Interval-based polygon geocoding for dggs
By converting barycentric coordinates to sequential indices and employing interval-based polygon geocoding, the method addresses the inefficiencies in DGGS systems for polygon data, improving computational efficiency and data compression for geospatial analysis.
Patent Information
- Application Number
- PCT/CA2026/050695
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2025-05-14
- Filing Date
- 2026-05-14
- Publication Date
- 2026-11-19
AI Technical Summary
Current DGGS systems struggle with efficient handling of polygon data, particularly in geocoding, which is essential for analyzing and visualizing geospatial data, due to the trade-off between accuracy and computational efficiency, and lack effective methods for handling vector data within these systems.
The method involves converting barycentric coordinates to sequential index coordinates, indexing cells within a DGGS using a numerical sequence based on common boundaries or vertices, and applying interval-based polygon geocoding to efficiently identify and compress cells of interest, allowing for efficient Boolean operations across multiple resolutions.
This approach enhances computational efficiency and data compression, enabling accurate and rapid analysis of geospatial data by allowing for efficient geocoding of polygons at various resolutions and facilitating direct comparison and operation between polygons with different resolutions.
Smart Images

Figure CA2026050695_19112026_PF_FP_ABST
Abstract
Description
INTERVAL-BASED POLYGON GEOCODING FOR DGGSRELATED APPLICATION
[0001] This application claims priority to U.S. Provisional Patent Application No. 63 / 805,612, filed 14 May 2025, the entire contents of which are incorporated herein by reference.FIELD
[0002] The disclosure relates to systems and methods for polygon geocoding for discrete global grid systems (DGGS). The systems and methods described enable efficient handling of geospatial areas mapped to a DGGS, including improvements in computation times, data compression and efficient Boolean operations with large data sets.BACKGROUND
[0003] With the ever-growing amount of information being collected about the Earth comes a need for tools for integrating, analyzing, simulating, distributing and visualizing this data to solve real-world problems. To analyze and simulate processes on the Earth, it is common to discretize the surface of the Earth into a grid to make spatial computations tractable. The grid allows for efficient storage and retrieval of data with spatial components and indexing their spatial neighbourhoods. One state-of-the-art technique for constructing such a grid on the Earth is a Discrete Global Grid System (DGGS).
[0004] In a DGGS, the Earth is discretized into a multi-resolution hierarchy of mostly regular cells where each cell is given a unique identifier for spatial data association. DGGS provides a unified framework for data integration, analysis and scientific visualization that allows the user to store and query many different types and formats of data with geospatial components, encoding their indices on polyhedral representation of the Earth.
[0005] Raster data (pixelated data associated with an area) is naturally handled in DGGS and can be analyzed using Map Algebra. Logical and Boolean operations are common examples of Map Algebra operations. Raster data can be stored at different resolutions of the DGGS depending on the desired representation and accuracy of the data. There is often a trade-off in efficiency and accuracy in spatial analysis, with finer spatial resolutions havingmore accurate computations at the cost of slower processing, while coarser resolutions having less accurate solutions but faster computation The trade-off is due, in part, to the number of area elements to process increases exponentially with higher resolutions. The storage, compression, and analysis of raster data is an area of active research.
[0006] Another common type of information GIS applications handle is vector data (e.g., points, polylines and polygons on the surface of Earth). In order to perform a broader analysis and harness the full potential of DGGS, it is common to discretize vector data onto the DGGS grid [6] [4], This process of inserting vector data into the DGGS grid to get a raster representation is called geocoding. Algorithms currently exist to geocode point and polyline data, with continued research; however, full treatment of polygon data is particularly difficult, and an area of active research. It is common in Digital Earth applications to analyze raster and vector datasets simultaneously.
[0007] For example, it is particularly interesting to examine a temperature dataset of the Earth-represented as a raster to understand the average temperature within a particular area such as a particular country-represented as a vector-data boundary.
[0008] If a raster dataset of temperature is laid out across the surface of the Earth and a vector dataset of country borders is configurable to the Earth, one can examine the data within a defined vector dataset to obtain information of interest, such as the approximate average temperature within a country’s borders (e.g., Canada).
[0009] The set of high-level operations to compute the average temperature using a DGGS is to find the unique DGGS cells in the raster inside the border of Canada, which is a trivial operation when the area of Canada is geocoded with the same DGGS. Analysis of the data associated with particular cells becomes more efficient with a consistent DGGS underlying the geocoding process, such as computing the average temperature in each of those cells.
