomposition l (Generally, wavenumbers of 0-5 represent low wavenumbers and large scales, while wavenumbers of 6-∞ represent high wavenumbers and small scales). When the complex plane waves corresponding to each gridded meteorological element field are within the range of 0-5, they are classified as low wavenumber large-scale fields; otherwise, they are classified as high wavenumber small-scale fields.
[0053] Step 4: Average the low wavenumber large-scale data, and superimpose any one of the high wavenumber small-scale data, then perform a two-dimensional inverse Fourier transform to obtain the fused multi-source gridded new meteorological element data.
[0054] Step 4 includes:
[0055] 4-1 Based on the number of data sources, the low wavenumber large-scale data corresponding to each wavenumber in all low wavenumber large-scale data are averaged to obtain the averaged data of the low wavenumber large-scale data corresponding to each wavenumber.
[0056] Averaging multiple large-scale low-wavenumber meteorological fields can be achieved using methods such as cumulative averaging or weighted averaging. The averaged data of the large-scale low-wavenumber meteorological fields is obtained, for example, if there are N meteorological fields:
[0057]
[0058] Among them, fft2[W average (k l [)] represents large-scale data with a low wavenumber of l and a wavenumber of l, obtained by averaging the two-dimensional discrete Fourier transform. l This represents large-scale data with a low wavenumber of l, where l ranges from 0 to 5, and fft2[W i (k l ] represents the i-th low-wavenumber large-scale data with wavenumber l in the two-dimensional discrete Fourier transform, and N represents the number of data sources.
[0059] 4-2 After the averaged data of all wavenumbers corresponding to the low wavenumber large-scale data are superimposed once, they are superimposed twice with any of the high wavenumber small-scale data, and the fused multi-source gridded new meteorological element data are obtained by two-dimensional inverse Fourier transform.
[0060] A high-wavenumber, small-scale meteorological field is selected and superimposed and fused with an averaged low-wavenumber, large-scale meteorological field.
[0061] 1) Arbitrarily select a high wavenumber small-scale meteorological field fft2[W(k 6-∞ )).
[0062] 2) The high-wavenumber, small-scale meteorological field is superimposed on the averaged low-wavenumber, large-scale meteorological field to obtain the superimposed field fft2[W new (k)], that is:
[0063] fft2[W new (k)]=fft2[W average (k l )]+fft2[W(k 6-∞ )]
[0064] 3) Set fft2[W new Perform a two-dimensional inverse Fourier transform on (k) (which can be implemented using Python's np.fft.ifft2 function) to obtain the fused new meteorological field W. new (x,y), where W new (x,y) represents the fused new meteorological data at two-dimensional spatial coordinates x and y, respectively, fft2[W new [(k)] represents the new stacked full-wavenumber scale meteorological data after averaging the low-wavenumber large-scale data and superimposing any high-wavenumber small-scale data after the two-dimensional discrete Fourier transform, and ifft2(·) represents the two-dimensional discrete Fourier inverse transform function.
[0065] Right now:
[0066] W new (x,y)=ifft2{fft2[W new (k)]}
[0067] Example 2:
[0068] This invention proposes a multi-source field fusion method for meteorological gridded data, such as... Figure 2 The steps shown are as follows:
[0069] This invention addresses the problem of poor performance of small-scale weather processes due to gridded averaging in ensemble forecasts. It proposes a multi-field fusion method based on two-dimensional Fourier analysis. Each two-dimensional forecast field at a certain altitude (the specific altitude and meteorological elements depend on the requirements; for example, wind fields at 70m altitude are typically required in the wind energy field, which is a typical two-dimensional forecast field) undergoes two-dimensional Fourier decomposition, resulting in large-scale and small-scale two-dimensional forecast fields. The large-scale two-dimensional forecast field is averaged using traditional ensemble forecasting methods, while the small-scale two-dimensional forecast field is not averaged. Instead, an arbitrary member is selected and superimposed on the averaged large-scale two-dimensional forecast field, ultimately yielding a fused multi-forecast field result. A schematic diagram of this process is shown below. Figure 2 As shown. This technique combines the advantages of traditional ensemble forecasting—namely, reducing errors through averaging multiple forecast fields—with the avoidance of smoothing out small-scale, high-frequency, severe weather events, thereby further improving forecast accuracy. This invention avoids smoothing out the effects of small-scale, high-frequency, severe weather events, thus further improving forecast accuracy. This invention is applicable not only to forecast fields but also to observation fields.
