GIS-based meteorological data fusion platform and construction method thereof

Abstract

The invention discloses a meteorological data fusion platform based on a GIS (Geographic Information System). The meteorological data fusion platform comprises a meteorological service module, a GIS service module, a data fusion module and a system management module, the meteorological service module provides meteorological information query and information early warning query services for users; the GIS service module is used for providing position information and map browsing service for a user, so that the user can carry out map operation and space analysis; the data fusion module removes error values of the meteorological data, analyzes traditional meteorological data, completes spatial interpolation, and carries out early warning and prediction on future weather data based on fused data; and the system management module is used for user permission distribution and background data maintenance. According to the method, the basic map of the district and county-level region can be well overlaid, so that the weather forecast data of the district and county-level region can be accurately positioned, the weather service requirement is met to the greatest extent, and the optimal auxiliary decision-making basis is provided for weather predictors.

Description

Technical Field

[0001] The present invention belongs to the field of meteorology, and particularly relates to a GIS-based meteorological data fusion platform and a construction method thereof in this field. Background Technique

[0002] The concept of data fusion was proposed in the 1970s of the 20th century, but it did not attract enough attention at that time. With the development of the times and the progress of science and technology, people have begun to increasingly recognize the importance of data fusion. Data fusion is a multi-level and multi-layer processing process, which involves detecting, interconnecting, correlating, estimating, and synthesizing multi-source data and information to obtain refined position estimates and attribute estimates, as well as complete and timely situation assessments.

[0003] Focusing on the field of meteorology, the close connection between meteorological services and geographical data determines that meteorological services have a wide range of requirements for GIS systems. Among the currently used meteorological systems, the most widely used is the MICAPS system. Whether it is the national meteorological bureau, provincial meteorological bureau, or district and county meteorological bureau, they all use the unified version of the MICAPS system for meteorological forecasting work. The MICAPS system is a system for analyzing and visually expressing meteorological data on a global and national scale, and has good reference value for weather forecasting in large areas. However, due to the fact that it does not use GIS technology, the expression accuracy of the MICAPS system is affected, and it cannot well overlay the basic maps of district and county areas, so it cannot accurately locate the meteorological forecasting data of district and county areas, and finally will affect the decision-making judgment and meteorological forecasting of meteorological forecasters in district and county meteorological departments. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a GIS-based meteorological data fusion platform and a construction method thereof, which can not only implement an efficient data management mechanism, but also effectively reduce the redundant transmission of a large amount of data and improve the utilization rate of network transmission.

[0005] The present invention adopts the following technical solutions:

[0006] A GIS-based meteorological data fusion platform, the improvement lies in: including a meteorological service module, a GIS service module, a data fusion module, and a system management module; the meteorological service module provides users with meteorological information query and information warning query services; the GIS service module provides users with location information and map browsing services for users to perform map operations and spatial analysis; the data fusion module removes error values from meteorological data, analyzes traditional meteorological data, completes spatial interpolation, and performs early warning and prediction on future weather data based on the fused data; the system management module is used for user permission allocation and background data maintenance.

[0007] Furthermore, the formula for spatial interpolation is:

[0008] α′ = α0 + Δα ij

[0009]

[0010] In the above formula, α′ is the correction value of variable α at grid point (i, j), α is any meteorological element, α0 is the first guess value of variable α at grid point (i, j), and Δα k is the difference between the observed value and the first guess value at observation point k, and W ijk is the weight factor, and K is the number of stations within the influence radius R;

[0011]

[0012] In the above formula, d ijk is the distance from grid point (i, j) to observation point K.