[0010] To perform operations like this efficiently, advanced geospatial data structures are required for robust and efficient geocoding which should also allow for efficient operations (e.g., Map Algebra) and compression of data after geocoding. Current Digital Earth and DGGS applications fulfill only a subset of the requirements, there has been a need for systems enabling faster and more efficient handling of data within DGGS systems across all requirements.SUMMARY
[0011] In accordance with the disclosure, a method of converting a barycentric coordinate to a sequential index coordinate, the method comprising the steps of:for a DGGS cell having barycentric coordinates and 1st, 2ndand 3rdbarycentric coordinate numbers (BCNs):• for each of the 1st, 2ndand 3rdBCNs, convert to binary number bits;• interleave the binary number bits from step a) into groups of triplets;• process each triplet to map a triplet to an index map, the index map having 4 designators, where each triplet is mapped to one of the 4 designators to form a converted triplet; and,• interpret each number from step c) in base 4 to form a sequential index coordinate.
[0012] In various embodiments:• the index map maps barycentric coordinates (0,1,1) to 0, (1,1 ,1) to 1, (1,1 ,0) to 2 and (1 ,0,1) to 2.• the steps are reversed to convert sequential index coordinates to barycentric coordinates.• for a list of sequentially indexed coordinates, determining if cells within the list of indexed cells are sequential or not, further comprising the steps of:sorting the list of indexed cells into an ascending or descending order; and,starting with the lowest or highest cell number, testing each successive cell to determine if it is sequential with its preceding cell to either create a list of successive numbers or to determine that a successive number is not successive.• if a number is not successive, further comprising the step of closing a current list of successive numbers.• further comprising the step of repeating the process to determine a next list of successive numbers until all numbers of the list have been processed.
[0013] In another aspect, a method of indexing cells of a discrete global grid system (DGGS) and identifying cells of interest within the DGGS is described, comprising the steps of:within a DGGS model having a plurality of triangular parent and child cells, wherein a parent cell contains the child cells:indexing each child cell within the parent cell as a numerical sequence of successive child cells sharing a common boundary (CB) or common child vertex (CCV) wherein the numerical sequence is a progression of numbers starting at a parent cell vertex wherein indexing includes:a. numbering successively first CC1, second CC2 and third CC3 child cells wherein CC1 and CC2 share a common boundary and CC2 and CC3 share a common boundary;b. numbering successively a fourth CC4 child cell sharing a common child vertex CCV with CC3 and a common child cell boundary with CC2 and wherein CC1-CC4 is a child numbering set;c. repeating steps a)-c) n times to successively number each child cell of the parent cell wherein numbering of a successive child number set is initiated across a common child cell boundary or common vertex, and,d. overlaying a first border on the DGGS wherein the first border overlays one or more child cells and determining child cells overlapping and within the first border.
[0014] In various embodiments:• Step d) includes: identifying a first cell intersecting the first border; and, checking each successively numbered cell and determining if each successively numbered cell is overlapping or within the first border.• The method further comprises the step of: for each successively numbered cell determined as overlapping or within the first border, marking each successively numbered cell as a group of cells.• The method further comprises the step of: tabulating each group of cells as single data points and wherein the single data points collectively define cells overlapping or within the first border.• The method further comprises the step of: of e) overlaying a second border wherein the second border overlays one or more child cells and determining cells overlapping and within the second border.• step e) includes identifying a first cell intersecting the second border; and, checking each successively numbered cell and determining if a successively numbered cell is overlapping or within the second border.• The method further comprises the step of: for each successively numbered cell determined as overlapping or within the second border, marking each successively numbered cell overlapping or within the second border as a second border group of cells.• The method further comprises the step of: tabulating each group of cells as single data points and wherein the single data points collectively define cells overlapping or within the second border.• The method further comprises the step of: conducting a Boolean union operation on the cells defined by the first and second borders utilizing the single data points.• The method further comprises the step of: conducting a Boolean intersection operation on the cells defined by the first and second borders utilizing the single data points.• The method further comprises the step of: conducting a Boolean difference operation on the cells defined by the first and second borders utilizing the single data points.
[0015] In another aspect, a digital global grid system (DGGS) is described, the DGGS comprising:a DGGS model defined by triangle parent and child cells at a plurality of resolutions wherein each resolution is defined as a degree of resolution and wherein a degree of resolution of a child cell is defined by a triangular cell within the parent cell and having a border length 50% of a border length of the parent cell;wherein a parent cell contains four child cells; and,wherein each child cell within a parent cell is indexed as a numerical sequence of successive child cells sharing a common boundary (CB) or common child vertex (CCV) wherein the numerical sequence is a progression of numbers starting at a parent cell vertex and wherein child cells within a parent cell are numbered successively as first, second and third child cells wherein the first and second child cells share a common boundary and the second and third child cells share a common boundary; and a fourth child cell shares a CCV with the third child cell and a common child cell boundary with the second child cell and wherein the first, second, third and fourth child cells is a child numbering set.
