CN113593006A - Meteorological data spatial interpolation refining method and system based on deep learning - Google Patents
Meteorological data spatial interpolation refining method and system based on deep learning Download PDFInfo
- Publication number
- CN113593006A CN113593006A CN202110663352.5A CN202110663352A CN113593006A CN 113593006 A CN113593006 A CN 113593006A CN 202110663352 A CN202110663352 A CN 202110663352A CN 113593006 A CN113593006 A CN 113593006A
- Authority
- CN
- China
- Prior art keywords
- data
- meteorological
- spatial
- analyzed
- grid
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 97
- 238000013135 deep learning Methods 0.000 title claims abstract description 42
- 238000007670 refining Methods 0.000 title claims abstract description 28
- 238000009826 distribution Methods 0.000 claims abstract description 75
- 238000010586 diagram Methods 0.000 claims abstract description 74
- 230000006870 function Effects 0.000 claims abstract description 58
- 238000001556 precipitation Methods 0.000 claims abstract description 43
- 239000003086 colorant Substances 0.000 claims description 18
- 238000012549 training Methods 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 8
- 238000004422 calculation algorithm Methods 0.000 claims description 7
- 238000005457 optimization Methods 0.000 claims description 7
- 238000007476 Maximum Likelihood Methods 0.000 claims description 6
- 238000002790 cross-validation Methods 0.000 claims description 6
- 238000012800 visualization Methods 0.000 claims description 6
- 238000010276 construction Methods 0.000 claims description 4
- 238000004458 analytical method Methods 0.000 abstract description 11
- 238000007405 data analysis Methods 0.000 abstract description 6
- 230000000007 visual effect Effects 0.000 abstract description 6
- 238000004590 computer program Methods 0.000 description 9
- 238000011160 research Methods 0.000 description 3
- 230000007613 environmental effect Effects 0.000 description 2
- 230000001788 irregular Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 239000002689 soil Substances 0.000 description 1
- 238000009827 uniform distribution Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
- G06T3/4007—Scaling of whole images or parts thereof, e.g. expanding or contracting based on interpolation, e.g. bilinear interpolation
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Graphics (AREA)
- Geometry (AREA)
- Software Systems (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a meteorological data spatial interpolation refining method and a meteorological data spatial interpolation refining system based on deep learning, wherein the method comprises the following steps: determining an area to be analyzed to obtain spatial data and meteorological data of a meteorological station, generating a grid point diagram to be interpolated, performing spatial interpolation by adopting a thin plate spline function method to obtain network surface data corresponding to the area to be analyzed, and generating a precipitation spatial distribution diagram and an air temperature spatial distribution diagram. According to the method, a grid point diagram to be interpolated is constructed based on spatial data of a region to be analyzed, refined spatial interpolation is carried out on grid points to be interpolated by adopting a thin plate spline function method based on meteorological data, each meteorological station is fitted to obtain network surface data which are suitable for real factors such as time, terrain, climate change and meteorological station distribution, a visual precipitation spatial distribution diagram and an air temperature spatial distribution diagram are constructed according to the network surface data, analysis errors are effectively reduced, spatial interpolation precision is improved, and grid data analysis efficiency is improved.
Description
Technical Field
The invention relates to the technical field of meteorological monitoring and computer graphics, in particular to a meteorological data spatial interpolation refining method and system based on deep learning.
Background
Meteorological elements are important parameters in the relevant research fields such as ecology, environmentality, soil science, and climate model research, as environmental factors. Theoretically, the spatial distribution of meteorological elements can be acquired by arranging high-density meteorological stations to form an observation station network; however, the meteorological stations cannot achieve completely uniform distribution in space, and the actual distribution is not uniform, so that the observation sequences are not uniform in length, and there are many uncertainties when a standard gridding model is directly applied. Therefore, in the research process, spatial distribution interpolation calculation is often performed on limited meteorological observation stations, and the point data observed by each meteorological observation station is utilized to assimilate the surface data of the observation station network, so that the operation efficiency of the gridding model is improved, and the refined display of the meteorological forecast result is facilitated.
Currently, common spatial Interpolation methods include Inverse Distance Weighted Interpolation (IDW), Local Polynomial Interpolation (LPI), Ordinary Kriging Interpolation (OK), Thin Plate Spline (TPS), and the like, and because the change of meteorological elements is complex, when the same meteorological observation data is processed, the processing results of the spatial Interpolation methods are greatly different; different optimal spatial interpolation methods exist for different time points or different environmental regions, and the existing spatial interpolation method and strategy cannot select the current optimal spatial interpolation method according to factors such as terrain, climate change characteristics, meteorological observation station point distribution and the like, so that the data analysis efficiency is low when gridding expression of meteorological observation data is carried out, and larger errors exist.
Disclosure of Invention
The embodiment of the invention discloses a meteorological data spatial interpolation refining method and system based on deep learning.
