Meteorological data processing method and device and storage medium

Abstract

The invention relates to the field of data processing, and discloses a meteorological data processing method and device, and a storage medium, and the method comprises the steps: determining a to-be-interpolated region corresponding to each to-be-interpolated algorithm according to the region range of a target region, the first latitude and longitude resolution of a large-scale region, and at least two to-be-selected interpolation algorithms; according to the second longitude and latitude resolution and each to-be-selected interpolation algorithm, performing interpolation processing on the meteorological data of the corresponding to-be-interpolated region to obtain a to-be-selected interpolation result, to-be-selected interpolation time and to-be-selected interpolation error corresponding to each to-be-selected interpolation algorithm; and determining a target interpolation algorithm according to the to-be-selected interpolation time and the to-be-selected interpolation error corresponding to each to-be-selected interpolation algorithm, and taking a to-be-selected interpolation result corresponding to the target interpolation algorithm as a target interpolation result. Through the technical scheme of the invention, the interpolation time and the interpolation precision of the interpolation algorithm are balanced, and the interpolation effect under the condition of no accurate target data is improved.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and in particular to a meteorological data processing method, device and storage medium. Background Art

[0002] Global Numerical Weather Prediction (NWP) uses computers to simulate atmospheric motion processes through a combination of physical processes and mathematical methods to predict weather conditions. Although the continuous development of meteorological data collection technology has made it increasingly easy to obtain large-scale low-resolution meteorological data around the world, it is still difficult to obtain high-resolution data. Therefore, it is necessary to interpolate low-resolution data into high-resolution data to meet the precise data needs of local areas.

[0003] Interpolation methods commonly used in meteorological data processing include bilinear interpolation, cubic spline interpolation, Kriging interpolation, etc. Using these interpolation methods alone cannot improve the interpolation accuracy in local areas. In addition, in the absence of accurate real data, it is difficult to judge the accuracy of the interpolation results of these interpolation algorithms. Summary of the invention

[0004] In order to solve the above technical problems, the present invention provides a meteorological data processing method, device and storage medium, which obtain the interpolation time and interpolation accuracy of the trade-off interpolation algorithm, improve the interpolation effect in the absence of accurate target data, and realize the accurate and efficient mapping of low-resolution large-scale regional meteorological observation data to local simulation area high-resolution data.

[0005] An embodiment of the present invention provides a method for processing meteorological data, the method comprising:

[0006] Determine, according to the area range of the target area, the first latitude and longitude resolution of the large-scale area, and at least two interpolation algorithms to be selected, an area to be interpolated corresponding to each interpolation algorithm to be selected;

[0007] According to the second longitude and latitude resolution and each candidate interpolation algorithm, interpolation processing is performed on the meteorological data of the corresponding region to be interpolated, so as to obtain a candidate interpolation result, a candidate interpolation time and a candidate interpolation error corresponding to each candidate interpolation algorithm; wherein the second longitude and latitude resolution is smaller than the first longitude and latitude resolution;

[0008] According to the candidate interpolation times and the candidate interpolation errors corresponding to the candidate interpolation algorithms, a target interpolation algorithm is determined, and the candidate interpolation results corresponding to the target interpolation algorithm are used as the target interpolation results.

[0009] An embodiment of the present invention provides an electronic device, the electronic device comprising:

[0010] Processor and memory;

[0011] The processor is used to execute the steps of the meteorological data processing method described in any embodiment by calling the program or instruction stored in the memory.

[0012] An embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a program or instruction, wherein the program or instruction enables a computer to execute the steps of the meteorological data processing method described in any embodiment.

[0013] The embodiments of the present invention have the following technical effects:

[0014] The interpolation area corresponding to each interpolation algorithm is determined according to the regional scope of the target area, the first longitude and latitude resolution of the large-scale area and at least two interpolation algorithms to be selected, that is, a suitable area is selected from the large-scale area to facilitate subsequent interpolation processing. According to the second longitude and latitude resolution and each interpolation algorithm to be selected, the meteorological data of the corresponding interpolation area is interpolated to obtain the interpolation results, interpolation time and interpolation error corresponding to each interpolation algorithm to be selected, so as to facilitate subsequent analysis according to accuracy and rate. According to the interpolation time and interpolation error corresponding to each interpolation algorithm to be selected, the target interpolation algorithm is determined, and the interpolation result corresponding to the target interpolation algorithm is used as the target interpolation result, so as to obtain the interpolation time and interpolation accuracy of the trade-off interpolation algorithm, improve the interpolation effect in the absence of accurate target data, and realize accurate and efficient mapping of low-resolution large-scale regional meteorological observation data to high-resolution data of local simulation area. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 is a flow chart of a meteorological data processing method provided by an embodiment of the present invention;

[0017] Figure 2 It is a structural schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0018] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be described clearly and completely below. Apparently, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0019] The meteorological data processing method provided by the embodiments of the present invention is mainly applicable to the situation where the time and accuracy of different interpolation algorithms can be measured without target data, and a suitable interpolation algorithm is selected for local interpolation processing. The meteorological data processing method provided by the embodiments of the present invention can be executed by an electronic device.

[0020] Embodiment 1

[0021] Figure 1 is a flowchart of a meteorological data processing method provided by the embodiments of the present invention. Refer to Figure 1 and the meteorological data processing method specifically includes:

[0022] S110. Determine the interpolation area corresponding to each interpolation algorithm to be used according to the area range of the target area, the first longitude and latitude resolution of the large-scale area, and at least two interpolation algorithms to be selected.

[0023] Among them, the target area is a small-scale area that requires interpolation to obtain high-resolution meteorological data. The large-scale area is an area with low-resolution meteorological data. It should be noted that the range of the target area is within the range of the large-scale area. The area range is an area determined by the maximum longitude, minimum longitude, maximum latitude, and minimum latitude. The first longitude and latitude resolution is the longitude and latitude resolution of the existing meteorological data in the large-scale area, specifically including the first longitude resolution and the first latitude resolution. The interpolation algorithms to be selected are the algorithms used for interpolation, and a suitable algorithm needs to be selected from the interpolation algorithms to be selected later to provide a high-resolution interpolation result.

