A rainstorm-like weather retrieval method based on height field

Abstract

The present invention relates to the field of meteorological forecasting technology, and proposes a method for retrieving similar rainstorm weather based on height field, including obtaining grid point data of current real-time weather; the grid point data includes multiple records, each record includes one-to-one corresponding longitude, latitude, height field data and flow field vector; extracting and segmenting the grid point data of current real-time weather to obtain multiple low-pressure vortices of current real-time weather; calculating the similarity between multiple low-pressure vortices of current real-time weather and low-pressure vortices of each historical weather to obtain the similarity of multiple historical weathers; sorting the multiple historical weathers according to the similarity of the historical weathers, and the sorting result is used to characterize the retrieval result of similar rainstorm weather. Through the above technical scheme, the problem of low accuracy of rainstorm weather prediction in the prior art is solved.

Description

Technical Field

[0001] The invention relates to the technical field of meteorological forecasting, and in particular to a method for retrieving rainstorm-like weather based on a height field. Background Art

[0002] With the development of the times, the level of science and technology has been continuously improved, the means of weather forecasting have become increasingly abundant, and the accuracy of forecasts has continued to improve. However, my country has a vast territory and diverse landforms, which has caused many weather changes in my country. For example, rainstorm disasters have a great impact on the economy, society, production and life. In order to minimize the harm of rainstorms, it is necessary to do preventive work in advance, which puts forward new and higher requirements for weather forecast services. Summary of the invention

[0003] The present invention proposes a rainstorm similar weather retrieval method based on height field, which solves the problem of low accuracy of rainstorm weather prediction in related technologies.

[0004] The technical solution of the present invention is as follows: obtaining grid point data of current real-time weather; the grid point data includes multiple records, each record includes one-to-one corresponding longitude, latitude, height field data and flow field vector;

[0005] Extract and segment the grid data of the current real-time weather to obtain multiple low-pressure vortices of the current real-time weather;

[0006] Calculate the similarity between multiple low-pressure vortices of current real-time weather and low-pressure vortices of each historical weather, and obtain the similarity of multiple historical weathers;

[0007] According to the similarity of historical weather, the plurality of historical weather are sorted, and the sorting result is used to characterize the retrieval result of similar weather to heavy rain;

[0008] The extraction and segmentation of the grid point data of the current real-time weather to obtain multiple low-pressure vortices of the current real-time weather specifically includes:

[0009] Constructed using the same grid data as the height field data Where V i is the height field data, <X i ,Y i > is the height field data V i At any grid point, X i is the longitude of the grid point, Y i is the latitude of the grid point;

[0010] Select any grid point as the initial anchor point, and filter out the grid points in the neighborhood of the initial anchor point; successively use the filtered grid points as new anchor points, filter out the grid points in the neighborhood of the new anchor point, until all grid points that can be used as anchor points are traversed, and the filtered grid points are used as a neighborhood set; filter out the neighborhood set that can form a closed curve, and obtain N mutually non-intersecting subsets G n , N mutually disjoint subsets G n The first curve set (G vor ,V i )={{<g1 <X i ,Y i >>,<g2 <X i ,Y i >>,…,<g N <X i ,Y i >>},V i}, where n = 1, 2, 3, ..., N;

[0011] The screening condition is: len|G n |≥3, and the subset G n At least three grid points are not on a straight line;

[0012] The flow field vector is superimposed on the grid points in the first curve set, and the grid points with cyclone conditions are selected from the first curve set according to the flow field vector to obtain the second curve set (G low_vor ,V i )={{<g1 <X i ,Y i >>,<g2 <X i ,Y i >>,…,<g M <X i ,Y i >>},V i};

[0013] For any set of second curves if If any closed curve p in If any closed curve q is found, it is judged that the closed curve p and the closed curve q form a low-pressure vortex containing multiple closed curves, and the number of closed curves in the low-pressure vortex is at least two; otherwise, it is judged that the closed curve p forms a low-pressure vortex containing a single closed curve;

[0014] If any closed curve p is completely contained in Any curve q in , specifically including:

[0015]

[0016] For any low-pressure vortex, when it is judged that the low-pressure vortex is multiple closed curves, the low-pressure vortex is saved as:

[0017] G pre ={(x in_min ,x in_max ,y in_min ,y in_max ,V in ),(x out_min ,x out_max ,y out_min ,y out_max ,V out )}