[0013] Furthermore, the process of early warning and prediction is:

[0014] Each training sample set trains a sub-SVM, and the training sample set:

[0015] D = {(x i , y i ) │ i = 1, 2,.., l}, x i ∈E, y i ∈R

[0016] In the above formula, E is the Euclidean space and R is the real number space;

[0017] The regression function of the SVM regression algorithm is:

[0018]

[0019] In the above formula, is a non-linear mapping and b is the threshold;

[0020] Compare the prediction accuracies of the ensemble learning models obtained with different numbers of clusters C, increment the number of classes C one by one starting from 1, and select C when the ensemble accuracy is the highest * as the optimal number of classes, and the final ensemble output result is:

[0021]

[0022] Furthermore, users can also perform meteorological data statistics and map management in the data fusion module.

[0023] Furthermore, users can also perform visual expression of meteorological data and contour drawing in the system management module.

[0024] A method for constructing a meteorological data fusion platform based on GIS, the improvement lies in that it includes the following steps:

[0025] Step 1, GIS data preparation:

[0026] Set up a data provider to receive various data sent from the data source; classify various data according to the feature type, and create a layer for each type of feature; organize and package the data of each layer to form the map data required by the platform;

[0027] Step 2, Convert MICAPS data into GIS meteorological data;

[0028] Step 3, Visualize GIS meteorological data;

[0029] Step 4, Perform data fusion for meteorological prediction:

[0030] Define the relative distance between the collected data as d ij :

[0031] d ij =|Zi―Zj| i,j = 1,2,…n

[0032] In the above formula, Zi is the collected data, Zj is the real data, and n is the number of collected data;

[0033] Define a support function r ij :

[0034]

[0035] Establish a support matrix R:

[0036]

[0037] Find the weight coefficient of the i-th data Z i in the overall data

[0038]

[0039] In the above formula, v1, v2,..., v n are a set of non-negative numbers;

[0040] The fusion result of n collected data is:

[0041]

[0042] Further, in step 1, the data formats that the data provider can directly access include Autodesk MapGuide SDF format, Autodesk DWG format, Oracle Spatial format, and ESRI SHP format.

[0043] Further, in step 2, the MICAPS data includes surface all-element mapping data, upper-air all-element mapping data, general mapping and discrete point isopleth data, and grid data.

[0044] Further, the conversion method for the surface all-element mapping data is as follows:

[0045] Extract all the data in the text file and read it in the order of the data storage format in the text file; then create empty point data in the database; then add the longitude and latitude spatial coordinate pairs of the stations to the sequentially read data according to the station numbers, and then add all the meteorological element values corresponding to the stations in the form of attributes. Repeat this process until all the spatial coordinate pairs and attribute values of all stations are successfully added, update the point data, and finally save the point data in the spatial database.

[0046] Further, the conversion method for the grid data is as follows: Extract the spatial element coordinate pairs and attribute data according to the content and format of the data; then create empty point data, add the spatial element data and attribute data to the point data in sequence and perform update processing; then perform spatial interpolation on the point data according to a certain meteorological attribute value in the point data to generate line data; finally, save the generated line data in the database, delete the point data used for spatial interpolation, and finally update the database.

[0047] The beneficial effects of the present invention are as follows:

[0048] The platform disclosed by the present invention, based on GIS (Geographic Information System) technology, can well overlay the basic maps of district and county-level regions, thereby accurately positioning the meteorological forecast data of district and county-level regions, meeting the meteorological business needs to the greatest extent, and providing the best auxiliary decision-making basis for meteorological forecasters.

[0049] The method disclosed by the present invention has a simple process and is easy to implement, and can quickly build a GIS-based meteorological data fusion platform, so that regions with a small number of meteorological observation stations can quickly improve the accuracy of meteorological prediction and fully meet the needs of different industries and departments for meteorological information. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 is the block diagram of the platform disclosed by the present invention;

[0051] Figure 2 is the flow schematic diagram of the method of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0052] In order to make the objectives, technical solutions and advantages of the present invention more clear and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0053] Embodiment 1 discloses a meteorological data fusion platform based on GIS, as Figure 1 shown, which includes a meteorological service module, a GIS service module, a data fusion module and a system management module; the meteorological service module provides users with meteorological information query and information warning query services; the GIS service module provides users with location information and map browsing services for users to perform map operations and spatial analysis; the data fusion module removes error values from meteorological data, analyzes traditional meteorological data, completes spatial interpolation, and performs early warning prediction on future weather data based on the fused data; the system management module is used for user permission allocation and background data maintenance.