[0016] In another embodiment, the system includes a plurality of adjacent parent cells and wherein sequential indexing of child cells within each parent cell progresses between adjacent parent cells.
[0017] In another aspect, a method of marking and compressing discrete global grid system (DGGS) data is described, the method comprising the steps of:in a DGGS system having n triangular child cells sharing common borders and vertexes and being numbered sequentially through the common boundary and vertexes and where a border is overlaid at least one child cell;• starting a cell n, checking if cell n has an overlaid border, andi. if yes marking cell n and setting a border condition to open; and,ii. if no, discarding cell n and set border condition to nil;1. increment n by 1 and check if cell n+1 has an overlaid border;a. if yes, performing step a) i)b. if no, repeating step a) ii);• if step a) is yes, incrementing n by 1 and checking if cell n+1 has an overlaid border, andi. if yes, setting a border condition to closed; and1. marking cell n+1 and change border condition to closed; and2. tabulate closed cells as a group;ii. if no, repeating step a) i);
[0018] In another embodiment, the method includes the step of confirming if all cells have been checked and if no, increment n by 1 and repeat step a) and, if no, stop.BRIEF DESCRIPTION OF THE DRAWINGS
[0019] Various objects, features and advantages of the disclosure will be apparent from the following description of particular embodiments, as illustrated in the accompanying drawings. The drawings are not necessarily to scale; emphasis is instead placed upon illustrating the principles of various disclosure embodiments. Similar reference numerals indicate similar components.Figures 1A-1B are figures showing a DGGS grid and polygon overlaid over a sphere. Figure 1A shows land / sea boundary data on the sphere and Figure 1B shows the associated mean temperature data associated with individual cells of the DGGS.Figure 1C illustrates a DGGS grid having a parent cell and multiple resolutions of child cells.Figures 2A-2C are figures showing a process for identifying DGGS cells bordering and within a polygon overlaid on a DGGS at different resolutions.Figures 3A-3I are figures illustrating a scanline process for identifying DGGS cells bordering and within a polygon.Figures 4A-4C are figures illustrating a scanline process for identifying cells bordering or within a polygon and a principle of compression.Figures 4D(1) and 4D(2) are figures comparing a scanline process and an alternate process for identifying DGGS cells bordering and within a polygon.Figures 5A-5B are figures illustrating different resolution cells bordering or within a polygon of DGGS cells.Figures 6A-6B are figures illustrating principles of Boolean operations with cells of a DGGS with overlain polygons.Figures 7A-7D illustrate the process of assigning portions of polygon overlaid a DGGS to different cells of the DGGS and the principles of projection.Figures 8A-8Q are figures illustrating processes of indexing triangular DGGS cells at varying resolutions, processes for identifying features and location of overlaid polygons within a DGGS indexing system and processes for compressing data associated with cells.Figure 8R illustrates an indexing method in accordance with one embodiment.Figure 8S is a flowchart showing a process for determining if cells are overlapping or within a border in accordance with one embodiment.Figure 8T is a graphical representation showing how groups of cells may be bracketed when analyzed to determine if cells are overlapping or are within a border in accordance with one embodiment.Figures 9A-9C are figures illustrating processes for cell data compression. Figures 9D(A)-9D(F) are figures illustrating a sequenced indexing process showing (A) a basic parent cell having 4 child cells; (B) 4 parent cells and how child cells of adjacent parent cells may be numbered similar to the number shown in (A); (C) sequential numbering of all child cells within 4 adjacent parent cells; (D) linkages between adjacent parent cells; (E) barycentric coordinates of a corresponding set of parent and child cells with an overlain polygon; and (F) sequentially indexed cells with an overlain polygon.Figure 9E shows an indexing map (upper left) and a grid of 16 child cells with a comparison of both barycentric and sequential indexing coordinates.Figure 9F shows a “walking” pattern showing how one parent cell may be sequentially numbered and how the walking pattern may be repeated across multiple adjacent parent and child cells.Figures 10A-10B are figures illustrating processes for conducting Boolean operations with compressed cell data.Figure 11A is a flowchart describing a process for index conversion in accordance with one embodiment.Figure 11B is a flowchart describing a process for bracketing cells within a list of cells to determine if cells within the list are sequentially numbered with respect to other cells in the list.Figure 11C is a flowchart describing a process for changing cell numbering for different resolutions.DETAILED DESCRIPTION
[0020] With reference to the figures, systems and methods for geocoding vector data on the surface of a discrete global grid system (DGGS) with efficient operations and compression are described.