[0018] Among them, x in_min is the minimum longitude coordinate of the innermost circle of the low-pressure vortex, x in_max is the maximum longitude coordinate of the innermost circle of the low-pressure vortex, y in_min is the minimum latitude coordinate of the innermost circle of the low-pressure vortex, y in_max is the maximum latitude coordinate of the innermost circle of the low-pressure vortex, V in is the height field data of the innermost circle of the low-pressure vortex;

[0019] x out_min is the minimum longitude coordinate of the outermost circle of the low-pressure vortex, x out_max is the maximum latitude coordinate of the outermost circle of the low-pressure vortex, y out_min is the minimum latitude coordinate of the outermost circle of the low-pressure vortex, y out_max is the maximum latitude coordinate of the outermost circle of the low-pressure vortex, V out is the height field data of the outermost circle of the low-pressure vortex;

[0020] When it is judged as a single closed curve, the low-pressure vortex is saved as:

[0021] G pre ={x min ,x max ,y min ,y max ,V i}, where x min is the minimum longitude of the curve, x max is the maximum longitude of the curve, y min is the minimum latitude of the curve, y max is the maximum latitude of the curve.

[0022] Further, the similarity between the multiple low-pressure vortices of the current real-time weather and the low-pressure vortices of each historical weather is calculated to obtain the similarity of multiple historical weathers, specifically including:

[0023] Any low pressure vortex in the current real-time weather

[0024] G pre ={(x in_min ,x in_max ,y in_min ,y in_max ,V in ),(x out_min ,x out_max ,y out_min ,y out_max ,V out )}, extract the position feature P, height field feature V and intensity difference C feature of the low-pressure vortex to form the low-pressure vortex feature set G pre_fea ={P,V,C}; where, when the low-pressure vortex is a plurality of closed curves,

[0025]

[0026] V=V in

[0027] C=V out -V in

[0028] When the low-pressure vortex is a single closed curve,

[0029]

[0030] The similarities between the multiple low-pressure vortex feature sets of the current real-time weather and the multiple low-pressure vortex feature sets of each historical weather are calculated to obtain the similarities of the multiple historical weathers; the similarity calculation process of any historical weather h includes: weighted calculation of the similarity components of the multiple low-pressure vortex feature sets of the current real-time weather to obtain the similarity of any historical weather h; wherein the calculation process of the similarity components includes:

[0031] The similarity candidate values ​​of the low-pressure vortex feature set and each low-pressure vortex feature set of the historical weather h are calculated respectively to obtain multiple similarity candidate values, and the largest similarity candidate value is selected as the similarity component of the low-pressure vortex feature set; the calculation process of any similarity candidate value includes:

[0032] Set the kth low-pressure vortex feature set of the current real-time weather to G pre_fea_k = {P k ,V k ,C k}, the mth low-pressure vortex characteristic set of any historical weather h is G pre_fea_hm = {P hm ,V hm ,C hm}, then the candidate similarity values ​​of the two low-pressure vortex feature sets are:

[0033] S=γ pS p +γ V S V +γ C S C

[0034] in,

[0035] coe is the coefficient, and

[0036] Furthermore, the calculation process of any similarity candidate value also includes:

[0037] Determine whether the latitude difference between the kth low-pressure vortex feature set of the current real-time weather and the mth low-pressure vortex feature set of any historical weather h is within 10°. If not, the calculation process of the similarity candidate value is not performed. The latitude difference is determined by P k and P hm Calculated.

[0038] Furthermore, according to the flow field vector, grid points with cyclonic conditions are screened out from the first curve set to obtain the second curve set, which specifically includes: for any closed curve in the first set, any two points on the closed curve are selected, and if the outer product of the flow field vectors of the two points is positive, then it is judged that the closed curve has cyclonic conditions, otherwise it is anticyclonic conditions.

[0039] The working principle and beneficial effects of the present invention are:

[0040] In the high-altitude weather map data, the area near the low-pressure vortex in the height field is a rainy area. The accurate positioning and analysis of the low-pressure vortex is the premise and condition for the accurate forecast of rainstorm weather. The present invention extracts multiple low-pressure vortices of the current real-time weather from the grid data, and uses the same method to obtain the low-pressure vortices of the historical weather. By calculating the similarity between the low-pressure vortices, similar historical weather can be found, which is conducive to the accurate forecast of weather conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0042] Figure 1 is a flow chart of the method of the present invention;

[0043] Figure 2 It is a schematic diagram of two curved low-pressure vortices in the present invention;

[0044] Figure 3 It is a schematic diagram of the four-curve low-pressure vortex in the present invention. DETAILED DESCRIPTION

[0045] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] like Figure 1 As shown, it is a flow chart of the rainstorm similar weather retrieval method of this embodiment, which specifically includes:

[0047] Step 1: Preprocess historical data

[0048] (1) According to the geographical location to be forecasted, valid data for similarity comparison is selected from the database as historical weather data. The historical weather data can be selected from heavy rain weather data with longitude and latitude close to the area to be forecasted.