[0024] The interpolation area to be used is an area expanded on the basis of the target area, and the purpose is to ensure that the edges in the target area can also have better interpolation results.

[0025] Specifically, since the target area is an area that requires high-resolution meteorological data, the data at the edges around the target area should also be accurately interpolated. Therefore, a certain degree of expansion is required to ensure that the data at the edges around the target area can also be interpolated. Thus, on the basis of the area range of the target area, the target area can be expanded according to the expansion range of the surrounding data required by different interpolation algorithms to be selected, in combination with the first longitude and latitude resolution of the large-scale area, to obtain the interpolation area corresponding to each interpolation algorithm to be used.

[0026] Based on the above examples, the following method can be used to determine the interpolation area corresponding to each interpolation algorithm to be selected according to the area range of the target area, the first longitude and latitude resolution of the large-scale area, and at least two interpolation algorithms to be selected:

[0027] According to the area range of the target area, determine the first longitude upper limit, the first longitude lower limit, the first latitude upper limit, and the first latitude lower limit corresponding to the target area;

[0028] For each interpolation algorithm to be selected, determine the second longitude upper limit and the second longitude lower limit according to the first longitude upper limit, the first longitude lower limit, the first longitude resolution in the first longitude and latitude resolution, and the longitude expansion range corresponding to the interpolation algorithm to be selected;

[0029] According to the first latitude upper limit, the first latitude lower limit, the first latitude resolution in the first longitude and latitude resolution, and the latitude expansion range corresponding to the interpolation algorithm to be selected, determine the second latitude upper limit and the second latitude lower limit;

[0030] According to the second longitude upper limit, the second longitude lower limit, the second latitude upper limit, and the second latitude lower limit, determine the interpolation area corresponding to the interpolation algorithm to be selected.

[0031] Among them, the first longitude upper limit, the first longitude lower limit, the first latitude upper limit, and the first latitude lower limit are the edge longitude and latitude information of the target area. The first longitude resolution is the resolution of meteorological data in the longitude direction in the large-scale area. The longitude expansion range is the amount of data (data points) required by the interpolation algorithm to be selected in the longitude direction for a certain interpolation point. The first latitude resolution is the resolution of meteorological data in the latitude direction in the large-scale area. The latitude expansion range is the amount of data (data points) required by the interpolation algorithm to be selected in the latitude direction for a certain interpolation point. The second longitude upper limit, the second longitude lower limit, the second latitude upper limit, and the second latitude lower limit are the edge longitude and latitude information of the interpolation area.

[0032] Specifically, analyze the regional scope of the target area to obtain the upper limit of the first longitude, the lower limit of the first longitude, the upper limit of the first latitude, and the lower limit of the first latitude corresponding to the target area. For each candidate interpolation algorithm, the algorithm can be analyzed to obtain the longitude expansion range and the latitude expansion range corresponding to the candidate interpolation algorithm. Based on the upper limit of the first longitude and the lower limit of the first longitude, expand the size of the product of the first longitude resolution and the longitude expansion range along the direction of expanding the longitude range of the target area to obtain the upper limit of the second longitude and the lower limit of the second longitude. Based on the upper limit of the first latitude and the lower limit of the first latitude, expand the size of the product of the first latitude resolution and the latitude expansion range along the direction of expanding the latitude range of the target area to obtain the upper limit of the second latitude and the lower limit of the second latitude. The area enclosed by the upper limit of the second longitude, the lower limit of the second longitude, the upper limit of the second latitude, and the lower limit of the second latitude is used as the interpolation area corresponding to the corresponding candidate interpolation algorithm.

[0033] After segmenting the low-resolution meteorological data of a large-scale area according to the target area range to obtain the interpolation area, interpolation operations are performed, that is, appropriate data slices (interpolation areas) are selected. This has important theoretical basis and practical significance in the interpolation process, especially in the context of large-scale meteorological data processing. The reasons are as follows: (1) Using data slices for interpolation can reduce the computational complexity and improve the computational efficiency, avoiding data redundancy. General meteorological data sets are very large, containing multiple regions, time periods, and variables. Directly interpolating the entire data set not only involves huge computational amounts but is also unnecessary in the case of accurate forecasting. Therefore, by selecting data slices, interpolation can be performed only on the data within the regional scope of the target area, reducing the unnecessary computational burden and improving the timeliness of actual forecasting; (2) Interpolation after data slicing can improve the interpolation accuracy. The interpolation accuracy depends on the quality and density of the original large-scale regional meteorological data used. Combining with the first fundamental theorem of geography, it can be known that if the data is too far from the interpolation points in the target area, the interpolation error will increase significantly. Moreover, according to the characteristics of meteorological data, considering the local nature of the changes in meteorological elements such as temperature, humidity, and air pressure, appropriate data slices can ensure that the local characteristics can be considered during the interpolation process, avoiding interference from cross-regional data mixing on the interpolation effect. Therefore, before interpolation, selecting an interpolation area (data slice) that is close to and covers the target area can reduce the computational burden, improve the interpolation accuracy, handle boundary problems, avoid redundant data, and retain the local characteristics of the data, significantly optimizing the interpolation process. A reasonable interpolation area can balance the interpolation accuracy and computational cost, especially in the scenario of large-scale meteorological data processing, ensuring that the interpolation process is both efficient and accurate.

[0034] Exemplarily, obtain the large-scale area of the meteorological data held by the user. The specific parameters include: the first longitude and latitude resolution , and are the first longitude resolution and the first latitude resolution respectively. The resolution of the meteorological data in the target area , and are the second longitude resolution and the second latitude resolution respectively. The data resolution r refers to the side length of the grid obtained by evenly dividing the regional data into grids.

