Dimension-separated regional non-rigid grid coding method

Abstract

The invention discloses a regional non-rigid grid coding method based on dimension separation, relates to the technical field of grid coding, and aims to develop research on an equidistant grid subdivision coding model oriented to a region in a city scene on the basis of the equidistant grid subdivision coding model. The regional non-rigid grid coding method based on dimension separation is provided, the cell coding method of an existing rigid tree-shaped grid is improved, and the uniqueness of administrative division identification is greatly improved. Aiming at the real situation that global administrative division data is diversified, the structure needs to be more elastic, unique identification of prefecture-level city administrative division data can be achieved through one code under a global grid unified framework, the code has spatial attributes, and range calculation and spatial relation operation can be rapidly achieved through operation of the code.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of grid coding, and particularly relates to a dimension-separated regional non-rigid grid coding method. BACKGROUND

[0002] The equidistant grid subdivision coding model is developed on the basis of the GeoSOT (Geographic coordinate Subdividing grid with One dimension integral coding on 2n Tree) earth subdivision framework. The GeoSOT grid has the characteristics of seamless and non-overlapping coverage of global longitude and latitude, can realize multi-scale expression and efficient coding calculation, and has important applications in disaster reduction and emergency, space remote sensing, land surveying and mapping, smart city, smart agriculture and the like. The GeoSOT grid has formed relevant standards such as national standards and military standards, and has played a role in the formulation of international standards such as the Open Geospatial Consortium (OGC), the International Organization for Standardization (ISO) and the Institute of Electrical and Electronics Engineers (IEEE).

[0003] The GeoSOT network realizes unified coding and efficient indexing of global spatial data through fixed-level recursive division, which plays a positive role in large-scale spatial analysis, spatial query and data sharing. However, the core advantage of the global unified model is built on the fixed and uniform grid division rule. This rule-based grid division structure has high efficiency and consistency when facing large-scale homogeneous regions, but when applied to urban scenes, problems gradually appear.

[0004] Firstly, the grid division of the rigid grid system is fixed, which will make the same city in two or more grids when identifying administrative divisions, such as the United States National Grid (USNG). It is difficult to accurately express the range of the region for the grid of fixed regional range and fixed level recursive subdivision, and many redundancies often occur. Since it is a rigid quadtree tree structure, adjacent grids can only be merged according to the rigid structure. When using this kind of grid to identify spatial objects, there are many problems such as the same object being identified in different grids, the same grid being identified in different objects, and the like. Figure 1 (a), Figure 1 (b), Figure 1(c) shown, the parent grid A contains the child grids A0, A1, A2 and A3, and the parent grids B, C and D are not shown at all. Although the size of the object does not change in the three cases, only the position changes. But in Figure 1 (a) shown, the four child grids occupied by the object have a merging relationship, so the object can be identified by the parent grid A. While in Figure 1 (b) and Figure 1 (c) shown, the four child grids have no merging relationship, which is not in the existing rigid grid structure and coding system. If the objects in the latter two cases are to be identified, there are two ways: one is to use multiple grid codes to identify the same object, and the other is to use a larger grid code to identify the object; the former way will increase the identification amount, which is not conducive to large-scale data management and storage; the latter way uses one code to identify the object, but the grid size and the matching degree of the object are low, and the redundancy is large, which is not conducive to spatial indexing.

[0005] Figure 1 (b) and Figure 1 (c) shown, the situation will occur in many locations around the world, and Figure 1 (a) has a relatively small probability of occurrence. As Figure 2 shown, when the global range is divided by the quadtree grid, the yellow circle marks the position of the four grids that cannot be merged into the upper level grid, and the red circle marks the position of the four grids that can be merged into the upper level grid. Many grids can be merged into larger grids, but they cannot be merged into the upper level grid of the non-rigid structure, such as the grid at the intersection of the equator and the prime meridian in the figure. It can be found from Figure 2 that a large number of four grids have the characteristics of adjacency and equal level, but they cannot be merged because they are not the intermediate nodes of the rigid quadtree structure, so Figure 1 (b) and Figure 1 (c) shown are inevitable. In addition, fixed grid division may cause large grid gaps in urban centers, which is also not conducive to the organization and management of grid data. SUMMARY

[0006] In view of the technical problems existing in the rigid fixed division grid in the administrative division identification, the present application proposes a dimensionally separated regional non-rigid grid coding method on the basis of the GeoSOT division and coding system.