[0021] Polygon geocoding refers to techniques used to convert a polygon on the surface of a sphere (with latitude / longitude points adjoined by great circle arcs) to a corresponding raster (set of DGGS cells covering the polygon). Each of these cells can also have locational data associated with it.DGGS and Associated Data
[0022] By way of additional background and initial description, the location of points on a spherical object, such as the Earth, can be represented by coordinate systems such as latitude and longitude. The spherical object can be represented by a discrete global grid system (DGGS), namely a polyhedra. The polyhedra represents the spherical object with planar surfaces representing the outer surface of the sphere. For the purposes of this description,while the grid of a DGGS can be represented by various base polyhedrons, the DGGS referenced herein comprises triangle cells that are equilateral triangle cells. Accordingly, it is understood that the systems and methods described herein can be applied to a DGGS represented by any triangle discretization and are not limited to those that are explicitly equilateral triangle cells. Further, the DGGS models are generally described assuming the DGGS is representing the Earth, however, it is understood that the principles described herein may be applied to DGGS’s of other spherical, ellipsoidal, or arbitrarily triangle-tessellated objects.
[0023] As shown in Figures 1A and 1 B, the Earth is represented by a polyhedra comprising triangular cells C. As shown, geospatial data may be associated with cells of a DGGS. For example, in the case of modelling the Earth, a DGGS can include land / water boundary data representing land and water areas of the Earth. In addition, other data may be associated with the cells of the DGGS.
[0024] Generally, any data can be associated with a cell. When the model is built, and data is associated with the model, a wide range of analyses can be undertaken when it is known what cells are contained within a region (e.g., a region defined by a polygon). Such analyses may be relatively simple to relatively complex.
[0025] Typically, when an analysis is undertaken, it is undertaken relative to an area of interest on the surface of the sphere. An area of interest can be defined as a polygon with vertices overlaid on the DGGS / Earth. Figure 1A shows that vertices V1-V6 can be defined using known coordinates (e.g., lat / long) with the boundary of the polygon defined by great circle arcs L1-L6 between the vertices.
[0026] In a representative initial example of how data associated with cells is searched, for the purposes of introduction, the polygon comprised of vertices and lines defines an enclosed area where the user wishes to determine the average surface temperature of the cells within the polygon.
[0027] As shown in Figure 1 B, the polygon is overlaid with the triangular cells of the DGGS. As can be seen, the polygon boundaries intersect various cells, with portions of each cell inside and outside the boundaries. Some cells have relatively small, enclosed areas that are inside the polygon boundary and larger areas that are outside the polygon boundary (e.g., Figure 1 B,cell (3,1,3)) whereas other cells may be fully contained within the polygon boundary or have relatively large areas that are inside the polygon boundary (e.g., Figure 1 B, cell (2,2,2)).
[0028] Data may be associated with each cell which, for example, Figure 1B shows temperature ranges from 1 °C in Greenland to 23-29 °C in the southern United States. As can be seen, the region of interest minimally intersects with the cells of the southern United States. In total, 12 cells are intersected by the polygon.
[0029] As such, at a first level the average temperature in the zone of interest can be calculated to be the sum of the average temperature of each cell divided by 12. As the polygon minimally intersects the southern US cells, the average temperature calculation is distorted as the three southern values represent 25% of the dataset but only about 1% of the area of interest.
[0030] As a result, to improve the accuracy of data analysis, data systems will use finer grids with a larger number of smaller cells to improve the average temperature calculation by more accurately only using data that substantially correlates to the area inside the area of interest. While the results will be more accurate, the resulting calculations will take longer and require greater computing power as explained below.Hierarchal Geocoding
[0031] Hierarchal geocoding is a process of organizing cells at various levels of resolution to obtain information refined to the level of resolution.
[0032] That is, in the above temperature example, any of the cells of Figures 1A and 1B may be sub-divided into a larger number of cells having a smaller area. Temperature data in each of the cells may be different; hence, if temperature data of a larger number of smaller cells is averaged, the calculation is more closely aligned with the true zone of interest.
[0033] In the example of Figure 1B, there are no cells fully contained within the boundary. However, in a typical analysis at a typical resolution, there will be many cells that are fully contained within the boundary.
[0034] Figure 1C shows a particular example of equilateral triangular DGGS cells, including a parent cell and multiple resolutions of child cells and the relative size relationship between each respective degree of resolution. As shown, parent cell P is shown arbitrarily as resolution 1 (1°). Additional resolutions progressively become smaller triangles that are fully contained within theparent cell. As such, each triangle at resolution is constructed by inserting a vertex at the midpoint of each edge of the parent P.