[0049] (2) Improve the data in the database, including unifying the formats of various meteorological elements and handling missing values ​​at each station.

[0050] Step 2: Obtain the grid data of the current real-time weather in the area to be forecasted, extract and segment the grid data of the current real-time weather, and obtain multiple low-pressure vortices of the current real-time weather.

[0051] (1) Extract and segment the height field data, and classify the grid points with the same height field data into Where V i is the heightened value, g(X i ,Y i ) is the high value V i The latitude and longitude coordinates of all points;

[0052] (2) Select any grid point as the initial anchor point and filter out the grid points in the neighborhood of the initial anchor point; successively use the filtered grid points as new anchor points and filter out the grid points in the neighborhood of the new anchor point until all grid points that can be used as anchor points are traversed, and the filtered grid points are used as a neighborhood set; repeat the above steps to obtain multiple neighborhood sets; filter out the neighborhood sets that can form a closed curve and obtain N mutually non-overlapping subsets G n , N mutually disjoint subsets G n The first curve set (G vor ,V i )={{<g1 <X i ,Y i >>,<g2 <X i ,Y i >>,…,<g N <X i ,Yi >>},V i}, where n = 1, 2, 3, ..., N;

[0053] If len|G n |≥3, and G n At least 3 points in the set are not on the same straight line, then it is proved that the points in the set can form a closed curve, and the set formed is:

[0054] (G vor ,V i )={{<g1 <X i ,Y i >>,<g2 <X i ,Y i >>,…,<g N <X i ,Y i >>},V i}.

[0055] (3) For any closed curve in the first curve set, superimpose the flow field vector data on the closed curve to determine whether the closed curve has cyclonic conditions. The determination method is: for any two points on the closed curve, if the outer product of the flow field vectors of the two points is positive, then the high-altitude vortex is determined to have cyclonic conditions, otherwise it is an anticyclonic condition. The vector outer product is defined as: for two vectors and Its vector outer product is written as The result of the vector outer product is also a vector, whose magnitude and direction are:

[0056] Mould length: Right now The length is numerically equal to The area of ​​the parallelogram formed by the angle θ.

[0057] direction: The direction is perpendicular to and The determined plane, The direction of the right-hand rule is from Steering To determine. vor ,V i )={{<g1 <X i ,Y i >>,<g2 <X i ,Y i >>,…,<g N <X i ,Y i >>},V i}, the points with cyclone conditions are selected to form the second curve set, denoted as (G low_vor ,V i )={{<g1 <X i ,Y i >>,<g2 <X i ,Y i >>,…,<g M <X i ,Y i >>},V i}.

[0058] (4) Determine whether the low-pressure vortex is a single closed curve or multiple closed curves. For any two second closed curve sets And V i <V j ,right If a closed curve p in If there is a closed curve q in the low-pressure vortex, then the low-pressure vortex is multiple closed curves, and the determination method is:

[0059]

[0060] In this embodiment, it is necessary to continue to determine whether the closed curve q is contained in another closed curve u, and so on, to obtain multiple closed curves with a containment relationship. Figure 2 As shown in the figure, it is an example of a low-pressure vortex composed of two closed curves. Figure 3 This is an example of a low-pressure vortex composed of four closed curves.

[0061] According to the above method, the low-pressure vortex can be defined as:

[0062] G pre ={(x in_min ,x in_max ,y in_min ,y in_max ,V in ),(x out_min ,x out_max ,y out_min ,y out_max ,V out )}

[0063] Among them, x in_min is the minimum longitude coordinate of the innermost circle of the low-pressure vortex, x in_max is the maximum longitude coordinate of the innermost circle of the low-pressure vortex, y in_min is the minimum latitude coordinate of the innermost circle of the low-pressure vortex, y in_max is the maximum latitude coordinate of the innermost circle of the low-pressure vortex, V in is the height field data of the innermost circle of the low-pressure vortex;

[0064] xout_min is the minimum longitude coordinate of the outermost circle of the low-pressure vortex, x out_max is the maximum latitude coordinate of the outermost circle of the low-pressure vortex, y out_min is the minimum latitude coordinate of the outermost circle of the low-pressure vortex, y out_max is the maximum latitude coordinate of the outermost circle of the low-pressure vortex, V out is the height field data of the outermost circle of the low-pressure vortex; if the low-pressure vortex is a single closed curve, save the low-pressure vortex as G pre ={x min ,x max ,y min ,y max ,V i}

[0065] Among them, x min is the minimum longitude of the curve, x max is the maximum longitude of the curve, y min is the minimum latitude of the curve, y max is the maximum latitude of the curve.