[0007] The application is a dimensional separation regional non-rigid grid coding method, taking administrative division as coding region, coding the center point of administrative division by non-rigid grid coding, which includes regional shape code and direction scale code; the regional shape code adopts the dissecting result of GeoSOT grid selection level, which is used to describe the geographical range and spatial characteristics of the region; the direction scale code represents several different level GeoSOT grid combination, which is used to describe the direction and scale information of the grid in the region.

[0008] Preferably, two dimensional quantities are designed to code different dimensional directions, so as to realize the control of the boundary of administrative division data.

[0009] Preferably, the calculation of the dissecting range includes the following process:

[0010] Step A1, calculating the outer package rectangle frame: calculating the minimum outer package rectangle Box of administrative division A, calculating the length B X and L X of Box in latitude and longitude direction respectively, the center of Box is (B0, L0), wherein B0 and L0 are the latitude and longitude of the center of Box respectively.

[0011] Step A2, determining the level of dimensional direction:

[0012] According to B X , selecting the grid level m suitable for the latitude direction of administrative division A, the length B m of the grid in latitude direction of m level grid satisfies the condition B m+1 <B X ≤B m , wherein B m refers to the grid size of GeoSOT grid in the mth level.

[0013] According to L X , selecting the grid level n suitable for the longitude direction of administrative division A, the length L n of the grid in longitude direction of n level grid satisfies the condition L n+1 <L X ≤L n , wherein L n refers to the grid size of GeoSOT grid in the nth level.

[0014] Step A3, determining the center of non-rigid grid region: in latitude direction, calculating the column coordinate J of B0 in GeoSOT specified level grid; in longitude direction, calculating the row coordinate I of L0 in GeoSOT specified level grid.

[0015] Step A4, determination of grid level: according to the non-rigid grid region center grid coordinates (I, J) and the corresponding grid level (m, n), the coverage range of the non-rigid grid is calculated to ensure that the grid can realize full coverage of administrative division;

[0016] Wherein, the latitude direction level determination method is as follows:

[0017] Step a, calculate the latitude center point coordinates B of the grid coordinates (I, J) center ;

[0018] Step b, according to the center point coordinates and level m corresponding to the grid, calculate the initial coverage range (B min , B max ), wherein B min =B center -B m / 2, B max =B center +B m / 2;

[0019] Step c, judge the coverage relationship of the initial grid and administrative division:

[0020] If B min <B0-B X / 2, and B max >B0+B X / 2, it satisfies the coverage of the grid to the administrative division, and m is the final grid level; Otherwise, m = m-1, continue to judge the coverage, and directly satisfy the condition, and the final grid level m is determined through iterative judgment;

[0021] The longitude direction level determination method is the same as the latitude direction.

[0022] Step A5, determination of direction scale code: according to the obtained grid level m and n, direction scale code code2=(m-1)×11+(n-1) is generated.

[0023] Step A6, generation of non-rigid grid code: based on the region center, the latitude direction side length B m , the longitude direction side length L n , the defined region range is the Box region CEDG non-rigid grid subdivision range.

[0024] Preferably, the coding method of region shape code: obtain the row coordinates I and column coordinates J of the non-rigid grid region center point on the GeoSOT grid specified level grid, calculate the G domain code Encoding method of direction scale code: obtain the network levels m and n of the non-rigid grid region in the latitude and longitude directions, and generate a direction scale code Code2=(m-1)×11+(n-1), wherein m and n are less than or equal to the designated level of the GeoSOT grid.

[0025] Preferably, in the process of equidistant grid encoding and conversion, for the case of overlap of the minimum bounding rectangles Box of two administrative divisions, the processing method includes the following steps:

[0026] Step B1, determining the overlapping region: determining the overlapping region of the equidistant grids of the two regions, obtaining the spatial range and boundary information thereof, and then unifying the two sets of grid data to the same coordinate system to ensure the consistency of the grid calculation basis.

[0027] Step B2, boundary data extraction: extracting the boundary node data at the boundary from the equidistant grids of the two regions in the overlapping region, respectively, cleaning and formatting the extracted boundary data to ensure the consistency and accuracy of the data;

[0028] Specifically, a scale factor ρ=A tgt / A src , wherein A tgt is the target grid scale, and A src is the source grid scale;

[0029] The source grid scale is smaller than the target grid scale, i.e., ρ≥2, the sample points of the target grid are all grid points or center points of the grid, the number of target grids is denoted as N tgt , the source grid takes part of the grid points or center points of the grid, and the sampling method adopts a random algorithm or an equidistant sampling method, and the number of source grids is N src =2N tgt / ρ;

[0030] The source grid scale is comparable to or larger than the target grid scale, i.e., 0<ρ<2, at this time, there is almost no information redundancy, and the sample points are extracted according to the type of grid data storage, and the sample points of the source grid and the target grid are all grid points or center points of the grid.