[0035] If the characteristics of the boundary problem can be determined for the selected interpolation area, the adjacent point range can be directly selected. If the selected interpolation area does not handle the boundary points of the target area smoothly, the possibility of large errors in interpolation at the boundary points will increase greatly. For example, when performing bilinear interpolation, if the selected interpolation area does not consider the regional boundary, it will result in a lack of data points at the boundary points and the interpolation cannot be carried out, and the error will increase greatly.

[0036] The strategy for selecting the adjacent point range for the interpolation area corresponding to the bilinear algorithm is as follows: Bilinear interpolation depends on the 4 grid points closest to the interpolation point, that is, the interpolation points in the target area need to select at least 4 points covering the points above, below, left, and right in the large-scale area, that is, at least one grid point (the first longitude resolution or the first latitude resolution) outside each boundary of the target area. Cubic spline interpolation is carried out by polynomial fitting and requires more adjacent grid point data. Taking two-dimensional data as an example, each interpolation point requires 8 grid points in its vicinity (4 grid points in each direction). Therefore, when slicing the large-scale area, the selection needs to be extended to the grid points in two rows and two columns outside the four sides of the target area, that is, the longitude expansion range and the latitude expansion range are 2, and the interpolation area is obtained.

[0037] For the bilinear interpolation algorithm, assume that the first longitude and latitude range of the target area is from (lat_B_min, lon_B_min) to (lat_B_max, lon_B_max), lat_B_min is the lower limit of the first longitude, lon_B_min is the lower limit of the first latitude, lat_B_max is the upper limit of the first longitude, and lon_B_max is the upper limit of the first latitude. The first longitude and latitude resolutions of the large-scale data are Δlat_A and Δlon_A, and Δlat_A and Δlon_A are the first longitude resolution and the first latitude resolution respectively. The area range of the interpolation area to be selected is: The second latitude range: [lat_B_min - Δlat_A, lat_B_max + Δlat_A]; The second longitude range: [lon_B_min - Δlon_A, lon_B_max + Δlon_A].

[0038] For cubic spline interpolation, in the above case, the area range of the area to be interpolated needs to be selected as follows: the second latitude range: [lat_B_min-2Δlat_A, lat_B_max+2Δlat_A]; the second longitude range: [lon_B_min-2Δlon_A, lon_B_max+2Δlon_A].

[0039] This ensures that when using the bilinear interpolation algorithm and the cubic spline interpolation algorithm, there is sufficient surrounding grid point data to participate in the interpolation for the edge points of the target area, thereby ensuring the accuracy of the interpolation.

[0040] S120, interpolating the meteorological data of the corresponding area to be interpolated according to the second longitude and latitude resolution and each interpolation algorithm to be selected, and obtaining a selected interpolation result, a selected interpolation time and a selected interpolation error corresponding to each interpolation algorithm to be selected.

[0041] Among them, the second longitude and latitude resolution is the longitude and latitude resolution required in the target area, and the second longitude and latitude resolution is smaller than the first longitude and latitude resolution. It can be understood that the first longitude and latitude resolution is the longitude and latitude grid size of the low-resolution large-scale area, and the second longitude and latitude resolution is the longitude and latitude grid size of the high-resolution local area. The candidate interpolation result is the part of the meteorological data in the target area after interpolation, the candidate interpolation time is the time required for the interpolation calculation to obtain the candidate interpolation result, and the candidate interpolation error is used to measure the accuracy of the candidate interpolation algorithm when performing interpolation operations.

[0042] Specifically, for each candidate interpolation algorithm, the meteorological data of the corresponding area to be interpolated is interpolated using the candidate interpolation algorithm according to the second longitude and latitude resolution to obtain a candidate interpolation result corresponding to the candidate interpolation algorithm, and the time used for the interpolation process, i.e., the candidate interpolation time, is recorded, and then the candidate interpolation error corresponding to the candidate interpolation result is calculated and determined.

[0043] Based on the above example, the interpolation algorithms to be selected include a bilinear interpolation algorithm and a cubic spline interpolation algorithm. The meteorological data of the corresponding interpolation area can be interpolated according to the second longitude and latitude resolution and each interpolation algorithm to be selected in the following manner to obtain the interpolation result, interpolation time and interpolation error corresponding to each interpolation algorithm to be selected:

[0044] Using a bilinear interpolation algorithm, according to a second longitude and latitude resolution, interpolation processing is performed on the meteorological data of the to-be-interpolated area corresponding to the bilinear interpolation algorithm to obtain a first interpolation result and a first interpolation time;

[0045] Determine a first interpolation error according to the first interpolation result by using a maximum gradient error algorithm;

[0046] Using a cubic spline interpolation algorithm, according to a second longitude and latitude resolution, interpolation processing is performed on the meteorological data of the to-be-interpolated area corresponding to the cubic spline interpolation algorithm to obtain a second interpolation result and a second interpolation time;

[0047] According to the second interpolation result and the longitude and latitude information corresponding to the second interpolation result, a second interpolation error is determined through a preset model and K-fold cross validation.

[0048] Among them, the first interpolation result is the result of interpolation processing using a bilinear interpolation algorithm. The first interpolation time is the time required for interpolation processing using a bilinear interpolation algorithm. The first interpolation error is the error data of the first interpolation result, which is used to measure the interpolation accuracy of the bilinear interpolation algorithm. The second interpolation result is the result of interpolation processing using a cubic spline interpolation algorithm. The second interpolation time is the time required for interpolation processing using a cubic spline interpolation algorithm. The second interpolation error is the error data of the second interpolation result, which is used to measure the interpolation accuracy of the cubic spline interpolation algorithm. The longitude and latitude information corresponding to the second interpolation result is the longitude and latitude of each interpolation point in the second interpolation result. The preset model can be a mathematical model trained using the interpolation result of this meteorological data, which is used to calculate the model corresponding to the meteorological data based on the longitude and latitude information.