[0054] The formula for spatial interpolation is:

[0055] α′ = α0 + Δα ij

[0056]

[0057] In the above formula, α′ is the correction value of the variable α at the grid point (i, j), α is any meteorological element, α0 is the first guess value of the variable α at the grid point (i, j), and Δα k is the difference between the observed value and the first guess value at the observation point k, and W ijk is the weight factor, and K is the number of stations within the influence radius R;

[0058]

[0059] In the above formula, d ijk is the distance from the grid point (i, j) to the observation point K.

[0060] The process of early warning prediction is as follows:

[0061] Each training sample set trains a sub-SVM, and the training sample set:

[0062] D = {(x i , y i ) │ i = 1, 2,.., l}, x i ∈ E, y i ∈ R

[0063] In the above formula, E is the Euclidean space and R is the real number space;

[0064] The regression function of the SVM regression algorithm is as follows:

[0065]

[0066] In the above formula, is a non - linear mapping, and b is the threshold;

[0067] Compare the prediction accuracies of the ensemble learning models obtained according to different numbers of clusters C. Starting from 1, increment the number of classes C one by one, and select C* with the highest ensemble accuracy as the optimal number of classes. The final ensemble output result is:

[0068]

[0069] Users can also perform meteorological data statistics and map management in the data fusion module.

[0070] Meteorological data statistics: Meteorological data is numerous and miscellaneous. To enable users to have a macroscopic understanding, the system provides a data statistics function to present the changes in meteorological data over time in the form of charts.

[0071] Map management: The maps in the system adopt a hierarchical management method and are stored as map files. The map management module provides basic geographic information system functions, including functions such as map zoom - in, zoom - out, pan, full - map display, previous view, next view, attribute query, and geographic measurement. The layers support functions such as moving up, moving down, annotation, rendering method setting, and layer deletion. To achieve the aesthetic appearance of raster data display, the system provides functions such as raster data clipping.

[0072] In terms of thematic map production, the system provides a rich symbol library, including a weather symbol library and a variety of thematic map production tools, which facilitate users to add map decoration elements such as north arrows, scales, legends, and text boxes. Users can save the mapping scheme as a template for convenient reuse. Finally, the mapping products can be output with high quality by saving as pictures, printing, etc.

[0073] In addition, for some meteorological forecast products, it supports forward or backward selection in chronological order, that is, users can choose to view the map at the previous time or the map at the next time. The map can be saved as pictures in formats such as JPG and PNG, and the dynamically played thematic map can also be saved as a Gif image for data viewing and use. It can also directly print the currently visualized map content.

[0074] Users can also perform meteorological data visualization expression and contour line drawing in the system management module.

[0075] Meteorological data visualization expression:

[0076] The visual expression of meteorological element query results and the visual expression of radar data mainly adopt the mapping method of range-segmented thematic maps in GIS thematic mapping, and then overlay some basic geographic information data to form the effect of a stipple map.

[0077] The real-time display of meteorological elements and the weather change warning map mainly adopt the methods of label thematic maps and custom label thematic maps. On the basis of combining basic geographic information data, multiple label thematic maps of different elements are overlaid to achieve visual expression.

[0078] In addition, for the realization of the spatial distribution map, first, the table connection of geographic data point data is realized through query results or data analysis results, so that the query results become part of the attribute information of geographic data. Then, the Cressman interpolation method commonly used in meteorology is used to generate raster data from discrete points, and the basic geographic data and raster data are overlaid, combined with map decoration elements, to complete the mapping of the stipple map.

[0079] Contour line drawing:

[0080] First, find all open contour lines, and then find the closed contour lines. The specific method for open contour lines is as follows:

[0081] Find all the triangles that contain a certain value (assumed to be 996) on at least two sides and put them in a list.