The embodiment of the invention discloses a meteorological data spatial interpolation refining method based on deep learning in a first aspect, which comprises the following steps:
determining boundary data of a region to be analyzed;
acquiring spatial data and meteorological data corresponding to a meteorological site in the area to be analyzed;
generating a grid point diagram to be interpolated based on the spatial data;
based on the meteorological data, carrying out spatial interpolation on the grid point diagram to be interpolated by adopting a thin plate spline function method to obtain net surface data corresponding to the area to be analyzed;
and generating a precipitation space distribution diagram and an air temperature space distribution diagram corresponding to the area to be analyzed based on the net surface data.
Preferably, the generating a grid point map to be interpolated based on the spatial data includes:
determining a precision index of spatial interpolation based on the spatial data, wherein the precision index comprises the density and the resolution of the grid points;
generating an initial grid point diagram to be interpolated based on the precision index, wherein grid points of the initial grid point diagram to be interpolated are uniformly distributed, grid intervals are equal, and the plane coordinate of each grid point is (x)i，yi) I is 1, 2, 3, … …, n, wherein n is a grid number;
and eliminating redundant grid points which exist outside the area to be analyzed in the initial grid point diagram to be interpolated to obtain the grid point diagram to be interpolated corresponding to the area to be analyzed.
Preferably, the performing spatial interpolation on the grid point diagram to be interpolated by using a thin-plate spline function method based on the meteorological data to obtain the mesh surface data corresponding to the area to be analyzed includes:
based on each said meteorological site pairCorresponding meteorological data and plane coordinates (x)j，yj) J is 1, 2, 3, … …, m, where m is the weather station number, and the weather data z (x, y) for each grid point is derived from the following formula:
wherein | represents the euclidean norm, ciAs a function of the number of the coefficients,
is the kernel function of the thin plate spline method, and has the following values:
ri＝(x-xi)2+(y-yi)2。
preferably, a plane is constructed by adopting a thin plate spline function method to fit each meteorological station, and based on a spline formed by combining a plurality of meteorological stations, a smooth plane which approximates to a control point corresponding to each meteorological station and has the minimum curvature is obtained through optimization and is used as the net surface data, wherein the smooth parameter of the smooth plane is determined by the minimization of a generalized cross validation method or the minimization of a generalized maximum likelihood;
assuming that t control points corresponding to meteorological sites are distributed in the area to be analyzed to form a known point set piAnd i is 1, 2, 3, … …, t, the coordinate of each control point is (x)i，yi，z(xi，yi) And when z (x)i，yi) With a second continuous derivative, the energy function is as follows:
and then determining the minimization of the energy function by adopting a thin plate spline function method:
Ztps＝argminE。
preferably, the generating a precipitation amount spatial distribution map and an air temperature spatial distribution map corresponding to the region to be analyzed based on the mesh surface data includes:
constructing a legend generation model by adopting a deep learning algorithm based on a geographic information system and a national weather industry standard;
training the legend generation model by a preset color gamut range and a historical meteorological image;
processing the net surface data by adopting the legend generation model to obtain the precipitation space distribution map and the air temperature space distribution map;
the precipitation space distribution map adopts different colors in the color gamut range to represent precipitation numerical values of each area; and the air temperature space distribution map represents the air temperature value of each region by adopting different colors in the color gamut range.
The second aspect of the embodiment of the invention discloses a system for refining meteorological data spatial interpolation based on deep learning, which comprises:
a boundary determining unit for determining boundary data of the region to be analyzed;
the data acquisition unit is used for acquiring spatial data and meteorological data corresponding to meteorological sites in the area to be analyzed;
the grid generating unit is used for generating a grid point diagram to be interpolated based on the spatial data;
the spatial interpolation unit is used for carrying out spatial interpolation on the grid point diagram to be interpolated by adopting a thin plate spline function method based on the meteorological data to obtain the mesh surface data corresponding to the area to be analyzed;
and the visualization unit is used for generating a precipitation space distribution map and an air temperature space distribution map corresponding to the area to be analyzed based on the network surface data.
Preferably, the mesh generation unit includes:
a precision determining subunit, configured to determine a precision index of spatial interpolation based on the spatial data, where the precision index includes density and resolution of a grid point;
an initial generating subunit, configured to generate an initial grid point map to be interpolated based on the accuracy index, where grid points of the initial grid point map to be interpolated are uniformly distributed, grid intervals are equal, and a plane coordinate of each grid point is (x)i，yi) I is 1, 2, 3, … …, n, wherein n is a grid number;
and the redundant eliminating subunit is used for eliminating redundant grid points which exist outside the area to be analyzed in the initial grid point diagram to be interpolated to obtain the grid point diagram to be interpolated corresponding to the area to be analyzed.