[0035] A typical prior state-of-the-art process of evaluating if cells are inside or outside the boundary is shown in Figures 2A-2C in a DGGS having square cells. The process includes the following:• Starting at the coarsest level of resolution of the cells:i. Individually test each cell to determine if all boundary lines BL intersect a cell of the DGGS. As shown in Figure 2A, the polygon PG intersects 14 cells, each referred to as a boundary cell BC. Cell X does not intersect the boundary but is fully within the boundary. A number of cells are shown to be fully outside the boundary.ii. Next, subdivide the boundary cells (i.e., increase cell resolution) and ignore the interior cell X. As shown in Figure 2B, each boundary cell has been sub-divided into 4 child cells.iii. Next, test each sub-divided cell to determine if the boundary line intersects with a sub-divided cell. As shown in Figure 2C, sub-divided cells outside the boundary are white. There are 7 sub-divided cells fully inside the boundary and shown as IC; 1 coarse cell X (not sub-divided) is fully within the boundary; and 26 sub-divided cells intersect the boundary BC.
[0036] Following this process, more refined and increased amounts of data can be accessed that more closely correlate to the area of interest.Scanline Process for Determining if a Cell Intersects a Boundary
[0037] Past processes for determining if a cell at a particular resolution intersects a boundary typically utilize a scanline process.
[0038] A typical scanline process is illustrated with reference to Figures 3A-3I.
[0039] As shown in Figure 3A, a zone of interest is illustrated by a triangle overlaid on the cells and defined by 3 vertices and 3 boundary lines L1-L3. In this example, the cells are laid out asa series of rows and columns. To determine which cells intersect and are within the boundary lines, the following process is followed:• Start with 1 vertex point (V1) and identify which cell contains vertex 1.• Follow boundary lines BL1 and BL2 away from V1, and for each cell that intersects the boundary line, label that cell (e.g., cells C1-C6) as intersected, as shown in Figures 3B and 3C.• For cells that are within the boundary but not intersecting (Figure 3D), these cells are identified as interior cells IC (e.g., IC1).• As shown in Figure 3E, this process includes conducting an analysis of each row (or column) wherein, for each row and starting at the cell that does intersect with the boundary, each cell of a row is successively reviewed in one direction. For example, cell C7 is determined to be intersecting BL2 and is the start of a new row R4. As such, the adjacent cell to right (as an example direction) is checked and cell C8 is determined to not intersect with boundary BL1; hence, must be within the boundaries. The next cell to the right is checked to see if intersects with a boundary. If no boundary is encountered, it is assumed that this cell is an interior cell and the process is repeated until another boundary cell is reached. In our example, cells C9 and C10 do intersect boundary BL1, whereas cell C* does not. The identification of cell C* stops the analysis for that row. Thus, each of cells C7-C10 are included as shown in Figure 3F.• The process is repeated for each row, as shown in Figures 3G, 3H and 3I, to identify all cells within the polygon or bordering the polygon.
[0040] While the foregoing can effectively locate all cells within a polygon, there is a need to be able to undertake such a review to:• Enable compression: The representation of the polygon as a raster should be as small as possible to enable efficient storage and transmission of raster data, as well as increase the speed of operations performed on rasters.• Enable multiple resolutions: Because of the trade-off between speed and accuracy when working with different raster resolutions, it is useful to be able togeocode at any resolution. However, if two polygons are geocoded at different resolutions, their rasters cannot be directly compared. Therefore, it is useful to be able to change the resolution of an already geocoded polygon to enable operations with other polygons geocoded at the same resolution.• Enable other operations to be conducted including Boolean operations.Interval Based Indexing and Compression
[0041] Turning to Figure 4A, the labelling / ordering of cells can be indexed according to a logical numbering scheme that, as explained below, can enable efficient compression of cells, maintain efficiency across multiple resolutions and enable efficient operations.
[0042] Figure 4A shows a numbering scheme for a 7-row by 9-column grid, with cells labelled 0-62, with a left to right sequential numbering across rows and a sequential number starting a new row. For the purpose of explanation, the grid is the cells of a DGGS at a first resolution. As with the cells of Figures 3A-3I, a simple triangular polygon is overlaid.
[0043] As shown, the triangle polygon of Figure 4B encompasses cells having numerical identifiers 9, 16, 17, 23, 24, 25, 30, 31, 32, 33, 37, 38, 39, 40, 45, 46 and 52 which includes cells touching the polygon boundary and cells within the polygon boundary.
[0044] For the purpose of efficiency when conducting analysis of cells, it is generally more efficient if a smaller number of cells are subject to processing at a given time. That is, processing 100 cells will take more time than processing 10 cells. Accordingly, the ability to compartmentalize groups of cells into packages that can be processed as a package instead of individual cells can greatly enhance processing efficiency.
[0045] For example, as shown in Figure 4G, cells that are successively numbered, for example, cells 16-17, 23-25, 30-33, 37-40, and 45-46, these successively numbered cells can be packaged into a group, package or interval of cells.