[0031] Step B3, grid reconstruction conversion: using the extracted boundary data and combining a grid interpolation model, converting the conversion problem of the edges between cities into a grid interpolation problem, generating a set of transition grid data with the characteristics of both sides in the overlapping region through interpolation processing, and realizing seamless sharing of data;

[0032] The Monte Carlo method is used to realize the gridding of continuous spatiotemporal data, and the earth subdivision grid unit Tile and the geographic coordinate and encoding conversion rule are given Assignment strategy Sample point number N and interpolation data The execution steps are as follows:

[0033] Step C1, according to the coordinate of the grid unit Tile and the encoding conversion rule Calculate the spatial range corresponding to the grid unit

[0034] Step C2, according to the assignment extraction strategy Extract N sample points within the coverage range of the grid unit

[0035] Step C3, calculate the interpolation data at the N sample points And calculate its mean As the gridding value of the partition network unit Tile.

[0036] The present application is based on the equidistant grid partition coding model, focusing on the urban scene, and researches around the regional equidistant grid partition coding model, and proposes a regional non-rigid grid coding method based on dimension separation, which improves the existing rigid tree grid coding method and greatly improves the uniqueness of administrative division identification. On the one hand, the fixed grid division of the rigid grid system solves the problem that the same city is in two or more grids when identifying administrative divisions (such as the MGRS grid system in the United States), causing inconvenience in use, such as data conversion and analysis; on the other hand, in view of the diverse reality of global administrative division data, the structure needs to be more flexible, and a coding can be used to realize the uniqueness of the identification of prefecture-level city administrative division data under the unified framework of global grid, and the coding has spatial attributes, which can quickly realize range calculation and spatial relationship calculation through coding operation.

[0037] At the same time, in the conversion process between other data and the regional non-rigid grid coding of the present application, the coordinate conflict, data misplacement and attribute fracture problems existing in the edge area of different zoning, regional or scale grids, a grid mutual conversion idea is proposed, which realizes data conversion between different grids through data extraction and grid reconstruction alignment, converts the conversion problem between cities into a grid interpolation problem, and generates a set of transition grid data with the characteristics of both sides in the overlapping area through interpolation processing, so as to realize seamless sharing of data. BRIEF DESCRIPTION OF DRAWINGS

[0038] Figure 1 A diagram for object association problem caused by rigid structure of partition grid, wherein Figure 1 (a) is that the objects in the region belong to one parent network, Figure 1 (b) is that the objects in the region cross two parent networks, Figure 1 ​(c) is the object in the region across four parent networks;

[0039] Figure 2 For the global rigidly dissected grid cannot contain the object's position diagram;

[0040] Figure 3 For the four-tree grid encoding method of GeoSOT network diagram;

[0041] Figure 4 For the region non-rigid grid encoding structure provided in embodiment 1 diagram;

[0042] Figure 5 For the adjacent administrative boundary grid coverage example diagram;

[0043] Figure 6 For the grid point storage sample point extraction method diagram;

[0044] Figure 7 For the grid surface storage grid sample point position example diagram. DETAILED DESCRIPTION

[0045] The application will be further described below in conjunction with the drawings and specific embodiments. The embodiments of the application are given for illustration and description only, and are not intended to be exhaustive or to limit the application to the forms disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art. Embodiments were chosen and described in order to best explain the principles of the application and its practical application, and to enable others skilled in the art to understand the application for various embodiments with various modifications as are suited to the particular use contemplated.

[0046] In the GeoSOT equidistant grid partitioning encoding model, the following key concepts and terms need to be defined:

[0047] ① Equidistant grid (Equidistant grid, EG): Equidistant grid is a discrete spatial region unit formed by partitioning the local surface space of the earth, and the distance between unit positioning points is equal. Equidistant grid is not strictly defined in geography, and in this article, equidistant grid refers to the establishment of a local rectangular coordinate system on the projected surface of the earth, with the projection line of the equator as the X axis and the projection of the zero meridian or a self-defined prime meridian as the Y axis. In this local coordinate system, the grid with the same span distance along the two coordinate axes is called equidistant grid.

[0048] ② Non-rigid grid (Non-rigid grid, NGR): Non-rigid grid is a non-fixed grid structure with different dimensional scales formed by realizing scale separation of grid cells in different dimensions with the grid framework unchanged.

[0049] ③Grid level: used to describe the resolution and accuracy of the grid, different levels of grid have different sizes and quantities.

[0050] ④Grid code: a coding rule that uniquely identifies each grid, including level code, location code, etc.