[0070] The overall technical approach of this patent is as follows: Figure 3 As shown, it consists of 3 steps:
[0071] 1. Based on the two-dimensional Fourier transform method, decompose gridded multi-meteorological fields.
[0072] 1) Determine the meteorological element field W(x,y) at a certain time and altitude that needs to be decomposed (the decomposition is performed separately for each time). For example, the gridded wind speed field at a height of 70m at a certain time, where x and y are two-dimensional spatial coordinates.
[0073] 2) Perform a two-dimensional Fourier transform on W(x,y) (which can be implemented using Python's np.fft.fft2 function) to obtain the meteorological field fft2[W(k) in the decomposed wavenumber space. l The meteorological field in wavenumber space is the complex plane wave corresponding to each gridded meteorological element field.
[0074] The two-dimensional discrete Fourier transform of an image (M*N) can transform the image from the spatial domain to the frequency domain. In the spatial domain, x and y represent spatial coordinates, and in the frequency domain, u and v represent frequencies. Here, f(x,y) represents the original image at two-dimensional spatial coordinates x and y (i.e., the gridded meteorological element data image M*N), F(u,v) represents the result after the Fourier transform at frequencies u and v (i.e., the complex plane wave data corresponding to each gridded meteorological element), M is the number of vertical data points in the image, and N is the number of horizontal data points in the image. The formula for the two-dimensional discrete Fourier transform is as follows:
[0075]
[0076] In Python, the NumPy library's `fft` module provides a pre-implemented 2D Discrete Fourier Transform function, `fft2`. It takes a grayscale image as input and outputs the result after a 2D Discrete Fourier Transform. However, the implementation doesn't directly use the formula mentioned above; instead, it uses the Fast Fourier Transform. The result needs to be visualized by taking the absolute value using `abs`, but the visual effect is not ideal because the Fourier spectrum range is very large. Therefore, a logarithmic transform is used to improve the visual effect.
[0077] When using the log function, it should be written as log(1+x) instead of directly using log(x) to avoid performing logarithmic operations on zero. Additionally, the origin of the image transformation needs to be moved to the center of the frequency domain rectangle, so the fftshift function should be used on the result of fft2. Finally, log can also be used to improve visualization.
[0078] 2. Averaging the decomposed large-scale meteorological fields with low wavenumbers.
[0079] 1) Extract the low-wavenumber large-scale meteorological field fft2[W(k) after two-dimensional Fourier decomposition. l(Generally, wavenumbers of 0-5 represent low wavenumbers and large scales, while wavenumbers of 6-∞ represent high wavenumbers and small scales). When the complex plane waves corresponding to each gridded meteorological element field are within the range of 0-5, they are classified as low wavenumber large-scale fields; otherwise, they are classified as high wavenumber small-scale fields.
[0080] 2) Averaging multiple low-wavenumber, large-scale meteorological fields can be performed using methods such as cumulative averaging or weighted averaging. For example, if there are N meteorological fields, then:
[0081]
[0082] Among them, fft2[W average (k l [)] represents large-scale data with a low wavenumber of l and a wavenumber of l, obtained by averaging the two-dimensional discrete Fourier transform. l This represents large-scale data with a low wavenumber of l, where l ranges from 0 to 5, and fft2[W i (k l ] represents the i-th low-wavenumber large-scale data with wavenumber l in the two-dimensional discrete Fourier transform, and N represents the number of data sources.
[0083] This yields the averaged low-wavenumber large-scale meteorological field fft2[W] average (k l )).
[0084] (3) Select a high wavenumber small-scale meteorological field and superimpose and fuse it with the averaged low wavenumber large-scale meteorological field.
[0085] 1) Arbitrarily select a high wavenumber small-scale meteorological field fft2[W(k 6-∞ )).