[0082] Start from a triangle in the list that contains a border edge of a triangular grid. Determine whether there is a point with the value of 996 on this edge. If so, record the current triangle and the current edge, and add this point to the list of points that make up the contour line. Then start searching for "exits" on the remaining two sides of this triangle.

[0083] After finding the first "exit", add this point to the list of points that make up the contour line, and mark the value of the current triangle as -1, indicating that it has been temporarily used. Then, with the current edge as the edge where the "exit" point is located, make the current triangle the other triangle to which the current edge belongs.

[0084] After obtaining the current triangle and the current edge, repeat the above step in a loop. Another point on the contour line will be found. According to this method, there will be two possibilities in the end: a certain edge belongs to only one triangle (reaching the border of the triangular grid); the value of the triangle reached is -1, indicating that this triangle has been used in the previous steps. At this time, exit the loop.

[0085] At this time, the search for the open contour line with the value of 996 has been completed. The method for searching the closed contour line is similar to that of the open contour line. The only difference is that its exit condition is that the "exit" found and the coordinates of the first point of the contour line must be the same.

[0086] When all the search work is completed, if the isolines are legal, that is, the open isolines reach the border and the closed isolines reach the starting point, then change the status value of the candidate triangles, making them roll back from the temporarily used state to the used state and no longer participate in the next search. Otherwise, perform a "rollback" operation to make them roll back from the temporarily used state to the unused state.

[0087] This embodiment also discloses a method for constructing a meteorological data fusion platform based on GIS, as Figure 2 shown, including the following steps:

[0088] Step 1, GIS data preparation:

[0089] Set up a data provider to receive various data sent from the data source; classify the various data according to the feature type and create a layer for each type of feature; organize and package the data of each layer to form the map data required by the platform.

[0090] The data formats that the data provider can directly access include Autodesk MapGuide SDF format, Autodesk DWG format, Oracle Spatial format, and ESRI SHP format.

[0091] Step 2, convert MICAPS data into GIS meteorological data;

[0092] MICAPS data includes surface complete element mapping data, upper-air complete element mapping data, general mapping and discrete point isoline data, and grid data.

[0093] The conversion method for surface complete element mapping data is as follows:

[0094] Extract all the data in the text file and read it in the order of the data storage format in the text file; then create empty point data in the database; then add the longitude and latitude spatial coordinate pairs of the stations to the data read in order according to the station number, and then add all the meteorological element values corresponding to the stations in the form of attributes. Repeat this process until all the spatial coordinate pairs and attribute values of the stations are successfully added, update the point data, and finally save the point data in the spatial database.

[0095] The conversion method for grid data is as follows: Extract the spatial element coordinate pairs and attribute data according to the content and format of the data; then create empty point data, add the spatial element data and attribute data to the point data in sequence and perform update processing; then perform spatial interpolation on the point data according to a certain meteorological attribute value in the point data to generate line data; finally, save the generated line data in the database, delete the point data used for spatial interpolation, and finally update the database.

[0096] Step 3, visually represent GIS meteorological data;

[0097] Step 4, perform data fusion for meteorological prediction:

[0098] Define the relative distance between the collected data as d ij :

[0099] d ij = |Zi − Zj| i,j = 1,2,…n

[0100] In the above formula, Zi is the collected data, Zj is the real data, and n is the number of collected data;

[0101] Define a support function r ij :

[0102]

[0103] Establish a support matrix R:

[0104]

[0105] Find the weight coefficient of the i-th data Z i in the overall data

[0106]

[0107] In the above formula, v1, v2, …, v n are a set of non-negative numbers;

[0108] The fusion result of n collected data is:

[0109]

Claims

1. A GIS-based meteorological data fusion platform, characterized by: It includes a meteorological service module, a GIS service module, a data fusion module and a system management module; the meteorological service module provides users with meteorological information query and information warning query services; the GIS service module provides users with location information and map browsing services, so that users can perform map operations and spatial analysis; the data fusion module removes erroneous values ​​from meteorological data, analyzes traditional meteorological data, completes spatial interpolation, and warns and predicts future weather data based on the fused data; the system management module is used for user authority allocation and background data maintenance.