Preferably, the spatial interpolation unit specifically includes:
based on the meteorological data and the plane coordinate (x) corresponding to each meteorological sitej，yj) J is 1, 2, 3, … …, m, where m is the weather station number, and the weather data z (x, y) for each grid point is derived from the following formula:
wherein | represents the euclidean norm, ciAs a function of the number of the coefficients,
is the kernel function of the thin plate spline method, and has the following values:
ri＝(x-xi)2+(y-yi)2。
preferably, a plane is constructed by adopting a thin plate spline function method to fit each meteorological station, and based on a spline formed by combining a plurality of meteorological stations, a smooth plane which approximates to a control point corresponding to each meteorological station and has the minimum curvature is obtained through optimization and is used as the net surface data, wherein the smooth parameter of the smooth plane is determined by the minimization of a generalized cross validation method or the minimization of a generalized maximum likelihood;
assuming that t control points corresponding to meteorological sites are distributed in the area to be analyzed to form a known point set piAnd i is 1, 2, 3, … …, t, the coordinate of each control point is (x)i，yi，z(xi，yi) And when z (x)i，yi) With a second continuous derivative, the energy function is as follows:
and then determining the minimization of the energy function by adopting a thin plate spline function method:
Ztps＝argminE。
preferably, the visualization unit includes:
the model construction subunit is used for constructing a legend generation model by adopting a deep learning algorithm based on the geographic information system and the national weather industry standard;
the training subunit is used for training the legend generation model by presetting a color gamut range and historical meteorological images;
a graph generating subunit, configured to process the mesh surface data by using the legend generation model to obtain the precipitation spatial distribution map and the air temperature spatial distribution map;
the precipitation space distribution map adopts different colors in the color gamut range to represent precipitation numerical values of each area; and the air temperature space distribution map represents the air temperature value of each region by adopting different colors in the color gamut range.
The third aspect of the embodiments of the present invention discloses a system for spatial interpolation refinement of meteorological data based on deep learning, including:
a memory storing executable program code;
a processor coupled with the memory;
the processor calls the executable program code stored in the memory to execute the meteorological data spatial interpolation refining method based on deep learning disclosed by the first aspect of the embodiment of the invention.
A fourth aspect of the embodiments of the present invention discloses a computer-readable storage medium storing a computer program, where the computer program enables a computer to execute the method for refining spatial interpolation of meteorological data based on deep learning disclosed in the first aspect of the embodiments of the present invention.
A fifth aspect of embodiments of the present invention discloses a computer program product, which, when run on a computer, causes the computer to perform some or all of the steps of any one of the methods of the first aspect.
A sixth aspect of the present embodiment discloses an application publishing platform, where the application publishing platform is configured to publish a computer program product, where the computer program product is configured to, when running on a computer, cause the computer to perform part or all of the steps of any one of the methods in the first aspect.
Compared with the prior art, the embodiment of the invention has the following beneficial effects:
a grid point diagram to be interpolated is constructed based on spatial data of a region to be analyzed, refined spatial interpolation is carried out on grid points to be interpolated by adopting a thin plate spline function method based on meteorological data, each meteorological station is fitted to obtain grid surface data adapted to real factors such as time, terrain, climate change and meteorological station distribution, a visual precipitation spatial distribution diagram and an air temperature spatial distribution diagram are constructed according to the grid surface data, analysis errors are effectively reduced, spatial interpolation precision is improved, and grid data analysis efficiency is improved.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to these drawings without creative efforts.
FIG. 1 is a schematic flow chart of a method for refining spatial interpolation of meteorological data based on deep learning according to an embodiment of the present invention;
FIG. 2 is a schematic representation diagram of a thin balcony function method in a meteorological data spatial interpolation refinement method based on deep learning according to an embodiment of the present invention;
FIG. 3 is a precipitation spatial distribution diagram and an air temperature spatial distribution diagram at a time of 2019-01-01 in a Meteorological data spatial interpolation refining method based on deep learning, which is disclosed by the embodiment of the invention;
FIG. 4 is a schematic structural diagram of a system for spatial interpolation refinement of meteorological data based on deep learning according to an embodiment of the present invention;
FIG. 5 is a schematic structural diagram of another system for refining spatial interpolation of meteorological data based on deep learning according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
It should be noted that the terms "first", "second", "third", "fourth", and the like in the description and the claims of the present invention are used for distinguishing different objects, and are not used for describing a specific order. The terms "comprises," "comprising," and any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
The embodiment of the invention discloses a meteorological data spatial interpolation refining method and system based on deep learning, wherein a grid point diagram to be interpolated is constructed based on spatial data of a region to be analyzed, then refined spatial interpolation is carried out on the grid point to be interpolated by adopting a thin plate spline function method based on the meteorological data, each meteorological station is fitted to obtain network surface data which are suitable for real factors such as time, terrain, climate change, meteorological station distribution and the like, and a visual precipitation spatial distribution diagram and air temperature spatial distribution diagram are constructed according to the network surface data, so that analysis errors are effectively reduced, the spatial interpolation precision is improved, and the gridding data analysis efficiency is improved.
Example one
Referring to fig. 1, fig. 1 is a schematic flowchart illustrating a method for refining spatial interpolation of meteorological data based on deep learning according to an embodiment of the present invention. As shown in FIG. 1, the method for refining the spatial interpolation of the meteorological data based on the deep learning can comprise the following steps.
101. Boundary data of the region to be analyzed is determined.
In the embodiment of the invention, for the area to be analyzed to be subjected to spatial interpolation, the boundary data such as the basin boundary, the administrative division boundary, the administrative center, the basin relation province and the like are determined by adopting a geographic information system, and the area to be analyzed is determined on the data level.