[0049] Specifically, by using a bilinear interpolation algorithm, the meteorological data of the area to be interpolated corresponding to the bilinear interpolation algorithm is interpolated according to the second longitude and latitude resolution, and a first interpolation result can be obtained, and the time required for the interpolation process is recorded, which is the first interpolation time. By using a maximum gradient error algorithm, the error of the first interpolation result is solved to obtain a first interpolation error. By using a cubic spline interpolation algorithm, the meteorological data of the area to be interpolated corresponding to the cubic spline interpolation algorithm is interpolated according to the second longitude and latitude resolution, and a second interpolation result can be obtained, and the time required for the interpolation process is recorded, which is the second interpolation time. Using the second interpolation result and the longitude and latitude information corresponding to the second interpolation result, it is divided into a test set and a training set multiple times and randomly, and the preset model is trained respectively and the corresponding errors are calculated. The second interpolation error can be obtained by combining these errors.

[0050] It is crucial to evaluate the accuracy of interpolation results, especially in the absence of accurate high-resolution data. Error estimation can effectively verify the accuracy of interpolation results and ensure the safety and reliability of interpolation data in practical applications. In meteorological fields such as weather forecasting, error estimation is necessary for numerical forecast model optimization and decision making. It can provide key information on forecast accuracy and help decision makers evaluate the credibility of results. Accurate error assessment helps prevent potential risks, especially when meteorological data is used for key decisions such as disaster prevention and mitigation, to avoid wrong decisions and safety problems caused by the use of unreliable data.

[0051] In the case of no accurate target data, the corresponding interpolation error is calculated according to the geometric features of the interpolation algorithm to be selected. The specific technical solution is: bilinear interpolation uses the maximum gradient error measurement method, and cubic spline interpolation uses the K-fold cross-validation method to measure the error size. Considering that the algorithm kernel of the bilinear interpolation algorithm uses linear functions to interpolate, when there is no target data comparison, the geometric features can be used to quantitatively calculate the size of the absolute error. For cubic spline interpolation, nonlinear polynomials are used to fit the interpolation, and the geometric features are not obvious. The K-fold cross-validation method that can achieve the same effect is selected to measure the absolute error. At this time, the error dimension is the same. Finally, the same formula can be used to calculate the average value to quantitatively compare the accuracy of each interpolation result.

[0052] Based on the above example, the first interpolation error may be determined according to the first interpolation result by using the maximum gradient error algorithm in the following manner:

[0053] Determining the maximum gradient of the interpolation point according to the meteorological data of each interpolation point in the first interpolation result;

[0054] For each interpolation point, the interpolation point error corresponding to the interpolation point is determined according to the maximum gradient of the interpolation point and the area of ​​the interpolation region corresponding to the interpolation point;

[0055] The average value of the errors at each interpolation point is taken as the first interpolation error.

[0056] Among them, the interpolation point is the data point added after the interpolation processing according to the second longitude and latitude resolution, and each interpolation point corresponds to an interpolated meteorological data. The maximum gradient is the maximum value of the gradient value of the meteorological data of the interpolation point in the direction of increasing longitude, decreasing longitude, increasing latitude and decreasing latitude. The area of ​​the interpolation region is the area of ​​the area corresponding to the interpolation point, the position points adjacent to the direction of increasing longitude and the direction of decreasing longitude, and the position points adjacent to the direction of increasing latitude and the direction of decreasing latitude. The interpolation point error is used to measure the accuracy of the interpolation point.

[0057] Specifically, for the meteorological data of each interpolation point in the first interpolation result, the gradient values ​​in the direction of increasing longitude, decreasing longitude, increasing latitude and decreasing latitude are calculated, and the maximum value of these gradient values ​​is used as the maximum gradient of the interpolation point. For each interpolation point, the area of ​​the area formed by the position points adjacent to the interpolation point in the direction of increasing longitude and decreasing longitude and the position points adjacent to the interpolation point in the direction of increasing latitude and decreasing latitude is determined as the interpolation area area, and the product of the maximum gradient of the interpolation point and the interpolation area area is used as the interpolation point error corresponding to the interpolation point. Then, the average value of the errors of each interpolation point is used as the first interpolation error.

[0058] It can be understood that the mathematical principle of using the maximum gradient and the area of ​​the region to estimate the first interpolation error of the first interpolation result corresponding to the bilinear interpolation algorithm is based on the rate of change (gradient) of the interpolation function (bilinear) and the size of the interpolation region. In bilinear interpolation, the error is approximately the product of the maximum gradient in the interpolation region and the area of ​​the interpolation region.

[0059] Assume that on a two-dimensional plane, there exists an interpolation function , whose gradient is expressed as , the interpolation point error corresponding to the i-th interpolation point The estimation formula is:

[0060]

[0061] in, is the maximum gradient of the interpolation point in the interpolation area, is the area of ​​the interpolation region. The error of this interpolation point is used as the absolute error. Since there is no accurate target data, the geometric characteristics of the interpolation region are used to estimate the error. The "maximum gradient error" is used as the error between the actual data value and the predicted data value, which is zero error under ideal conditions.

[0062] The average value of the interpolation point error of each interpolation point is used as the first interpolation error MAE1, based on which the quality of the first interpolation result can be estimated:

[0063]

[0064] in, is the interpolation point error of the i-th interpolation point, n is the number of interpolation points, is the first interpolation error.