2. The GIS-based meteorological data fusion platform according to claim 1, characterized in that: The formula for spatial interpolation is: α′=α0+Δα ij In the above formula, α′ is the corrected value of variable α at grid point (i, j), α is any meteorological element, α0 is the first guess value of variable α at grid point (i, j), Δα k is the difference between the observed value at observation point k and the first guess, W ijk is the weight factor, K is the number of stations within the influence radius R; In the above formula, d ijk is the distance from the grid point (i, j) to the observation point K.

3. The GIS-based meteorological data fusion platform according to claim 1 is characterized by: The process of early warning and prediction is: Each training sample set trains a sub-SVM, training sample set: D={(x i ,y i )│i=1,2,..,l},x i ∈E,y i ∈R In the above formula, E is Euclidean space and R is real number space; The regression function of the SVM regression algorithm is: In the above formula, is a nonlinear mapping, b is the threshold; Compare the prediction accuracy of the ensemble learning model obtained with different cluster numbers C, increase the number of categories C one by one starting from 1, and select the C with the highest ensemble accuracy * As the optimal number of categories, the final integrated output result is:

4. The GIS-based meteorological data fusion platform according to claim 1 is characterized by: Users can also perform meteorological data statistics and map management in the data fusion module.

5. The GIS-based meteorological data fusion platform according to claim 1 is characterized by: Users can also visualize meteorological data and draw contour lines in the system management module.

6. A method for constructing a meteorological data fusion platform based on GIS, characterized in that: The steps include: Step 1, GIS data preparation: Set up a data provider to receive various data sent from the data source; classify various data according to feature types and create a layer for each feature; organize and package the data of each layer to form the map data required by the platform; Step 2, convert MICAPS data into GIS meteorological data; Step 3: Visualize GIS meteorological data; Step 4: Perform data fusion for weather forecasting: Define the relative distance between collected data as d ij : d ij =|Zi―Zj|i,j=1,2,…n In the above formula, Zi is the collected data, Zj is the real data, and n is the number of collected data; Define a support function r ij : Establish the support matrix R: Find the i-th data Z i Weight coefficient in the overall data In the above formula, v1, v2, ..., v n is a set of non-negative numbers; The fusion result of n collected data is:

7. The method for constructing a meteorological data fusion platform based on GIS according to claim 6, characterized in that: In step 1, the data formats that the data provider can directly access include Autodesk MapGuide SDF format, Autodesk DWG format, Oracle Spatial format, and ESRI SHP format.

8. The method for constructing a meteorological data fusion platform based on GIS according to claim 6, characterized in that: In step 2, MICAPS data includes ground full-factor mapping data, high-altitude full-factor mapping data, general mapping and discrete point contour data, and grid data.

9. The method for constructing a meteorological data fusion platform based on GIS according to claim 8, characterized in that: The conversion method of ground full-factor mapping data is as follows: Extract all the data in the text file and read them sequentially according to the storage format of the data in the text file; then create empty point data in the database; then add the latitude and longitude spatial coordinate pairs of the area station to the sequentially read data in the order of the area station number, and then add all the meteorological element values ​​corresponding to the area station in the form of attributes. Repeat this process until the spatial coordinate pairs and attribute values ​​of all area stations are successfully added, update the point data, and finally save the point data in the spatial database.

10. The method for constructing a meteorological data fusion platform based on GIS according to claim 8, characterized in that: The conversion method of grid data is as follows: extract spatial element coordinate pairs and attribute data according to the content and format of the data; then create empty point data, add spatial element data and attribute data to the point data in turn and update them; then perform spatial interpolation on the point data according to a certain meteorological attribute value in the point data to generate line data; finally, save the generated line data in the database, delete the point data used for spatial interpolation, and finally update the database.

Images