102. And acquiring spatial data and meteorological data corresponding to the meteorological station in the area to be analyzed.
In the embodiment of the invention, the Chinese national weather data center extracts the spatial data of weather stations in the area to be analyzed, such as longitude and latitude coordinates, altitude and the like, and weather data of precipitation, air temperature and the like monitored by each weather station.
103. And generating a grid point diagram to be interpolated based on the spatial data.
In the embodiment of the invention, the area to be analyzed is subjected to initial gridding.
As an alternative implementation, the spatial data-based determinationDetermining the precision index of spatial interpolation, wherein the precision index comprises the density and the resolution of a grid point; generating an initial grid point diagram to be interpolated based on the precision index, wherein grid points of the initial grid point diagram to be interpolated are uniformly distributed, grid intervals are equal, and the plane coordinate of each grid point is (x)i，yi) I is 1, 2, 3, … …, n, wherein n is a grid number; and eliminating redundant grid points existing outside the area to be analyzed in the initial grid point diagram to be interpolated to obtain a grid point diagram to be interpolated corresponding to the area to be analyzed. Specifically, the higher the density and the resolution of the grid points, the more refined the spatial interpolation is, the more accurate the analysis result is, but at the same time, the data volume involved in the processing process will also increase sharply, so that a suitable accuracy index is determined according to the actual analysis requirement, and accordingly, an initial grid point diagram to be interpolated is generated in the area to be analyzed, wherein the grid points are distributed uniformly and the grid intervals are equal; because the geographical boundary of the region to be analyzed is usually an irregular curve, redundant grid points are generated outside the geographical boundary in the grid point generation process, so that grid points outside the geographical boundary are removed based on the boundary data, only the grid points in the region to be analyzed are reserved, and a grid point diagram to be interpolated is obtained.
104. And based on the meteorological data, carrying out spatial interpolation on the grid point diagram to be interpolated by adopting a thin plate spline function method to obtain the mesh surface data corresponding to the area to be analyzed.
In the embodiment of the invention, the spatial interpolation is refined based on a thin plate spline function method.
As an optional implementation mode, the method is based on the meteorological data and the plane coordinate (x) corresponding to each meteorological sitej，yi) J is 1, 2, 3, … …, m, where m is the weather station number, and the weather data z (x, y) for each grid point is derived from the following formula:
wherein | | x | represents the euclidean norm, ciAs a function of the number of the coefficients,
is the kernel function of the thin plate spline method, and has the following values:
ri＝(x-xi)2+(y-yi)2。
referring to FIG. 2, a plane is constructed by adopting a thin plate spline function method to fit each meteorological site, and based on a spline formed by combining a plurality of meteorological sites, a smooth plane which approximates to a control point corresponding to each meteorological site and has the minimum curvature is obtained through optimization and is used as network surface data, wherein the smooth parameter of the smooth plane is determined by the minimization of a generalized cross validation method or the minimization of a generalized maximum likelihood;
suppose that t control points corresponding to meteorological sites are distributed in an area to be analyzed to form a known point set piAnd i is 1, 2, 3, … …, t, the coordinate of each control point is (x)i，yi，z(xi，yi) And when z (x)i，yi) With a second continuous derivative, the energy function is as follows:
and then determining the minimization of the energy function by adopting a thin plate spline function method:
Ztps＝argminE。
in fig. 2, a point p is a corresponding position of the meteorological station in the grid point diagram to be interpolated, a point q is an actual position of the meteorological station after optimization fitting by adopting a thin plate spline function method, and finally, plane and line distorted network surface data are obtained through processing.
Therefore, under the fitting of the thin plate spline function method, the position of each meteorological station on the net surface is adjusted, the net surface data are matched with the actual conditions of terrain, climate change, meteorological station distribution and the like in the area to be analyzed, most errors are eliminated, and the net surface data are clear and accurate.
105. And generating a precipitation space distribution diagram and an air temperature space distribution diagram corresponding to the area to be analyzed based on the net surface data.
In the embodiment of the invention, the network surface data is marked for visual display.
As an optional implementation, a deep learning algorithm is adopted to construct a legend generation model based on a geographic information system and national weather industry standards; training a legend generation model by using a preset color gamut range and a historical meteorological image; processing the network surface data by adopting a legend generation model to obtain a precipitation space distribution map and an air temperature space distribution map; the precipitation space distribution map adopts different colors in a color gamut range to represent precipitation numerical values of each area; and the air temperature space distribution map represents the air temperature value of each region by adopting different colors in the color gamut range. Specifically, referring to fig. 3, taking the precipitation amount spatial distribution diagram as an example, the rainfall values of 0 to 250mm are respectively represented by the gradual colors from white to blue, so that different areas on the precipitation amount spatial distribution diagram are marked by different colors, the precipitation amount distribution condition and the rainfall values of the areas can be visually known, and the meteorological analysis efficiency is improved.