[0065] Based on the above example, the second interpolation error can be determined according to the second interpolation result and the latitude and longitude information corresponding to the second interpolation result, through a preset model and K-fold cross validation in the following manner:

[0066] Initialize the current number of times, and randomly divide the meteorological data of each interpolation point in the second interpolation result and the longitude and latitude information corresponding to each interpolation point in the second interpolation result into a training set and a test set according to a preset ratio;

[0067] Train the preset model according to the training set to obtain the target model;

[0068] Input each test longitude and latitude in the test set into the target model to obtain the model output results corresponding to each test longitude and latitude;

[0069] Determine the current error based on the model output results corresponding to each test longitude and latitude and the test interpolation results;

[0070] In response to the current number being less than the preset number, the current number is updated, and the step of randomly dividing the second interpolation result and the longitude and latitude information corresponding to the second interpolation result into a training set and a test set according to a preset ratio is returned to be executed until the current number is equal to the preset number;

[0071] In response to the current number being equal to the preset number, an average value of the current errors is used as the second interpolation error.

[0072] Among them, the current number is the number of times the current error is solved. The training set includes each training longitude and latitude and the training interpolation result corresponding to each training longitude and latitude. The test set includes each test longitude and latitude and the test interpolation result corresponding to each test longitude and latitude. The training set and the test set can be divided according to a preset ratio, for example, the training set: data set is 8:2, etc. The target model is the model obtained after training with the training set, and the target models obtained after training with different training sets may be different. The model output result is the result obtained by processing the test longitude and latitude with the target model. The current error is the error value between the target model solution and the interpolation result corresponding to the current number.

[0073] Specifically, the current number is initialized, that is, the current number is set to 1. Then, the meteorological data of each interpolation point in the second interpolation result and the longitude and latitude information corresponding to each interpolation point in the second interpolation result are randomly divided into a training set and a test set according to a preset ratio. The preset model is trained using the training set to obtain a target model. Each test longitude and latitude in the test set is input into the target model, and the model calculation is performed on each test longitude and latitude through the target model to obtain the model output result corresponding to each test longitude and latitude. According to the model output result corresponding to each test longitude and latitude and the test interpolation result, the error is solved to obtain the current error corresponding to the current number. If the current number is less than the preset number, it means that the number of iterative solutions is insufficient and the current error needs to be solved. Therefore, the current number is updated, that is, the current number is increased by one, and the second interpolation result and the longitude and latitude information corresponding to the second interpolation result are randomly divided into a training set and a test set according to a preset ratio. The current error corresponding to the new current number is solved until the current number is equal to the preset number. If the current number is equal to the preset number, it means that the current number of errors is sufficient, so the average value of the errors calculated by the preset number of times is used as the second interpolation error. .

[0074] S130 , determining a target interpolation algorithm according to the candidate interpolation times and the candidate interpolation errors corresponding to the candidate interpolation algorithms, and using the candidate interpolation results corresponding to the target interpolation algorithm as the target interpolation results.

[0075] The target interpolation algorithm is an interpolation algorithm selected in combination with the interpolation rate and the interpolation accuracy for performing interpolation processing. The target interpolation result is the interpolation result obtained for the target area using the target interpolation algorithm.

[0076] Specifically, according to the preset interpolation requirements, such as accuracy priority, rate priority, both accuracy and rate priority, etc., the candidate interpolation times and candidate interpolation errors corresponding to the candidate interpolation algorithms are compared, and the target interpolation algorithm that meets the preset interpolation requirements is determined, and the candidate interpolation results corresponding to the target interpolation algorithm are used as the target interpolation results.

[0077] Based on the above example, the target interpolation algorithm can be determined according to the candidate interpolation time and the candidate interpolation error corresponding to each candidate interpolation algorithm in the following manner:

[0078] The candidate interpolation algorithm corresponding to the minimum value among the candidate interpolation times is used as the first target algorithm;

[0079] The interpolation algorithm to be selected corresponding to the minimum value among the interpolation errors to be selected is used as the second target algorithm;

[0080] In response to the first target algorithm being the same as the second target algorithm, using the first target algorithm as the target interpolation algorithm;

[0081] In response to the first target algorithm being different from the second target algorithm, a target interpolation algorithm is determined from the first target algorithm and the second target algorithm.

[0082] The first target algorithm is a candidate interpolation algorithm with the smallest candidate interpolation time, and the second target algorithm is a candidate interpolation algorithm with the smallest candidate interpolation error.

[0083] Specifically, according to the principle of rate priority, the candidate interpolation algorithm corresponding to the minimum value in the candidate interpolation time is used as the first target algorithm. According to the principle of accuracy priority, the candidate interpolation algorithm corresponding to the minimum value in the candidate interpolation error is used as the second target algorithm. If the first target algorithm is the same as the second target algorithm, it means that both accuracy and rate are optimal, so the first target algorithm is used as the target interpolation algorithm. If the first target algorithm is different from the second target algorithm, the target interpolation algorithm is selected from the first target algorithm and the second target algorithm according to user needs, or the relevant information of the first target algorithm and the second target algorithm (interpolation time and interpolation error) is displayed to the user, allowing the user to make a selection, and the user's selection is used as the target interpolation algorithm.

[0084] Based on the above example, the target interpolation algorithm can be determined from the first target algorithm and the second target algorithm in the following manner:

[0085] The interpolation result to be selected, the interpolation time to be selected, and the interpolation error to be selected corresponding to the first target algorithm are used as first candidate items, and the interpolation result to be selected, the interpolation time to be selected, and the interpolation error to be selected corresponding to the second target algorithm are used as second candidate items;

[0086] On the target terminal, displaying the first candidate item and the second candidate item;

[0087] Receive candidate item confirmation information fed back from the target terminal, determine the target candidate item according to the candidate item confirmation information, and use the candidate interpolation algorithm corresponding to the target candidate item as the target interpolation algorithm.

[0088] Among them, the first candidate is the interpolation result to be selected, the interpolation time to be selected, and the interpolation error to be selected corresponding to the first target algorithm. The second candidate is the interpolation result to be selected, the interpolation time to be selected, and the interpolation error to be selected corresponding to the second target algorithm. The target terminal is a terminal device provided to the user for viewing and use, and has display and receiving functions. The candidate confirmation information is the information for the user to select from the first candidate and the second candidate. The target candidate is the candidate carried in the candidate confirmation information, and is one of the first candidate and the second candidate.