[0051] In the equidistant grid earth partition space, the earth surface is divided into multiple levels of discrete areas with similar area and shape, without gaps and overlaps, and each discrete area is an equidistant grid. Let the grid system G = {G0G1···G i ···G n}, G i = {G i0 G i1 ···G ij ···G ini} represents all grids of the i-th level, and for any two grids G is , G it (s≠t) of the i-th level, we have If the coverage area of a geographic entity Entity is E, and the level is l, then the entity in the partition space can be described as a set of l-th level grids where G lj represents the partition grid of level l, and n r represents the number of grids of the r-th level covered by the entity.

[0052] For the equidistant grid G dl encoded as code, the equidistant grid computing model M ld (t) can be defined as

[0053]

[0054] where d is the spatial dimension d ∈ {2, 3}, t ∈ T is the time slice, f θ is the grid data assignment calculation function, which is determined by the specific calculation scenario and algorithm model, θ is the function parameter, is the i-th data participating in the operation in the grid encoded as code in the t time slice 1≤i≤m.

[0055] The basic element of the equidistant grid partition coding model is the partition coding. Let then and have the following relationship:

[0056]

[0057] where L represents the highest level of equidistant grid partition, and respectively represent the same longitude and latitude but different partition levels layer and layer+n. The above formula shows that the partition grid G father and G son have an empty inclusion relationship; and The first layer bits of the encoding are the same, indicating that the partition encoding has multi-scale.

[0058] The GeoSOT partition framework uses the China Geodetic Coordinate System 2000 (CGCS2000) as the reference ellipsoid system, with the coordinate origin located at the intersection of the three faces of the Earth reference ellipsoid, the prime meridian, and the equatorial plane, and has a unified conversion basis and standard. GeoSOT takes the longitude direction as the X-axis, with the prime meridian east as positive and west as negative. The latitude direction is taken as the Y-axis direction, with the equator north as positive and south as negative. GeoSOT partitions the longitude and latitude space of the entire Earth surface, with the segmentation direction along the meridian direction and the latitude direction, thereby forming a multi-level grid system covering the entire globe. Through three times of Earth expansion, the entire Earth is expanded to 512°x512°, 1°x1° is expanded to 64'x64', and 1'x1' is expanded to 64" x64". The purpose of three times of spatial expansion is shown in Figure 1 (a), Figure 1 (b), Figure 1 (c), thereby realizing global, integral, and integral four-tree partitioning, forming a multi-scale four-tree grid system from the Earth level (0 level, G level) to the centimeter level cell (32 level), see Figure 2 . With each increase in partition level, the positioning accuracy of the grid also improves, and the accuracy information of the data is implied in the encoding process. Table 1 shows the detailed information of the GeoSOT grid level and partition characteristics.

[0059] Table 1

[0060]

[0061]

[0062] GeoSOT adopts quadtree recursive subdivision for each grid, and its grid coding rule is based on quadtree algorithm and Z-order space-filling curve. The coding starts from the global 0-level grid (coded as G), and each level of grid divides the current grid into four sub-grids, using the numbers "0", "1", "2", "3" to represent different sub-regions. With the increase of the level, the grid code is refined level by level, and the coding form is a number string that increases layer by layer. In the coding process, the coordinates of the sub-regions are alternately arranged in the order of latitude first and longitude second, forming a quadtree one-dimensional sequence Morton code. The 32-level GeoSOT grid coding format is Gddddddddd-mmmmmm-ssssss.uuuuuuuuuuu, and the quadtree grid coding mode is shown in Figure 3 .

[0063] Embodiment 1

[0064] When identifying administrative divisions, due to the diversity of the shapes of administrative divisions and the regularity of the regular grid itself, when identifying the boundary profile, especially for administrative division data with nested relationship, repeated coding is inevitable.

[0065] Therefore, the embodiment proposes a dimensionally separated regional non-rigid grid coding method, which takes administrative division as the coding region and codes the center point of the administrative division in a non-rigid grid. The non-rigid grid here refers to replacing the rigid fixed tree structure between grid levels with a tree structure with variable hierarchical relationship, forming a hierarchical grid structure with multiple types of grid units across grid points and grid edges, and using a unified coding method for identification and calculation.

[0066] The advantage of the non-rigid grid is that, on the one hand, the grid division of the rigid grid system is fixed, and when identifying administrative divisions, the same city will be in two or more grids (such as the MGRS grid system in the United States), causing inconvenience in use, such as data conversion and analysis; on the other hand, in view of the diverse reality of global administrative division data, the structure needs to be more flexible, enabling a global grid unified framework to achieve unique identification of prefecture-level city administrative division data using one code, and the code has spatial attributes, enabling quick range calculation and spatial relationship calculation through code operation.