In conclusion, a grid point diagram to be interpolated is constructed based on the spatial data of the region to be analyzed, refined spatial interpolation is performed on the grid points to be interpolated by adopting a thin plate spline function method based on meteorological data, and each meteorological station is fitted to obtain the grid surface data which is suitable for the real factors such as time, terrain, climate change, meteorological station distribution and the like, and a visual precipitation spatial distribution diagram and an air temperature spatial distribution diagram are constructed according to the grid surface data, so that the analysis error is effectively reduced, the spatial interpolation precision is improved, and the grid data analysis efficiency is improved.
Example two
Referring to fig. 2, fig. 3 and fig. 4, fig. 4 is a schematic structural diagram of a system for refining spatial interpolation of meteorological data based on deep learning according to an embodiment of the present invention. As shown in FIG. 4, the system for refining the spatial interpolation of meteorological data based on deep learning can comprise the following contents.
A boundary determining unit 401, configured to determine boundary data of the region to be analyzed;
a data obtaining unit 402, configured to obtain spatial data and meteorological data corresponding to a meteorological site in an area to be analyzed;
a grid generating unit 403, configured to generate a grid point diagram to be interpolated based on the spatial data;
the spatial interpolation unit 404 is configured to perform spatial interpolation on the grid point diagram to be interpolated by using a thin-plate spline function method based on the meteorological data to obtain mesh surface data corresponding to the area to be analyzed;
and a visualization unit 405 for generating a precipitation volume spatial distribution map and an air temperature spatial distribution map corresponding to the region to be analyzed based on the mesh surface data.
The grid generating unit 403 includes:
a precision determining subunit 4031, configured to determine a precision index of spatial interpolation based on the spatial data, where the precision index includes density and resolution of a grid point;
an initial generating subunit 4032, configured to generate an initial grid point map to be interpolated based on the precision index, where grid points of the initial grid point map to be interpolated are uniformly distributed, grid distances are equal, and a plane coordinate of each grid point is (x)i，yi) I is 1, 2, 3, … …, n, wherein n is a grid number;
and the redundant eliminating subunit 4033 is configured to eliminate redundant grid points existing outside the area to be analyzed in the initial grid point map to be interpolated to obtain the grid point map to be interpolated corresponding to the area to be analyzed.
And, the spatial interpolation unit 404 specifically includes:
based on meteorological data and plane coordinates (x) corresponding to each meteorological sitej，yj) J is 1, 2, 3, … …, m, where m is the weather station number, and the weather data z (x, y) for each grid point is derived from the following formula:
wherein | | x | represents the euclidean norm, ciAs a function of the number of the coefficients,
is the kernel function of the thin plate spline method, and has the following values:
ri＝(x-xi)2+(y-yi)2。
further, the visualization unit 405 includes:
the model construction subunit 4051 is used for constructing a legend generation model by adopting a deep learning algorithm based on the geographic information system and the national weather industry standard;
a training subunit 4052, configured to train the legend generation model with a preset color gamut range and a historical meteorological image;
a graph generation subunit 4053, configured to process the mesh data using the legend generation model to obtain a precipitation spatial distribution map and an air temperature spatial distribution map;
the precipitation space distribution map adopts different colors in a color gamut range to represent precipitation numerical values of each area; and the air temperature space distribution map represents the air temperature value of each region by adopting different colors in the color gamut range.
As an alternative embodiment, the precision determination subunit 4031 determines a precision index of spatial interpolation based on the spatial data, where the precision index includes density and resolution of a grid point; the initial generation subunit 4032 generates an initial grid point map to be interpolated based on the precision index, the grid points of the initial grid point map to be interpolated are uniformly distributed, the grid intervals are equal, and the plane coordinate of each grid point is (x)i，yi) I is 1, 2, 3, … …, n, wherein n is a grid number; the redundancy elimination subunit 4033 eliminates the redundant grid points existing outside the area to be analyzed in the initial grid point diagram to be interpolatedAnd dividing to obtain a grid point diagram to be interpolated corresponding to the area to be analyzed. Specifically, the higher the density and the resolution of the grid points, the more refined the spatial interpolation is, the more accurate the analysis result is, but at the same time, the data volume involved in the processing process will also increase sharply, so that a suitable accuracy index is determined according to the actual analysis requirement, and accordingly, an initial grid point diagram to be interpolated is generated in the area to be analyzed, wherein the grid points are distributed uniformly and the grid intervals are equal; because the geographical boundary of the region to be analyzed is usually an irregular curve, redundant grid points are generated outside the geographical boundary in the grid point generation process, so that grid points outside the geographical boundary are removed based on the boundary data, only the grid points in the region to be analyzed are reserved, and a grid point diagram to be interpolated is obtained.