[0046] As shown in Figures 4D(1) and 4(D), the analysis pathway to create cell packages can be conducted using a different process to that described in Figures 3A-3I utilizing different evaluation and / or movement rules. Figures 4D(1) and 4D(2) compare a row-based analysis vs an arbitrary indexing of the cells, optimal for this particular polygon. However, this choice is arbitrary and the best choice depends on the properties one may wish to achieve.
[0047] Figure 4D(1) shows a row-by-row analysis as described above, where the cell numbers are defined based on adjacent cells on a line which in this case is a row of a grid. Figure 4D(2) shows a different ordering of the cells where a series of movement rules can identify all the cells touching the polygon. In this example, the rules may evaluate each cell around a cell that intersects a boundary line to identify all cells on the boundary and those within the boundary. In the example of Figure 4D(2), the movement along the boundary lines (to determine if a boundary line intersects a cell) requires testing of at least 4 adjacent cells to confirm or reject the inclusion of a cell, which is computationally inefficient.
[0048] Importantly, other ordering techniques can be advantageous to ultimately improve efficiency, compression and multiple-resolution properties.Multi-Resolution
[0049] Figures 5A and 5B show a polygon on a grid at different resolutions. It would be highly advantageous if there were a simple relationship between the geocoding of the polygon at one resolution and the geocoding of the same polygon at a different resolution wherein cells are numbered such that there is such a simple relationship.Operations
[0050] When conducting spatially-correlated data analysis of cells, Boolean operations are often required and conducted to enable a variety of conclusions from various types of data. These Boolean operations are illustrated with reference to Figure 6A, showing two different polygons PG1 and PG2 both overlaid on the cells and all cells from both polygons identified.
[0051] Typical Boolean operations can be conducted with respect to these polygons, including union (i.e., the set of all cells overlapping either polygon or both), intersection (i.e., the set of only those cells overlapping both polygons) and difference (i.e., all cells overlapping one polygon but not the other).
[0052] For reference, Figure 6B illustrates and identifies the cells of an intersection analysis. Importantly, as explained below, operations are ideally conducted with improved efficiency. In this case, polygonl is defined by polygon boundary lines PG1 and polygon2 is defined by polygon boundary lines PG2. The cells listed as the intersection cells are the cells common to both polygons.Interval Based Polygon Geocoding for DGGS
[0053] With the foregoing background, in accordance with one embodiment, interval-based polygon geocoding (IBPG) methods are now described.
[0054] The IBPG methods are described referencing the previously described concept of a polygonal area of interest on a sphere and a DGGS representation of that sphere.
[0055] As shown in Figures 7A and 7B, and discussed above, the area of interest may extend over a boundary between adjacent cells.
[0056] In accordance with one embodiment, for any polygon in more than one base cell, the polygon may be cut as shown in Figures 7C and 7D.Indexing Triangle Grid Cells via Interval Based Indexing
[0057] The foregoing examples have generally described indexing of square cells of DGGS.
[0058] Indexing of a DGGS having triangular cells is described with reference to Figures 8A-8Q. Figure 8A shows indexing of a representative triangular parent cell PC and child cells CC, e.g., CC1, and CC2 cells. Parent cell is shown with a boundary length R1, CC1 is shown with boundary length R2 (50% of R1) and CC2 is shown with boundary length R3 (50% of R2; 25% of R1). Each of the 16 child cells may be labelled according to rows and columns of the child cells within the parent with 3 coordinate numbers. In this example, starting at the lower left vertex V1 , there are 4 successive rows extending upward and to the right and labelled as rows 0-3 and being the first number of the index identifier shown in brackets (e.g., (0,3,3). The second number is derived from the rows extending downward from the top vertex V2 and the third number is derived from the rows extending upward and to the left from the lower right vertex V3. As such, each of the 16 CC2 cells has a unique index identifier.
[0059] As in previous sections, Figure 8B shows a representative polygon having 6 vertices (PV1-PV6) overlaid on the cells. The position of each of the 6 polygon vertices can be defined in terms of the cell that it is contained within and refined by its relative position to a row edge as shown in Figure 8C. For example, V4 is 40% into the 3rdrow from V1 expressed as 3.4; 80% into the 3rdrow from V2 expressed as 3.8 and 80% into the 1strow from V3 expressed as 0.8. Thus, PV4 is expressed as (3.4, 3.8, 0.8).
[0060] Each line of the polygon can be expressed by an equation, parametrically defining the points along the line:+ r bZbwhere parameter t is the percentage of the position between two points as shown in Figure 8D and expressed as:a—x.t= -X2~X1
[0061] Figure 8E is an example showing the calculation of t between 2 points where the boundary point BP1 is calculated as being 20% from PV3 to PV4, where t=0.2. This calculation is referred to as linear interpolation (or colloquially, “lerp”) and allows the identification of cell boundaries along a line.