[0089] Specifically, the interpolation result, interpolation time and interpolation error corresponding to the first target algorithm are used as the first candidate, and the interpolation result, interpolation time and interpolation error corresponding to the second target algorithm are used as the second candidate. On the target terminal, the first candidate and the second candidate are displayed to the user, so that the user can choose one that meets their needs. After the user makes the selection, the candidate confirmation information will be fed back. The candidate confirmation information fed back on the target terminal is received, and the candidate carried in the candidate confirmation information is parsed as the target candidate, and the interpolation algorithm corresponding to the target candidate is used as the target interpolation algorithm.

[0090] It should be noted that if the number of candidate interpolation algorithms exceeds two, the candidate interpolation result, candidate interpolation time and candidate interpolation error corresponding to each candidate interpolation algorithm can be used as candidate items; on the target terminal, each candidate item is displayed, and candidate item confirmation information fed back on the target terminal is received, and the target candidate item is determined based on the candidate item confirmation information, and the candidate interpolation algorithm corresponding to the target candidate item is used as the target interpolation algorithm.

[0091] It is understandable that the reason for selecting bilinear interpolation and cubic spline interpolation, which are often used in the meteorological field, as candidate interpolation algorithms for interpolation is that they can effectively reduce the contingency and uncertainty brought by a single interpolation method, and the comparison of interpolation results of different methods can improve the credibility of the interpolation results. In addition, meteorological data are usually nonlinear, complex and highly dependent on spatial location. Therefore, the distribution characteristics of different meteorological data are not the same. Therefore, different candidate interpolation algorithms may have different performances on the spatial distribution characteristics of data. From the algorithm level, bilinear interpolation may produce large errors in the boundary area because it only considers the linear relationship between adjacent data points. In contrast, cubic spline interpolation can generate smoother boundaries by considering a larger range of neighboring points, which can effectively reduce the occurrence of discontinuity or mutation of boundaries. In terms of accuracy and computational efficiency, bilinear interpolation is relatively simple and has a fast calculation speed, which is suitable for large-scale data sets and real-time application scenarios (such as weather forecasting). However, cubic spline interpolation has a larger amount of calculation and higher complexity, especially when processing large-scale meteorological data, it is likely to increase the calculation time. Therefore, it is very important to analyze the balance between the calculation speed and accuracy of the interpolation algorithm to be used for meteorological forecasting and analysis. In addition, for some application scenarios, such as short-term weather forecasting, it may be required to generate results quickly, while other applications, such as climate change research, require higher accuracy. Therefore, the target interpolation algorithm can be determined according to different requirements.

[0092] Taking the bilinear interpolation algorithm and the cubic spline interpolation algorithm as the candidate interpolation algorithms, by comparing the first interpolation time of the bilinear interpolation algorithm with the second interpolation time of the cubic spline interpolation algorithm, and comparing the first interpolation error of the bilinear interpolation algorithm with the second interpolation error of the cubic spline interpolation algorithm, to measure and judge whether the candidate interpolation method used is reasonable. Among them, the interpolation error is represented by MAE (Mean Absolute Error), and the specific calculation method has been described in the above example and will not be elaborated here. It is often used to evaluate performance, especially in cases where the incremental deviation is more sensitive. It can provide the deviation of the absolute deviation, and its unit is the same as the original data. The time for the bilinear interpolation algorithm and the cubic spline interpolation algorithm to perform the interpolation operation are respectively recorded as the first interpolation time T1 and the second interpolation time T2, and the mean absolute errors are respectively recorded as the first interpolation error e1 and the second interpolation error e2. If T1>T2 and e1>e2, it means that the accuracy and time efficiency of the bilinear interpolation algorithm are both better than those of the cubic spline interpolation algorithm, and it can be concluded that the first interpolation result corresponding to the bilinear interpolation algorithm is comprehensively better than the second interpolation result corresponding to the cubic spline interpolation algorithm. If T1<T2 and e1<e2, it means that the accuracy and time efficiency of the cubic spline interpolation algorithm are both better than those of the bilinear interpolation algorithm, and it can be concluded that the second interpolation result corresponding to the cubic spline interpolation algorithm is comprehensively better than the first interpolation result corresponding to the bilinear interpolation algorithm. In both of these cases, the optimal interpolation scheme, that is, the target interpolation algorithm, can be determined. If neither of the above sets of conditions is met, at this time, the optimal interpolation scheme cannot be directly determined and can be left to the user to choose, making a trade-off between accuracy and computational efficiency, and making a decision based on the specific application scenario (weather forecasting, weather analysis, etc.) or weighing the pros and cons of balancing accuracy and computational efficiency.

[0093] Based on the above example, after taking the candidate interpolation result corresponding to the target interpolation algorithm as the target interpolation result, the following method can be used for visual analysis, specifically:

[0094] Determine the type of analysis chart according to the meteorological data type;

[0095] For each type of analysis chart, determine the target analysis chart according to the target interpolation result;

[0096] On the target terminal, display each target analysis chart.

[0097] Among them, the meteorological data type includes temperature, humidity, etc., which are the types of meteorological data. The type of analysis chart is the type of visualization chart, such as a numerical distribution chart, a spatial distribution chart, a line chart, a bar chart, etc. combined with geographical information. The target analysis chart is a visualization chart drawn according to the target interpolation result in accordance with the type of analysis chart.

[0098] Specifically, by analyzing the meteorological data type, a visualization type that is convenient for reflecting the meteorological data type can be determined, that is, an analysis chart type is determined. For each analysis chart type, the target interpolation result is substituted for drawing to obtain a target analysis chart corresponding to the analysis chart type. Then, each target analysis chart is displayed on the target terminal so that the user can intuitively see the interpolated meteorological data in the target area.