As an optional implementation mode, the method is based on the meteorological data and the plane coordinate (x) corresponding to each meteorological sitej，yj) Where j is 1, 2, 3, … …, m, where m is the weather station number, the spatial interpolation unit 404 derives the weather data z (x, y) for each grid point from the following equation:
wherein | | x | represents the euclidean norm, ciAs a function of the number of the coefficients,
is the kernel function of the thin plate spline method, and has the following values:
ri＝(x-xi)2+(y-yi)2。
referring to FIG. 2, a plane is constructed by adopting a thin plate spline function method to fit each meteorological site, and based on a spline formed by combining a plurality of meteorological sites, a smooth plane which approximates to a control point corresponding to each meteorological site and has the minimum curvature is obtained through optimization and is used as network surface data, wherein the smooth parameter of the smooth plane is determined by the minimization of a generalized cross validation method or the minimization of a generalized maximum likelihood;
suppose that t control points corresponding to meteorological sites are distributed in an area to be analyzed to form a known point set piAnd i is 1, 2, 3, … …, t, the coordinate of each control point is (x)i，yi，z(xi，yi) And when z (x)i，yi) With a second continuous derivative, the energy function is as follows:
and then determining the minimization of the energy function by adopting a thin plate spline function method:
Ztps＝argminE。
in fig. 2, the point p is the corresponding position of the meteorological station in the grid point diagram to be interpolated, the point q is the actual position of the meteorological station after the optimal fitting by adopting the thin plate spline function method, and finally, the net surface data with distorted planes and lines is obtained by processing
Therefore, under the fitting of the thin plate spline function method, the position of each meteorological station on the net surface is adjusted, the net surface data are matched with the actual conditions of terrain, climate change, meteorological station distribution and the like in the area to be analyzed, most errors are eliminated, and the net surface data are clear and accurate.
As an optional implementation manner, the model construction subunit 4051 constructs a legend generation model by using a deep learning algorithm based on geographic information systems and national weather industry standards; the training subunit 4052 trains the legend generation model with a preset color gamut range and a historical meteorological image; the graph generation subunit 4053 processes the mesh data by using the graph generation model to obtain a precipitation spatial distribution map and an air temperature spatial distribution map; the precipitation space distribution map adopts different colors in a color gamut range to represent precipitation numerical values of each area; and the air temperature space distribution map represents the air temperature value of each region by adopting different colors in the color gamut range. Specifically, referring to fig. 3, taking the precipitation amount spatial distribution diagram as an example, the rainfall values of 0 to 250mm are respectively represented by the gradual colors from white to blue, so that different areas on the precipitation amount spatial distribution diagram are marked by different colors, the precipitation amount distribution condition and the rainfall values of the areas can be visually known, and the meteorological analysis efficiency is improved.
In conclusion, a grid point diagram to be interpolated is constructed based on the spatial data of the region to be analyzed, refined spatial interpolation is performed on the grid points to be interpolated by adopting a thin plate spline function method based on meteorological data, and each meteorological station is fitted to obtain the grid surface data which is suitable for the real factors such as time, terrain, climate change, meteorological station distribution and the like, and a visual precipitation spatial distribution diagram and an air temperature spatial distribution diagram are constructed according to the grid surface data, so that the analysis error is effectively reduced, the spatial interpolation precision is improved, and the grid data analysis efficiency is improved.
EXAMPLE III
Referring to fig. 5, fig. 5 is a schematic structural diagram of another system for refining spatial interpolation of meteorological data based on deep learning according to an embodiment of the present invention. As shown in fig. 5, the system for refining the spatial interpolation of meteorological data based on deep learning may include:
a memory 501 in which executable program code is stored;
a processor 502 coupled to a memory 501;
the processor 502 calls the executable program code stored in the memory 501 to execute the method for refining the spatial interpolation of the meteorological data based on the deep learning of fig. 1.
The embodiment of the invention discloses a computer-readable storage medium which stores a computer program, wherein the computer program enables a computer to execute the meteorological data spatial interpolation refining method based on deep learning in the figure 1.
Embodiments of the present invention also disclose a computer program product, wherein, when the computer program product is run on a computer, the computer is caused to execute part or all of the steps of the method as in the above method embodiments.
It will be understood by those skilled in the art that all or part of the steps in the methods of the embodiments described above may be implemented by program instructions associated with hardware, and the program may be stored in a computer-readable storage medium, which includes Read-Only Memory (ROM), Random Access Memory (RAM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), One-time Programmable Read-Only Memory (OTPROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Compact Disc-Read-Only Memory (CD-ROM), or other Memory, magnetic disk, magnetic tape, or magnetic tape, Or any other medium which can be used to carry or store data and which can be read by a computer.
The method and the system for refining the meteorological data spatial interpolation based on the deep learning disclosed by the embodiment of the invention are described in detail, a specific embodiment is applied in the method to explain the principle and the implementation mode of the invention, and the description of the embodiment is only used for helping to understand the method and the core idea of the invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present invention.
Claims (10)
1. A meteorological data spatial interpolation refining method based on deep learning is characterized by comprising the following steps:
determining boundary data of a region to be analyzed;
acquiring spatial data and meteorological data corresponding to a meteorological site in the area to be analyzed;
generating a grid point diagram to be interpolated based on the spatial data;
based on the meteorological data, carrying out spatial interpolation on the grid point diagram to be interpolated by adopting a thin plate spline function method to obtain net surface data corresponding to the area to be analyzed;
and generating a precipitation space distribution diagram and an air temperature space distribution diagram corresponding to the area to be analyzed based on the net surface data.