[0067] The non-rigid grid coding includes a region shape code and a direction scale code, as shown in Figure 4 The region shape code uses the subdivision results of the selected level of GeoSOT grid to describe the geographical range and spatial characteristics of the region; the direction scale code represents the combination of several different levels of GeoSOT grid, and is used to describe the directionality and scale information of the grid in the region.

[0068] At the same time, two dimensional quantities are designed to encode different dimensional directions in the embodiment, realizing the control of administrative division data boundary, and the calculation of the range of dissection includes the following processes:

[0069] 1. Calculate the minimum outer package rectangle Box of administrative division A, and calculate the length B X and L X of Box in latitude and longitude direction respectively, the center of Box is (B0, L0), wherein B0 and L0 are the latitude and longitude of the center of Box respectively.

[0070] 2. Determine the level of dimensional direction:

[0071] According to B X , select the grid level m suitable for the latitude direction of administrative division A, the length B m of the grid in latitude direction at m level satisfies the condition B m+1 <B X ≤B m , wherein B m refers to the grid size of GeoSOT grid at m level.

[0072] According to L X , select the grid level n suitable for the longitude direction of administrative division A, the length L n of the grid in longitude direction at n level satisfies the condition L n+1 <L X ≤L n , wherein L n refers to the grid size of GeoSOT grid at n level.

[0073] 3. Determine the center of non-rigid grid area: in latitude direction, calculate the column coordinate J of B0 in GeoSOT designated level grid; in longitude direction, calculate the row coordinate I of L0 in GeoSOT designated level grid.

[0074] 4. Determine the grid level: according to the grid coordinate (I, J) of non-rigid grid area center and the corresponding grid level (m, n), calculate the coverage range of non-rigid grid, to ensure that the grid can realize the full coverage of administrative division;

[0075] Wherein, the latitude direction level determination method is as follows:

[0076] Step a, calculate the latitude center point coordinate B center corresponding to grid coordinate (I, J);

[0077] Step b, according to the center point coordinate and level m corresponding to the grid, calculate the initial coverage range (B min , B max ), wherein Bmin = B center -B m / 2, B max = B center + B m / 2;

[0078] Step c, judging the covering relationship between the initial grid and the administrative division:

[0079] If B min <B0-B X / 2, and B max >B0+B X / 2, the grid covers the administrative division, and m is the final grid level; otherwise, m = m-1, continue to judge the covering, and the direct satisfaction of the condition is the main one, and the final grid level m is determined through iterative judgment;

[0080] The longitude direction level determination method is the same as the latitude direction.

[0081] 5. Determination of direction scale code: according to the obtained grid levels m and n, a direction scale code code2 = (m-1) x 11 + (n-1) is generated.

[0082] 6. Generation of non-rigid grid code: based on the region center, the latitude direction side length B m , the longitude direction side length L n , the defined region range is the Box region CEDG non-rigid grid division range.

[0083] For the determined CEDG non-rigid grid division range of the administrative division, the coding method of the region shape code: the row coordinate I and the column coordinate J of the non-rigid grid region center point on the 11th layer grid of GeoSOT are obtained, and the G domain code of the non-rigid grid region is calculated That is, the G domain code adopts the division result of the 11th level of GeoSOT, and there are 22 bits in total.

[0084] The advantages of the above G domain coding method are:

[0085] First, uniqueness. Since the center points of all administrative division data are different, by establishing a coding mapping relationship based on the center points of the administrative division data It can be ensured that each administrative division unit obtains a unique identification code, and this mechanism effectively avoids the occurrence of repeated coding.

[0086] Second, accuracy, thanks to the GeoSOT encoding system centimeter level of minimum resolution characteristics, theoretically through the method of calculating the center point can be infinitely close to the actual center point, to achieve centimeter level precision matching. However, considering the optimization needs of the design of the encoding system, this embodiment through the overall design of CEDG encoding, only 22 bits are used for center point encoding identification, and the center point error is effectively controlled within 0.087°. The error value is the maximum value of the 11th layer grid scale of GeoSOT, which is usually smaller in actual application.

[0087] In the design of the scale direction (i.e. T domain), 11 layers of encoding structure are used to generate 121 combinations, corresponding to 7-bit encoding space, which has good consistency with the actual situation, ensuring the encoding accuracy and optimizing the encoding efficiency. The direction scale code encoding method provided by the embodiment: obtaining the network levels m, n of the non-rigid grid region in the latitude and longitude direction, generating the direction scale code Code2=(m-1) x 11+(n-1), wherein m, n are less than or equal to the specified level of GeoSOT grid. Since the specified level of this embodiment is 11, the value range of m, n is an integer between 1 and 11.