[0099] It is understandable that after completing the interpolation work and judging the selection of appropriate algorithms and results, based on the purpose of making the interpolation effect clearer to the user and better interpreting the target interpolation results for the product, the obtained target interpolation results are visualized and analyzed, and the meteorological data corresponding to the target interpolation results are presented in the form of pictures and pushed to the path or web page specified by the user. Specifically, it can be to determine the input interpolated data file (target interpolation result), obtain user needs, and select the appropriate picture type (analysis chart type) in combination with the meteorological data type of the target interpolation result. The specific form can be set according to the general tool, and further output the picture (target analysis chart) for analysis. In order for subsequent users to be able to further analyze efficiently or use similar visualization analysis methods for other meteorological data, a corresponding program can be provided. After the data is input, the processing results appear in the form of pictures.

[0100] The above examples achieve dual evaluation and improvement of interpolation accuracy and computational efficiency, and provide a complete process for judging the quality of the results of each candidate interpolation method. The data slicing method is used to expand the interpolation area corresponding to the target area, which improves the accuracy of interpolation of local high-resolution data using wide-area low-resolution data based on supercomputers in big data scenarios. An error analysis method is proposed in the absence of accurate target data, which provides a scientific basis for decision makers in the meteorological field to judge the feasibility of interpolation results, so as to meet the needs of decision makers or scientists in the field of meteorological forecasting and meteorological analysis. Finally, a visualization solution is provided to facilitate product interpretation and further analysis, which has very important practical application value and market prospects.

[0101] The present invention has the following technical effects: by determining the to-be-interpolated area corresponding to each to-be-interpolated algorithm according to the regional scope of the target area, the first longitude and latitude resolution of the large-scale area and at least two to-be-interpolated algorithms, that is, selecting a suitable area from the large-scale area, so as to facilitate subsequent interpolation processing; interpolating the meteorological data of the corresponding to-be-interpolated area according to the second longitude and latitude resolution and each to-be-interpolated algorithm, obtaining the to-be-interpolated result, the to-be-interpolated time and the to-be-interpolated error corresponding to each to-be-interpolated algorithm, so as to facilitate subsequent analysis according to accuracy and rate; determining the target interpolation algorithm according to the to-be-interpolated time and the to-be-interpolated error corresponding to each to-be-interpolated algorithm, and taking the to-be-interpolated result corresponding to the target interpolation algorithm as the target interpolation result, obtaining the interpolation time and interpolation accuracy of the weighed interpolation algorithm, improving the interpolation effect in the absence of accurate target data, and realizing accurate and efficient mapping of low-resolution large-scale regional meteorological observation data to high-resolution data of a local simulation area.

[0102] Embodiment 2

[0103] Figure 2 Schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. Figure 2 As shown, the electronic device 200 includes one or more processors 201 and a memory 202 .

[0104] The processor 201 may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device 200 to perform desired functions.

[0105] The memory 202 may include one or more computer program products, and the computer program product may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, a random access memory (RAM) and / or a cache memory (cache), etc. The non-volatile memory may include, for example, a read-only memory (ROM), a hard disk, a flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 201 may run the program instructions to implement the meteorological data processing method of any embodiment of the present invention described above and / or other desired functions. Various contents such as initial external parameters, thresholds, etc. may also be stored in the computer-readable storage medium.

[0106] In one example, the electronic device 200 may further include: an input device 203 and an output device 204, which are interconnected via a bus system and / or other forms of connection mechanisms (not shown). The input device 203 may include, for example, a keyboard, a mouse, etc. The output device 204 may output various information to the outside, including early warning prompt information, braking force, etc. The output device 204 may include, for example, a display, a speaker, a printer, a communication network and a remote output device connected thereto, etc.

[0107] Of course, to simplify, Figure 2 Only some of the components related to the present invention in the electronic device 200 are shown, and components such as a bus, an input / output interface, etc. are omitted. In addition, according to specific application conditions, the electronic device 200 may also include any other appropriate components.

[0108] In addition to the above methods and devices, an embodiment of the present invention may also be a computer program product, which includes computer program instructions, which, when executed by a processor, enable the processor to execute the steps of the meteorological data processing method provided by any embodiment of the present invention.

[0109] The computer program product may be written in any combination of one or more programming languages ​​to write program code for performing the operations of the embodiments of the present invention, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0110] In addition, an embodiment of the present invention may also be a computer-readable storage medium having computer program instructions stored thereon, which, when executed by a processor, enables the processor to execute the steps of the meteorological data processing method provided by any embodiment of the present invention.

[0111] The computer readable storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can include, for example, but is not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0112] It should be noted that the terms used in the present invention are only for describing specific embodiments, rather than limiting the scope of the present application. As shown in the present specification, unless the context clearly indicates an exception, the words "one", "a", "a kind of" and / or "the" do not specifically refer to the singular, but may also include the plural. The terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that the process, method or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method or device. In the absence of further restrictions, the elements defined by the sentence "include one..." do not exclude the presence of other identical elements in the process, method or device including the elements.

[0113] It should also be noted that the terms "center", "up", "down", "left", "right", "vertical", "horizontal", "inside", "outside", etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. Unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", etc. should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be an indirect connection through an intermediate medium, or it can be a connection between the two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0114] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the technical solutions of the embodiments of the present invention.

Claims

1. A meteorological data processing method, characterized in that: include: Determine, according to the area range of the target area, the first latitude and longitude resolution of the large-scale area, and at least two interpolation algorithms to be selected, an area to be interpolated corresponding to each interpolation algorithm to be selected; According to the second longitude and latitude resolution and each candidate interpolation algorithm, interpolation processing is performed on the meteorological data of the corresponding region to be interpolated, so as to obtain a candidate interpolation result, a candidate interpolation time and a candidate interpolation error corresponding to each candidate interpolation algorithm; wherein the second longitude and latitude resolution is smaller than the first longitude and latitude resolution; According to the candidate interpolation times and the candidate interpolation errors corresponding to the candidate interpolation algorithms, a target interpolation algorithm is determined, and the candidate interpolation results corresponding to the target interpolation algorithm are used as the target interpolation results.