2. The method for refining spatial interpolation of meteorological data based on deep learning of claim 1, wherein the generating a grid point diagram to be interpolated based on the spatial data comprises:
determining a precision index of spatial interpolation based on the spatial data, wherein the precision index comprises the density and the resolution of the grid points;
generating an initial grid point diagram to be interpolated based on the precision index, wherein grid points of the initial grid point diagram to be interpolated are uniformly distributed, grid intervals are equal, and the plane coordinate of each grid point is (x)i，yi) I is 1, 2, 3, … …, n, wherein n is a grid number;
and eliminating redundant grid points which exist outside the area to be analyzed in the initial grid point diagram to be interpolated to obtain the grid point diagram to be interpolated corresponding to the area to be analyzed.
3. The method for refining the spatial interpolation of the meteorological data based on the deep learning of claim 2, wherein the spatial interpolation of the grid point diagram to be interpolated by using a thin-plate spline function method based on the meteorological data to obtain the mesh data corresponding to the area to be analyzed comprises:
based on the meteorological data and the plane coordinate (x) corresponding to each meteorological sitej，yj) J is 1, 2, 3, … …, m, where m is the weather station number, and the weather data z (x, y) for each grid point is derived from the following formula:
wherein | represents the euclidean norm, ciAs a function of the number of the coefficients,
is the kernel function of the thin plate spline method, and has the following values:
ri＝(x-xi)2+(y-yi)2。
4. the method for spatial interpolation refinement of meteorological data based on deep learning of claim 3, wherein a plane is constructed by adopting a thin-plate spline function method to fit each meteorological site, and based on a spline formed by combining a plurality of meteorological sites, a smooth plane which approximates to a control point corresponding to each meteorological site and has the smallest curvature is obtained through optimization and is used as the network surface data, wherein the smooth parameter of the smooth plane is determined by the minimization of a generalized cross validation method or the minimization of a generalized maximum likelihood;
assuming that t control points corresponding to meteorological sites are distributed in the area to be analyzed to form a known point set piAnd i is 1, 2, 3, … …, t, the coordinate of each control point is (x)i，yi，z(xi，yi) And when z (x)i，yi) With a second continuous derivative, the energy function is as follows:
and then determining the minimization of the energy function by adopting a thin plate spline function method:
Ztps＝argminE。
5. the method for refining the spatial interpolation of the meteorological data based on the deep learning of claim 1, wherein the generating the precipitation spatial distribution map and the air temperature spatial distribution map corresponding to the region to be analyzed based on the mesh surface data comprises:
constructing a legend generation model by adopting a deep learning algorithm based on a geographic information system and a national weather industry standard;
training the legend generation model by a preset color gamut range and a historical meteorological image;
processing the net surface data by adopting the legend generation model to obtain the precipitation space distribution map and the air temperature space distribution map;
the precipitation space distribution map adopts different colors in the color gamut range to represent precipitation numerical values of each area; and the air temperature space distribution map represents the air temperature value of each region by adopting different colors in the color gamut range.
6. A meteorological data spatial interpolation refining method based on deep learning is characterized in that the system comprises the following steps:
a boundary determining unit for determining boundary data of the region to be analyzed;
the data acquisition unit is used for acquiring spatial data and meteorological data corresponding to meteorological sites in the area to be analyzed;
the grid generating unit is used for generating a grid point diagram to be interpolated based on the spatial data;
the spatial interpolation unit is used for carrying out spatial interpolation on the grid point diagram to be interpolated by adopting a thin plate spline function method based on the meteorological data to obtain the mesh surface data corresponding to the area to be analyzed;
and the visualization unit is used for generating a precipitation space distribution map and an air temperature space distribution map corresponding to the area to be analyzed based on the network surface data.
7. The deep learning based meteorological data spatial interpolation refinement system of claim 6, wherein the grid generation unit comprises:
a precision determining subunit, configured to determine a precision index of spatial interpolation based on the spatial data, where the precision index includes density and resolution of a grid point;
an initial generating subunit, configured to generate an initial grid point map to be interpolated based on the accuracy index, where grid points of the initial grid point map to be interpolated are uniformly distributed, grid intervals are equal, and a plane coordinate of each grid point is (x)i，yi) I is 1, 2, 3, … …, n, wherein n is a grid number;
and the redundant eliminating subunit is used for eliminating redundant grid points which exist outside the area to be analyzed in the initial grid point diagram to be interpolated to obtain the grid point diagram to be interpolated corresponding to the area to be analyzed.