[0088] Embodiment 2

[0089] In the context of multi-source heterogeneous spatial data integration, how to efficiently convert data from different sources in the same spatial expression system has always been the focus of researchers. Spatial interpolation, as a key means of converting discrete data into continuous data, plays a crucial role in grid data conversion. Spatial interpolation method can estimate the attribute value of unknown position based on limited observed data by considering the correlation between spatial data, thereby realizing high-precision reconstruction of discrete data.

[0090] The conversion method between the geometric features of raster data and vector data (including point, line, surface, and body objects) and the regional non-rigid grid encoding disclosed in embodiment 1 is not the focus of this embodiment. This embodiment focuses on the coordinate conflict, data misplacement, and attribute fracture problems existing in the edge area of different zoning, regional, or scale grids. This embodiment proposes a grid conversion idea, which realizes data conversion between different grids through data extraction and grid reconstruction alignment.

[0091] In the equidistant grid subdivision model, due to the differences in coordinate system, data alignment, and local consistency, etc., different zoning, different regions, or different scale equidistant grid data at the junction (edge area) will lead to data discontinuity and significant local differences. This phenomenon is referred to as "grid edge problem". Figure 5As shown in the figure, the green grid matrix represents the equidistant grid of Beijing, the orange grid matrix represents the equidistant grid matrix of Tianjin, and the red part represents the overlapping part of the two equidistant grids. The core of solving the grid edge problem lies in realizing the smooth transition and seamless fusion of the data in the edge area.

[0092] The processing method provided by the embodiment includes the following steps:

[0093] 1. Determine the overlapping area: determine the overlapping area of the equidistant grids of the two areas, obtain the spatial range and boundary information, and then unify the two sets of grid data to the same coordinate system to ensure the consistency of the grid calculation basis.

[0094] 2. Boundary data extraction: In the overlapping area, the boundary node data at the boundary is extracted from the equidistant grids of the two areas respectively, and the extracted boundary data is cleaned and formatted to ensure the consistency and accuracy of the data.

[0095] Extracting sample point data from the grid is the basis of the grid conversion method, and the quality and representativeness of the extracted samples are crucial to data reconstruction. Grid data extraction refers to the conversion of grid data to latitude and longitude data, that is, the data is converted from the form of <grid ID, data> to the form of <latitude and longitude, data> through sampling method. In the grid data model, the selection of data storage method is essentially the mapping of spatial feature cognition and calculation logic. Grid data is usually stored in grid points or grid surfaces.

[0096] For the inherent properties with physical continuity (such as terrain elevation, real-time temperature and humidity), grid point storage records the measured value of each node in array form by discretizing the spatial coordinates. This storage method follows the Nyquist sampling theorem to capture the spatial gradient of continuous field with fixed resolution, which can provide accurate grid point coordinate position and attribute value corresponding to the grid point, and support interpolation, differentiation and other accurate calculations based on point value. Three typical methods of extracting sample points are shown in the figure, wherein the blue points are sample points and the red points are extracted sample points. Figure 6

[0097] ​For statistical aggregate data (such as population density, average precipitation), the nature of its attributes is a regional derived value, generated by unit-level statistics (such as area weighting, sample average), rather than point-by-point measurement of physical space. The cell storage binds the attribute value to the geometric range of the grid cell, following the assumption of regional homogeneity, and the attribute value within each cell is a "virtual average" in a statistical sense. This storage method has significant advantages in land use classification and environmental monitoring average scenarios, both by sharing attribute values to reduce storage redundancy and by ensuring consistency between data production logic (such as "first zoning, then statistics") and storage structure. Such grid data cannot directly obtain the location coordinates and attribute values of the sample points, and needs to be preprocessed to determine the sample points. The location of the sample point can be the geometric center of the grid, a random sampling point in the grid, or a regular point array in the grid, as shown in Figure 7 .

[0098] This embodiment defines the scale factor ρ = A tgt / A src , that is, the ratio of the area of the target grid cell to the source grid cell. According to the scale difference between the source grid and the target grid, the sample point extraction method and extraction density are determined, which can be divided into the following two cases:

[0099] ① The source grid scale is smaller than the target grid scale, that is, ρ ≥ 2, the sample points of the target grid are the grid points or center points of all grids, and the number of target grids is N tgt , and the source grid takes part of the grid with grid points or center points, and the sampling method uses a random algorithm or equidistant sampling, and the number of source grids is N src = 2N tgt / ρ;

[0100] ② The scale of the source grid is comparable to or larger than the scale of the target grid, that is, 0 < ρ < 2, at this time there is almost no information redundancy, and the sample points are extracted according to the type of grid data storage, and the sample points of the source grid and the target grid are all grid points or center points of all grids. When the source grid scale is small, the spatial interpolation model should be fully utilized to generate sample point data of the target grid.