[0066] Step 3: Similarity measurement method of low-pressure vortex in height field

[0067] According to the characteristics that affect the similarity of low-pressure vortices, a similarity measurement method is designed. In this embodiment, the low-pressure vortex characteristics are selected as: position, height field data, and intensity difference. The specific definition is:

[0068] Position P: the center of the rectangle circumscribing the innermost circle of the low-pressure vortex;

[0069] Height field data V; Height field data of the innermost circle isobars;

[0070] Intensity difference C: the intensity difference between the outermost circle height field data and the innermost circle height field data. If it is a single closed curve, this feature is not compared.

[0071] Calculate the similarity between multiple low-pressure vortices of the current real-time weather and the low-pressure vortices of each historical weather, and obtain the similarity of multiple historical weathers, specifically including:

[0072] (1) Any low-pressure vortex in the current real-time weather

[0073] G pre ={(x in_min ,x in_max ,y in_min ,y in_max ,V in ),(x out_min ,x out_max ,y out_min ,y out_max ,V out)}, extract the position feature P, height field feature V and intensity difference C feature of the low-pressure vortex to form the low-pressure vortex feature set G pre_fea ={P,V,C}; where, when the low-pressure vortex is a plurality of closed curves,

[0074]

[0075] V=V in

[0076] C=V out -V in

[0077] When the low-pressure vortex is a single closed curve,

[0078]

[0079] (2) calculating the similarity between multiple low-pressure vortex feature sets of the current real-time weather and multiple low-pressure vortex feature sets of each historical weather, and obtaining the similarity of multiple historical weathers; the similarity calculation process of any historical weather h includes: performing weighted calculation on the similarity components of the multiple low-pressure vortex feature sets of the current real-time weather, and obtaining the similarity of any historical weather h; wherein the calculation process of the similarity components includes:

[0080] The similarity candidate values ​​of the low-pressure vortex feature set and each low-pressure vortex feature set of the historical weather h are calculated respectively to obtain multiple similarity candidate values, and the largest similarity candidate value is selected as the similarity component of the low-pressure vortex feature set; the calculation process of any similarity candidate value includes:

[0081] Set the kth low-pressure vortex feature set of the current real-time weather to G pre_fea_k = {P k ,V k ,C k}, the mth low-pressure vortex characteristic set of any historical weather h is G pre_fea_hm = {P hm ,V hm ,C hm}; First, compare the latitude difference between the two low-pressure vortex feature sets. If the latitude difference is greater than 10°, the similarity candidate values ​​of the two low-pressure vortex feature sets are calculated as:

[0082] S=γ p S p +γ V S V +γ C S C

[0083] in,

[0084]

[0085] Among them, coe can be adjusted according to actual needs. In this embodiment, coe=20.

[0086] Step 4: Calculate the results

[0087] According to the similarity measurement method, the similarity between the current weather and all historical weather can be calculated and arranged from high to low according to the similarity.