8. The system for spatial interpolation refinement of meteorological data based on deep learning of claim 7, wherein the spatial interpolation unit comprises:
based on the meteorological data and the plane coordinate (x) corresponding to each meteorological sitej，yj) J is 1, 2, 3, … …, m, where m is the weather station number, and the weather data z (x, y) for each grid point is derived from the following formula:
wherein | represents the euclidean norm, ciAs a function of the number of the coefficients,
is the kernel function of the thin plate spline method, and has the following values:
ri＝(x-xi)2+(y-yi)2。
9. the system for spatial interpolation refinement of meteorological data based on deep learning of claim 8, wherein a plane is constructed by adopting a thin-plate spline function method to fit each meteorological site, and based on a spline formed by combining a plurality of meteorological sites, a smooth plane which approximates a control point corresponding to each meteorological site and has a minimum curvature is obtained through optimization and is used as the net surface data, wherein a smooth parameter of the smooth plane is determined by minimization of a generalized cross validation method or minimization of a generalized maximum likelihood;
assuming that t control points corresponding to meteorological sites are distributed in the area to be analyzed to form a known point set piAnd i is 1, 2, 3, … …, t, the coordinate of each control point is (x)i，yi，z(xi，yi) And when z (x)i，yi) With a second continuous derivative, the energy function is as follows:
and then determining the minimization of the energy function by adopting a thin plate spline function method:
Ztps＝argminE。
10. the deep learning based meteorological data spatial interpolation refinement system of claim 6, wherein the visualization unit comprises:
the model construction subunit is used for constructing a legend generation model by adopting a deep learning algorithm based on the geographic information system and the national weather industry standard;
the training subunit is used for training the legend generation model by presetting a color gamut range and historical meteorological images;
a graph generating subunit, configured to process the mesh surface data by using the legend generation model to obtain the precipitation spatial distribution map and the air temperature spatial distribution map;
the precipitation space distribution map adopts different colors in the color gamut range to represent precipitation numerical values of each area; and the air temperature space distribution map represents the air temperature value of each region by adopting different colors in the color gamut range.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110663352.5A CN113593006A (en) | 2021-06-15 | 2021-06-15 | Meteorological data spatial interpolation refining method and system based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110663352.5A CN113593006A (en) | 2021-06-15 | 2021-06-15 | Meteorological data spatial interpolation refining method and system based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113593006A true CN113593006A (en) | 2021-11-02 |
Family
ID=78243751
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110663352.5A Pending CN113593006A (en) | 2021-06-15 | 2021-06-15 | Meteorological data spatial interpolation refining method and system based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113593006A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115269945A (en) * | 2022-09-29 | 2022-11-01 | 北京长河数智科技有限责任公司 | Big data visualization analysis method and device |
| CN115841628A (en) * | 2022-12-12 | 2023-03-24 | 中国水利水电科学研究院 | High-precision irrigation district water consumption information automatic interpretation system |
-
2021
- 2021-06-15 CN CN202110663352.5A patent/CN113593006A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115269945A (en) * | 2022-09-29 | 2022-11-01 | 北京长河数智科技有限责任公司 | Big data visualization analysis method and device |
| CN115841628A (en) * | 2022-12-12 | 2023-03-24 | 中国水利水电科学研究院 | High-precision irrigation district water consumption information automatic interpretation system |
| CN115841628B (en) * | 2022-12-12 | 2023-09-19 | 中国水利水电科学研究院 | Automatic interpretation system for high-precision irrigation area water consumption information |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Crochemore et al. | Bias correcting precipitation forecasts to improve the skill of seasonal streamflow forecasts | |
| CN108227041B (en) | Horizontal visibility forecasting method based on site measured data and mode result | |
| CN110705115B (en) | Weather forecast method and system based on deep belief network | |
| CN113593006A (en) | Meteorological data spatial interpolation refining method and system based on deep learning | |
| CN113344291B (en) | Urban inland inundation range forecasting method, device, medium and equipment | |
| CN111105068B (en) | Numerical mode correction method based on sequential regression learning | |
| CN109782374A (en) | Moisture content by assimilating inverting optimizes the method and device of numerical weather forecast | |
| CN112069449B (en) | Weather forecasting method and device based on initial value set | |
| CN111324936B (en) | Fan wind speed prediction method, computer readable storage medium and computing device | |
| CN113222019A (en) | Meteorological forecast data processing method, device and equipment for power transmission line tower | |
| KR101870263B1 (en) | The system for dynamic downscaling and display climate change senario with a high spatial resoultion | |
| CN115761303A (en) | Ground object classification method based on airborne laser radar point cloud and remote sensing image data | |
| CN112613633A (en) | Meteorological element prediction method and device for wind power plant | |
| CN112100922A (en) | Wind resource prediction method based on WRF and CNN convolutional neural network | |
| CN109033181B (en) | Wind field geographic numerical simulation method for complex terrain area | |
| CN110442836A (en) | Short-term wind speed forecasting method of wind farm and system | |
| CN116071651B (en) | Voltage equalizing field identification method and device, storage medium and terminal | |
| CN116609859A (en) | Weather disaster high-resolution regional mode forecasting system and method | |
| CN116341841A (en) | Runoff forecast error correction method, apparatus, device, medium and program product | |
| Young et al. | Initialization and setup of the Coastal Model Test Bed: integrated bathymetry | |
| CN112734684A (en) | Ocean environment information superposition and fusion display method based on electronic chart | |
| CN113988349A (en) | Numerical model forecasting product objective release method based on space-time local model | |
| Guan et al. | Rain fall predict and comparing research based on Arcgis and BP neural network | |
| KR102350075B1 (en) | A Numerical Method for Predicting High Resolution Near-Surface Air Temperature Using Fine-scale Geographical Information | |
| CN111814371B (en) | Seawater temperature field construction method based on shipborne actual measurement data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20211102 |