[0101] 3. Grid reconstruction conversion: using the extracted boundary data, combining the grid interpolation model, converting the conversion problem between cities into a grid interpolation problem, generating a set of transition grid data with the characteristics of both sides in the overlapping area through interpolation processing, so as to realize seamless sharing of data.

[0102] Data gridding refers to converting latitude and longitude data into grid data, that is, converting data from the form of <latitude and longitude, data> to the form of <grid ID, data> through sampling methods. There are many methods to convert continuous point data in latitude and longitude form into discrete point data in grid form.

[0103] The process of data gridding is essentially to establish the mapping of data points to grid data, which can be symbolized as follows. The longitude and latitude in the coordinate point are the data point p i , where λ is the longitude value, is the latitude value, and the grid data is represented as G(g), g is the grid code under a certain grid system. Data gridding can be abstracted as G(g) = f(p1, p2, …, p n ). All data gridding methods can be considered as specific expressions of this abstract formula and can be expressed under this expression framework.

[0104] Monte Carlo method is a numerical simulation and calculation method guided by probability and statistics theory. This method solves many problems that are difficult to calculate explicitly through formulas by using random numbers or pseudo-random numbers in computers. Monte Carlo method usually includes two types of problems. The first type of problem is to solve the problem itself with inherent randomness, and to simulate the numerical simulation by constructing a random process model. The second type is to convert the problem to be solved into a certain random distribution characteristic number, and then estimate the probability of the frequency of random events based on the idea of random sampling, and then take the frequency as the solution to the problem.

[0105] The Monte Carlo method is used to realize the gridding of continuous spatio-temporal data. Given the earth subdivision grid unit Tile and the geographic coordinate and coding conversion rule assignment strategy the number of sample points N and the interpolated data The execution steps are as follows:

[0106] Step C1, according to the coordinate and coding conversion rule of the grid unit Tile calculate the spatial range corresponding to the grid unit

[0107] Step C2, according to the assignment extraction strategy extract N sample points within the coverage range of the grid unit

[0108] Step C3, calculate the interpolated data at the N sample points and calculate its mean as the gridding value of the subdivision network unit Tile. Considering the importance of sample points and the non-uniformity of spatial distribution, a weight function can be introduced in the statistical estimation.

[0109] ​Since the spatio-temporal data is continuously changing in certain time and space, Monte Carlo method can better capture the continuity and variation characteristics of the data through random sampling and statistical analysis, and can better ensure the accuracy and authenticity of the data in numerical calculation. When the sample number N→∞, the estimated value converges to the true value S(Tile).

[0110] Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by ordinary skilled in the art and related fields without creative labor shall belong to the scope of protection of the present application.

Claims

1. A dimensionally separated region non-rigid mesh encoding method, characterized in that, Using administrative divisions as the coding area, the center point of the administrative division is coded using a non-rigid grid. This non-rigid grid coding includes a region shape code and a direction scale code. The region shape code uses the subdivision results of the selected level of the GeoSOT grid to describe the geographical range and spatial characteristics of the region. The orientation scale code represents a combination of several different GeoSOT grids at different levels, used to describe the orientation and scale information of the grids within a region.

2. The regional non-rigid grid coding method according to claim 1, characterized in that, The design incorporates two dimensions for encoding in different directions to control the boundaries of administrative division data.

3. The regional non-rigid grid coding method according to claim 2, characterized in that, The calculation of the partition range includes the following process: Step A1, Calculate the bounding rectangle: Calculate the minimum bounding rectangle Box of administrative division A, and calculate the length B of Box in the latitude and longitude directions respectively. X and L X The center of Box is (B0, L0), where B0 and L0 are the latitude and longitude of the center of Box, respectively. Step A2, Determining the dimensional direction hierarchy: According to B X Select a grid level m that corresponds to the latitude of administrative division A, and the latitudinal side length B of the m-level grid. m Condition B is satisfied m+1 <B X ≤B m B m This refers to the mesh size of the GeoSOT mesh at layer m; According to L X Choose a grid level n that corresponds to the longitude direction of administrative division A, and let the longitude side length L of the n-level grid be... n Condition L is satisfied n+1 <L X ≤L n L n This refers to the grid size of the GeoSOT mesh at layer n; Step A3, Determining the center of the non-rigid grid region: In the latitudinal direction, calculate the column coordinate J of B0 in the GeoSOT specified level grid; In the longitude direction, calculate the row coordinates I of L0 in the GeoSOT grid at the specified level; Step A4, Determining the grid level: Calculate the coverage area of ​​the non-rigid grid based on the center grid coordinates (I,J) and the corresponding grid level (m,n) of the non-rigid grid region to ensure that the grid can achieve full coverage of the administrative divisions; Step A5, Determination of orientation scale code: Based on the obtained grid levels m and n, generate orientation scale code code2 = (m-1) × 11 + (n-1); Step A6, Generation of non-rigid grid encoding: Based on the region center, the latitudinal side length B m Longitude side length L n The defined region is the range of the Box region CEDG non-rigid mesh.