[0062] To solve for the PV3 and PV4 cells, all t values at all integer coordinates between the two points are calculated which in this example, is from (2.2, 3.6, 2.3) to (3.4, 3.8, 0.8). In the first coordinate we go from 2.2 to 3.4 and we pass the integer 3 so we solve for t using the above equation where a=3, x1=2.2, x2=3.4. In the second coordinate we go from 3.6 to 3.8 so we don’t pass any integers. In the third coordinate we go from 2.3 to 0.8 and we pass 2 and 1 so we solve fort using the above equation where a=2, x1=2.3, x2=0.8 and also fort where a=1, x1=2.3, x2=0.8. This gives us three t values 0.666, 0.2, 0.866, respectively. From here we sort the t values to obtain 0.2, 0.666, 0.866. Since t=0.2 is first in our sorted list, and we solved for it using the third coordinate and to get from 2.3 to 0.8 we go down in the third coordinate, this means we decrement the third coordinate of our cell number to go from (2,3,2) to (2,3,1). We process each subsequent t value in the sorted list similarly. As shown in Figure 8F, PV3 to PV4 cells are successively (2,3,2) -^(2,3,1) -^(3,3,1) -^(3,3,0).
[0063] As shown in Figure 8G, the identification of all cells touching the polygon can be determined.
[0064] As shown in Figure 8H, depending on the resolution, cells in the interior that do not touch the boundary may not be identified. As discussed above, this may be addressed by a scanline.Sequential Index and Compression
[0065] As shown in Figures 8J and 8K, if the indexing of cells is patterned to be numerically sequential for cells bordering another cell either through an edge or vertex, as shown in Figure 8K, one can “walk” through successive cells either across edges or common vertices as shown by walk-line WL without any discontinuities or self-intersections. To progress to higher resolutions, a copy of Figure 81 is placed into each cell at the current resolution. When going from the resolution in 81 to the resolution in 8J, the copy of 81 placed in cell 1 of 8I must be flipped upside down and the copy of 8I placed in cell 3 of 8I must be flipped left to right. This index is referred to herein as the “sequential index” or “interval based index”.
[0066] Figures 8L and 8M show the indexing systems with the example polygon overlaid. Figure 8L shows both a coordinate indexing system and sequential indexing system overlaid and Figure 8M shows both indexing systems with boundary cells of a generic overlaid polygon identified.
[0067] Importantly, the sequential index can provide computational efficiencies. The reasoning for computational efficiency and processes of converting barycentric coordinates as shown for example in Figures 8A-8G to a sequential index are described below.
[0068] Figures 8N and 80 show the cells on the boundary for the two indexing systems. Figure 8N shows that the number of cells can be compressed to a smaller group with the sequential indexing system; in this case, the 11 border cells being reduced to 2 sets, intervals or runs of grouped namely cells 4-11 and 13-15.
[0069] Importantly, as shown in Figures 8P and 8Q, based on a sequential index, all cells subsequently numbered to a cell on the boundary are necessarily within the polygon, until another boundary cell is encountered. For example, cell 20 is a boundary cell and shares an edge with cell 21, cell 21 shares an edge with cell 22 and cell 22 shares a vertex with cell 23 which is a boundary cell. As such, cells 20-23 can be considered as a single index group. Similarly, cell 34 is boundary cell with the next boundary cell being cell 38. Hence, cells 34-38 are a single index group.
[0070] Figure 8Q also shows labelling of the cells of the polygon into a list wherein all cells on the boundary are listed and cells that start an index group where the cells are inside the polygon have a negative sign.
[0071] Figure 8R graphically shows a parent cell P having vertexes V1 , V2 and V3 and 16 child cells (CC1-CC16) and the sequential numbering process for numbering each child cell starting at vertex V1. As shown, sequential numbering numbers child cells in a repeating pattern where cells 1-3 are numbered across a common boundary CB and cell 4 is numbered across a common child vertex CCV with a common boundary shared with the 2ndcell of the quartet of cells. This pattern of indexing is repeated across all child cells. While this indexing process is preferred, it is understood that variations may be introduced to provide for exceptions that may exist within a particular parent / child cell configuration. Figure 8R also shows a generic border B overlaid with the DGGS.
[0072] Figure 8S is a flowchart showing a representative process for identifying indexed cells as shown in Figure 8R to determine if cells are overlapping or within a border. Figure 8T is a representation of a border bracket routine (BBR). The BBR is a process utilized to set one of 3 conditions, including:• NIL - a condition of cells that are not an OPEN or CLOSED condition and can follow a CLOSED condition.• OPEN - a condition where a previous cell has a border B (i.e. , analogous to an “opening” bracket) and a second “closing” bracket has not been encountered. When the condition is OPEN, all cells up to and including a “closing” bracket are marked M.• CLOSED -a condition following an OPEN condition and is triggered when a cell has encountered a closing bracket.