[0086] 2) The high-wavenumber, small-scale meteorological field is superimposed on the averaged low-wavenumber, large-scale meteorological field to obtain the superimposed field fft2[W new (k)], that is:
[0087] fft2[W new (k)]=fft2[W average (k l )]+fft2[W(k 6-∞ )]
[0088] 3) Set fft2[W new Perform a two-dimensional inverse Fourier transform on (k) (which can be implemented using Python's np.fft.ifft2 function) to obtain the fused new meteorological field W. new (x,y), where W new(x,y) represents the fused new meteorological data at two-dimensional spatial coordinates x and y, respectively, fft2[W new [(k)] represents the new stacked full-wavenumber scale meteorological data after averaging the low-wavenumber large-scale data and superimposing any high-wavenumber small-scale data after the two-dimensional discrete Fourier transform, and ifft2(·) represents the two-dimensional discrete Fourier inverse transform function.
[0089] Right now:
[0090] W new (x,y)=ifft2{fft2[W new (k)]}
[0091] Example 3:
[0092] Based on the same inventive concept, this invention also provides a multi-source field fusion system for meteorological gridded data, such as... Figure 4 As shown, it includes: a data acquisition module, a calculation module, a partitioning module, and a fusion module;
[0093] The acquisition module is used to collect gridded meteorological element data from multiple data sources within the same time and space range;
[0094] The calculation module is used to perform two-dimensional Fourier transform on the gridded meteorological element data of each data source to obtain the complex plane wave data corresponding to each gridded meteorological element data.
[0095] The partitioning module is used to divide the complex plane wave data corresponding to each gridded meteorological element data into low wavenumber large-scale data and high wavenumber small-scale data according to a preset scheme.
[0096] The fusion module is used to average the low wavenumber large-scale data, and after superimposing any one of the high wavenumber small-scale data, it performs a two-dimensional inverse Fourier transform to obtain the fused multi-source gridded new meteorological element data.
[0097] The partitioning module includes: a partitioning judgment submodule;
[0098] The segmentation and judgment submodule is used to classify complex plane wave data as low wavenumber large-scale data if the wavenumber is within a preset threshold range, and otherwise classify it as high wavenumber small-scale data.
[0099] The preset threshold range is 0 to 5.
[0100] The fusion module includes: an averaging submodule and an overlay submodule;
[0101] The averaging submodule is used to average the low wavenumber large-scale data corresponding to each wavenumber in all low wavenumber large-scale data based on the number of data sources, so as to obtain the averaged data of low wavenumber large-scale data corresponding to each wavenumber.
[0102] The overlay submodule is used to overlay the averaged data of all wavenumber-corresponding low wavenumber large-scale data once, and then overlay it a second time with any one of the high wavenumber small-scale data. Finally, through a two-dimensional inverse Fourier transform, the fused multi-source gridded new meteorological data is obtained.
[0103] The formula for calculating the averaged data is as follows:
[0104]
[0105] Among them, fft2[W average (k l [)] represents the averaged data of large-scale low-wavenumber data with wavenumber l, k l This represents large-scale data with a low wavenumber of l, where l ranges from 0 to 5, and fft2[W i (k l ] represents the i-th low-wavenumber large-scale data with wavenumber l in the two-dimensional discrete Fourier transform, and N represents the number of data sources.
[0106] The averages include: cumulative averages or weighted averages.
[0107] The calculation formula for the fused multi-source gridded new meteorological element data is as follows:
[0108] W new (x,y)=ifft2{fft2[W new (k)]}
[0109] Among them, W new (x,y) represents the new meteorological data after fusion at x and y in the two-dimensional spatial coordinate system of the gridded meteorological element data, fft2[W new [(k)] represents the new stacked full-wavenumber scale meteorological data after averaging the low-wavenumber large-scale data and superimposing any high-wavenumber small-scale data after the two-dimensional discrete Fourier transform, and ifft2(·) represents the two-dimensional discrete Fourier inverse transform function.
[0110] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations 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, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0111] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0112] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0113] 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 its scope of protection. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that after reading the present invention, they can still make various changes, modifications or equivalent substitutions to the specific implementation of the invention, but these changes, modifications or equivalent substitutions are all within the scope of protection of the pending claims of the invention.