4. The regional non-rigid grid encoding method according to claim 3, characterized in that, In step 4, the method for determining the latitude direction hierarchy is as follows: Step a, calculate the latitude center point coordinates B corresponding to the grid coordinates (I,J). center ; Step b: Calculate the initial coverage area (B) based on the center point coordinates and layer m of the grid. min B max ), where B min =B center -B m / 2, B max =B center +B m / 2; Step c, determine the coverage relationship between the initial grid and the administrative divisions: If B min <B0-B X / 2, and B max >B0+B X If the value is 2, then the grid covers the administrative division, and m is the final grid level; otherwise, m = m-1, and the coverage judgment continues, with the condition being the primary one, and the final grid level m is determined through iterative judgment. The method for determining the longitude level is the same as that for the latitude level.

5. The regional non-rigid grid encoding method according to claim 3, characterized in that, Encoding method for region shape codes: Obtain the row coordinates I and column coordinates J of the center point of the non-rigid mesh region on the GeoSOT mesh at a specified level, and calculate the G-domain code that forms the non-rigid mesh region.

6. The regional non-rigid mesh coding method according to claim 4, characterized in that, The encoding method of the orientation scale code is as follows: obtain the network level m and n of the non-rigid grid region in the latitude and longitude directions, and generate the orientation scale code Code2 = (m-1)×11+(n-1), where m and n are less than or equal to the specified level of the GeoSOT grid.

7. The regional non-rigid mesh coding method according to claim 6, characterized in that, During the equidistant grid encoding conversion process, the handling of cases where the minimum bounding rectangles (Box) of two administrative divisions overlap includes the following steps: Step B1, Determine the overlapping area: Identify the overlapping area of ​​the two equidistant grids, obtain its spatial range and boundary information, and then unify the two sets of grid data into the same coordinate system to ensure that the basis of grid calculation is consistent; Step B2, Boundary Data Extraction: Within the overlapping area, extract the boundary node data at the junction from the equidistant grids of the two areas respectively. Clean and format the extracted boundary data to ensure data consistency and accuracy. Step B3, Grid Reconstruction and Transformation: Using the extracted boundary data and combined with the grid interpolation model, the problem of transforming the boundary between cities is transformed into a grid interpolation problem. Through interpolation processing, a set of transitional grid data compatible with the characteristics of both sides is generated in the overlapping area, thereby achieving seamless data sharing.

8. The regional non-rigid mesh encoding method according to claim 7, characterized in that, In step B2, the scaling factor ρ = A is defined. tgt / A src A tgt For the target grid scale, A src The source grid scale; The source grid scale is smaller than the target grid scale, i.e., ρ≥2. The sample points of the target grid are the grid points or center points of all grids. The number of target grids is denoted as N. tgt The source mesh is a portion of the grid points or center point of the grid. The sampling method uses a random algorithm or equal-interval sampling. The number of source meshes is N. src =2N tgt / ρ; If the scale of the source grid is similar to or larger than that of the target grid (i.e., 0 < ρ < 2), there is almost no information redundancy. Sample points are extracted according to the data storage type of the grid. The sample points of both the source and target grids are grid points or center points of all grids.

9. The regional non-rigid mesh encoding method according to claim 7, characterized in that, In step B3, the Monte Carlo method is used to grid the continuous spatiotemporal data, given the Earth's grid unit tiles and geographic coordinate and encoding conversion rules. Assignment strategy Number of sample points N and interpolated data The execution steps are as follows: Step C1, according to the coordinate and encoding conversion rules of the grid cell Tile. Calculate the spatial extent corresponding to this grid cell. Step C2, extract values ​​according to the assignment strategy Within the coverage of grid cells Extract N sample points from within Step C3: Calculate the interpolated data at N sample points. And calculate its mean. As the meshing value of the Tile, a subdivided network unit.

Images