[0088] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A rainstorm similar weather retrieval method based on height field, characterized in that: include: Obtaining grid point data of current real-time weather; the grid point data includes multiple records, each record including one-to-one corresponding longitude, latitude, height field data and flow field vector; Extract and segment the grid data of the current real-time weather to obtain multiple low-pressure vortices of the current real-time weather; Calculate the similarity between multiple low-pressure vortices of current real-time weather and low-pressure vortices of each historical weather, and obtain the similarity of multiple historical weathers; According to the similarity of historical weather, the plurality of historical weather are sorted, and the sorting result is used to characterize the retrieval result of similar weather to heavy rain; The extraction and segmentation of the grid point data of the current real-time weather to obtain multiple low-pressure vortices of the current real-time weather specifically includes: Constructed using the same grid data as the height field data ,in is the height field data, The height field data is Any grid point of is the longitude of the grid point, is the latitude of the grid point; is the latitude and longitude coordinate set of all points whose height field data is Vi; Traversal All the grid points in , multiple neighborhood sets are obtained; one of the neighborhood sets is specifically: Select any grid point as the initial anchor point, and filter out the grid points in the neighborhood of the initial anchor point; successively use the filtered grid points as new anchor points, filter out the grid points in the neighborhood of the new anchor point, until all grid points that can be used as anchor points are traversed, and the filtered grid points are used as a neighborhood set; filter out the neighborhood set that can form a closed curve, and obtain N mutually non-intersecting subsets , N mutually disjoint subsets Composing the first curve set ,in ; Subset The filter criteria are: , and the subset At least three grid points in the first curve set are not on a straight line; superimposing flow field vectors on the grid points in the first curve set, and selecting grid points with cyclone conditions from the first curve set according to the flow field vectors to obtain a second curve set ; For any set of second curves ,if Any closed curve p in If any closed curve q is found, it is judged that the closed curve p and the closed curve q form a low-pressure vortex containing multiple closed curves, and the number of closed curves in the low-pressure vortex is at least two; otherwise, it is judged that the closed curve p forms a low-pressure vortex containing a single closed curve; Any closed curve p is completely contained in Any curve q in , specifically including: For any low-pressure vortex, when the low-pressure vortex contains multiple closed curves, the saved low-pressure vortex is: in, is the minimum longitude coordinate of the innermost circle of the low-pressure vortex, is the maximum longitude coordinate of the innermost circle of the low-pressure vortex, is the minimum latitude coordinate of the innermost circle of the low-pressure vortex, is the maximum latitude coordinate of the innermost circle of the low-pressure vortex, is the height field data of the innermost circle of the low-pressure vortex; is the minimum longitude coordinate of the outermost circle of the low-pressure vortex, is the maximum latitude coordinate of the outermost circle of the low-pressure vortex, is the minimum latitude coordinate of the outermost circle of the low-pressure vortex, is the maximum latitude coordinate of the outermost circle of the low-pressure vortex, is the height field data of the outermost circle of the low-pressure vortex; When the low-pressure vortex is judged to be a single closed curve, the low-pressure vortex is saved as: ,in, is the minimum longitude of the curve, is the maximum longitude of the curve, is the minimum latitude of the curve, is the maximum latitude of the curve; The method of calculating the similarity between the multiple low-pressure vortices of the current real-time weather and the low-pressure vortices of each historical weather to obtain the similarity of the multiple historical weathers specifically includes: Any low pressure vortex in the current real-time weather , extract the position feature P, height field feature V and intensity difference C feature of the low-pressure vortex to form a low-pressure vortex feature set ; Among them, when the low-pressure vortex is a plurality of closed curves, When the low-pressure vortex is a single closed curve, ； The similarities between the multiple low-pressure vortex feature sets of the current real-time weather and the multiple low-pressure vortex feature sets of each historical weather are calculated to obtain the similarities of the multiple historical weathers; the similarity calculation process of any historical weather h includes: weighted calculation of the similarity components of the multiple low-pressure vortex feature sets of the current real-time weather to obtain the similarity of any historical weather h; wherein the calculation process of the similarity components includes: The similarity candidate values ​​of the low-pressure vortex feature set and each low-pressure vortex feature set of the historical weather h are calculated respectively to obtain multiple similarity candidate values, and the largest similarity candidate value is selected as the similarity component of the low-pressure vortex feature set; the calculation process of any similarity candidate value includes: Set the kth low-pressure vortex feature set of the current real-time weather to , the mth low-pressure vortex feature set of any historical weather h is , then the candidate similarity values ​​of the two low-pressure vortex feature sets are: in, , , coe is the coefficient, and .

2. A rainstorm similar weather retrieval method based on height field according to claim 1, characterized in that: The calculation process of any similarity candidate value also includes: Determine whether the latitude difference between the kth low-pressure vortex feature set of the current real-time weather and the mth low-pressure vortex feature set of any historical weather h is within 10°. If not, the calculation process of the similarity candidate value is not performed. The latitude difference is given by and Calculated.

3. A rainstorm similar weather retrieval method based on height field according to claim 1, characterized in that: The method of selecting grid points with cyclonic conditions from the first curve set according to the flow field vector to obtain the second curve set specifically includes: for any closed curve in the first curve set, selecting any two points on the closed curve, and if the outer product of the flow field vectors of the two points is positive, then it is judged that the closed curve has cyclonic conditions, otherwise it is anticyclonic conditions.

Images