Claims
1. A multi-source field fusion method for meteorological gridded data, characterized in that, include: Collect gridded meteorological element data from multiple data sources within the same time and spatial range; Two-dimensional Fourier transforms were performed on the gridded meteorological element data from each data source to obtain the complex plane wave data corresponding to each gridded meteorological element data. According to the preset plan, the complex plane wave data corresponding to each gridded meteorological element data are divided into low wavenumber large-scale data and high wavenumber small-scale data. The low wavenumber large-scale data is averaged and superimposed with any one of the high wavenumber small-scale data. Then, a two-dimensional inverse Fourier transform is performed to obtain the fused multi-source gridded new meteorological element data. The step of dividing the complex plane wave data corresponding to each gridded meteorological element data into low wavenumber large-scale data and high wavenumber small-scale data according to a preset scheme includes: If the wavenumber of the complex plane wave data is within the preset threshold range, it is classified as low wavenumber large-scale data; otherwise, it is classified as high wavenumber small-scale data. The process of averaging the low-wavenumber large-scale data, superimposing any one of the high-wavenumber small-scale data, and then performing a two-dimensional inverse Fourier transform to obtain the fused multi-source gridded new meteorological element data includes: Based on the number of data sources, the low wavenumber large-scale data corresponding to each wavenumber in all low wavenumber large-scale data are averaged to obtain the averaged data of low wavenumber large-scale data for each wavenumber. After the averaged data of all wavenumbers corresponding to the large-scale low wavenumber data are superimposed once, they are superimposed a second time with any of the high wavenumber small-scale data, and then the fused multi-source gridded new meteorological element data are obtained by two-dimensional inverse Fourier transform.
2. The method according to claim 1, characterized in that, The preset threshold range is 0 to 5.
3. The method according to claim 1, characterized in that, The formula for calculating the averaged data is as follows: in, Indicates wave number Averaged data of low wavenumber large-scale data. Indicates wave number Low wavenumber large-scale data, The value range is 0-5. The first two-dimensional discrete Fourier transform represents the... The number of waves is Low-wavenumber, large-scale data, where N represents the number of data sources.
4. The method according to claim 1, characterized in that, The averages include: cumulative averages or weighted averages.
5. The method according to claim 1, characterized in that, The calculation formula for the fused multi-source gridded new meteorological element data is as follows: in, This represents the new meteorological data after fusion at x and y points in the two-dimensional spatial coordinate system of the gridded meteorological element data. This represents the new stacked full-wavenumber scale meteorological data obtained by averaging the low-wavenumber large-scale data and superimposing it with any high-wavenumber small-scale data after a two-dimensional discrete Fourier transform. This represents the two-dimensional discrete Fourier inverse transform function.
6. A multi-source field fusion system for meteorological gridded data, characterized in that, include: The module comprises an acquisition module, a calculation module, a partitioning module, and a fusion module. The acquisition module is used to acquire gridded meteorological element data from multiple data sources within the same time and spatial range; The calculation module is used to perform two-dimensional Fourier transform on the gridded meteorological element data of each data source to obtain the complex plane wave data corresponding to each gridded meteorological element data. The partitioning module is used to divide the complex plane wave data corresponding to each gridded meteorological element data into low wavenumber large-scale data and high wavenumber small-scale data according to a preset scheme. The fusion module is used to average the low wavenumber large-scale data, and after superimposing any one of the high wavenumber small-scale data, it performs a two-dimensional inverse Fourier transform to obtain the fused multi-source gridded new meteorological element data. The partitioning module includes: a partitioning judgment submodule; The division and judgment submodule is used to classify complex plane wave data as low wavenumber large-scale data if the wavenumber is within a preset threshold range, and otherwise classify it as high wavenumber small-scale data. The fusion module includes: Based on the number of data sources, the low wavenumber large-scale data corresponding to each wavenumber in all low wavenumber large-scale data are averaged to obtain the averaged data of low wavenumber large-scale data for each wavenumber. After the averaged data of all wavenumbers corresponding to the large-scale low wavenumber data are superimposed once, they are superimposed a second time with any of the high wavenumber small-scale data, and then the fused multi-source gridded new meteorological element data are obtained by two-dimensional inverse Fourier transform.
7. The system according to claim 6, characterized in that, The preset threshold range is 0 to 5.