2. The method according to claim 1, characterized in that The step of determining the interpolation area corresponding to each interpolation algorithm according to the area range of the target area, the first latitude and longitude resolution of the large-scale area, and at least two interpolation algorithms to be selected includes: Determine, according to the area range of the target area, a first longitude upper limit, a first longitude lower limit, a first latitude upper limit, and a first latitude lower limit corresponding to the target area; For each candidate interpolation algorithm, determine a second longitude upper limit and a second longitude lower limit according to the first longitude upper limit, the first longitude lower limit, the first longitude resolution in the first longitude and longitude resolution, and the longitude expansion range corresponding to the candidate interpolation algorithm; Determine a second latitude upper limit and a second latitude lower limit according to the first latitude upper limit, the first latitude lower limit, the first latitude resolution in the first longitude and latitude resolution, and the latitude expansion range corresponding to the selected interpolation algorithm; The interpolation area corresponding to the interpolation algorithm is determined according to the second longitude upper limit, the second longitude lower limit, the second latitude upper limit, and the second latitude lower limit.

3. The method according to claim 1, characterized in that The interpolation algorithm to be selected includes a bilinear interpolation algorithm and a cubic spline interpolation algorithm; The interpolation process is performed on the meteorological data of the corresponding interpolation area according to the second longitude and latitude resolution and each interpolation algorithm to be selected, and the interpolation result, interpolation time and interpolation error corresponding to each interpolation algorithm to be selected are obtained, including: Using a bilinear interpolation algorithm, according to the second longitude and latitude resolution, interpolate the meteorological data of the to-be-interpolated area corresponding to the bilinear interpolation algorithm to obtain a first interpolation result and a first interpolation time; Determine a first interpolation error according to the first interpolation result by using a maximum gradient error algorithm; Using a cubic spline interpolation algorithm, according to the second longitude and latitude resolution, interpolate the meteorological data of the interpolation area corresponding to the cubic spline interpolation algorithm to obtain a second interpolation result and a second interpolation time; According to the second interpolation result and the longitude and latitude information corresponding to the second interpolation result, a second interpolation error is determined through a preset model and K-fold cross validation.

4. The method according to claim 3, characterized in that The determining the first interpolation error according to the first interpolation result by using a maximum gradient error algorithm includes: Determining the maximum gradient of each interpolation point according to the meteorological data of each interpolation point in the first interpolation result; For each interpolation point, determine the interpolation point error corresponding to the interpolation point according to the maximum gradient of the interpolation point and the area of ​​the interpolation region corresponding to the interpolation point; The average value of the errors at each interpolation point is taken as the first interpolation error.

5. The method according to claim 3, characterized in that: The determining, according to the second interpolation result and the longitude and latitude information corresponding to the second interpolation result, through a preset model and K-fold cross validation, a second interpolation error includes: Initialize the current number of times, and randomly divide the meteorological data of each interpolation point in the second interpolation result and the longitude and latitude information corresponding to each interpolation point in the second interpolation result into a training set and a test set according to a preset ratio; wherein the test set includes each test longitude and latitude and the test interpolation result corresponding to each test longitude and latitude; Training the preset model according to the training set to obtain a target model; Input each test longitude and latitude in the test set into the target model to obtain the model output result corresponding to each test longitude and latitude; Determine the current error based on the model output results corresponding to each test longitude and latitude and the test interpolation results; In response to the current number being less than the preset number, the current number is updated, and the step of randomly dividing the second interpolation result and the longitude and latitude information corresponding to the second interpolation result into a training set and a test set according to a preset ratio is returned to be executed until the current number is equal to the preset number; In response to the current number being equal to the preset number, an average value of the current errors is used as the second interpolation error.

6. The method according to claim 1, characterized in that The step of determining a target interpolation algorithm according to the interpolation time and interpolation error corresponding to each interpolation algorithm includes: The interpolation algorithm to be selected corresponding to the minimum value among the interpolation times to be selected is used as the first target algorithm; The interpolation algorithm to be selected corresponding to the minimum value among the interpolation errors to be selected is used as the second target algorithm; In response to the first target algorithm being the same as the second target algorithm, using the first target algorithm as a target interpolation algorithm; In response to the first target algorithm being different from the second target algorithm, a target interpolation algorithm is determined from the first target algorithm and the second target algorithm.

7. The method according to claim 6, characterized in that Determining a target interpolation algorithm from the first target algorithm and the second target algorithm includes: Taking the interpolation result to be selected, the interpolation time to be selected, and the interpolation error to be selected corresponding to the first target algorithm as the first candidate item, and taking the interpolation result to be selected, the interpolation time to be selected, and the interpolation error to be selected corresponding to the second target algorithm as the second candidate item; On the target terminal, displaying the first candidate item and the second candidate item; Receive candidate item confirmation information fed back from the target terminal, determine a target candidate item according to the candidate item confirmation information, and use the candidate interpolation algorithm corresponding to the target candidate item as the target interpolation algorithm.

8. The method according to claim 1, characterized in that After taking the candidate interpolation result corresponding to the target interpolation algorithm as the target interpolation result, the method further includes: Determine the type of analysis chart based on the type of meteorological data; For each analysis graph type, determining a target analysis graph according to the target interpolation result; On the target terminal, the analysis diagram of each target is displayed.

9. An electronic device, characterized in that: The electronic device comprises: Processor and memory; The processor is used to execute the steps of the meteorological data processing method according to any one of claims 1 to 8 by calling the program or instruction stored in the memory.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program or instruction, and the program or instruction enables a computer to execute the steps of the meteorological data processing method according to any one of claims 1 to 8.

Images