[0073] After a border has been overlaid in a grid, all cells are checked to determine if they are on or within the border. Starting at the lowest child cell number 80, the first cell n is checked to see if it has an overlaid border 82. If yes, the cell is marked, n is incremented by 1 and the BBR condition is set to OPEN 84.
[0074] Thereafter, n is incremented by 1 86 and the n+1 cell is checked for a border 88. If no at 88, the cell is marked and BBR condition is set to OPEN 84. If yes at 88, the cell is marked and the BBR condition is set to CLOSED 90. All CLOSED cells are tabulated as a group 92.The process determines if all cells have been checked 94, and if yes, at 94 stops the process 98. If no at 94, n is incremented by 1 96 and the new n cell is checked for a border 82.
[0075] If at 82 cell n does not have a border, cell n is discarded and the BBR condition is set to NIL 104. n is incremented by 1 102 and the new n+1 cell is checked for a border 100. If yes at 100, the cell is marked and the BBR condition is set to OPEN 84 and the process continues. If no at 100, the cell is discarded and the BBR condition is set to NIL 104 and the process continues.
[0076] Figure 8T further illustrates a process for progressively checking how each of the NIL, OPEN and CLOSED conditions are triggered. As shown, a NIL condition exists for the first 5 cells being checked. The 6thcell encounters a border which triggers the OPEN condition. The 6thcell is marked and the subsequent 7th-9thcells maintain the OPEN condition and are marked. The 10thcell encounters a border which triggers marking that cell and the CLOSED condition. The 11th- 14thcells are NIL condition cells. The CLOSED cells (e.g., cells 6-10) are tabulated as a group.
[0077] Further, as shown in Figure 8Q, there are 39 cells bordering or within the polygon. Using the compression method described, this is compressed into 3 index groups (6 cells for the starts and ends of each group).Multi-Resolution
[0078] The compression method is further described with a representative example for a triangle parent with a total of 64 child cells.
[0079] Figures 9A-9C show how to change the resolution of a geocoded polygon. To go to a coarser resolution, we can simply divide the start and end of each index group by 4 and round down. When we do this, we may need to compress adjacent index groups. As shown in Figure 9A, a total of 64 cells are indexed as cells 0-63, with cells 16-41, 43-47 and 53-60 shown as 3 groups of cells touching or within the polygon. Figure 9C shows the same triangle at a coarser resolution where the sequential index for each cell has been divided by 4, thus showing 16 indexed cells. Figure 9C shows that additional compression is possible whereby dividing the cell numbers by 4, the 3 groups of cells can be reduced to 2 groups of cells. For example, groups [4-10] and [10-11] from Figures 9A and 9B can simply be represented as [4-11] as shown in Figure 9C. Since this procedure for changing the resolution of a geocoded polygon is trivial, we can perform operations and analysis on polygons at any resolution we want allowingthe user to choose the trade-off they want between speed and accuracy when working with rasters. This also allows Boolean operations between two polygons geocoded at different resolutions because we can change the resolutions to match.Barycentric to Sequential Indexing Conversion Process
[0080] As described above, cells may be indexed with different indexing systems, including barycentric coordinates and sequential numbering coordinates.
[0081] In various applications, it is desirable to convert barycentric coordinates to sequential numbering coordinates in order to take advantage of the various processing advantages described herein.
[0082] For the purpose of describing this process, Figures 9D(A)-9D(F) illustrate aspects of a sequential indexing method and how barycentric coordinates can be converted to sequential numbering coordinates and vice versa.
[0083] Figure 9D(A) shows a basic sequential indexing process for a parent cell PC having 4 child cells CC and the movement steps to number child cells 0, 1, 2 and 3 as described above.
[0084] Figure 9D(B) shows how the process is repeated with 4 parent cells (labelled 1-4) wherein after the child cells within a first parent cell 1 are numbered 0-3 (as in Figure 9D(A)), each of the 3 adjacent parent and corresponding child cells are indexed following a rule pattern for indexing. As shown, sequential numbering within each parent cell includes numbering the first 3 child cells (child cells 0, 1 and 2) across abutting borders and numbering the 4thchild (child cell 3) cell across a vertex.
[0085] As can be seen in Figure 9B, parent cells 1 and 3 are similar in terms of orientation, in that the 0thcell is at the “bottom left” of the parent cell, whereas for parent cell 2, the 0thchild cell is at the “top left” and for parent cell 4, the 0thchild cell is at the “bottom right”. As such, relative to parent cell 1 , the orientation of parent cell 2 is “rotated” 90 degrees relative to parent cell 1 and the orientation of parent cell 4 is “reflected” across a vertical mirror axis relative to parent cell 1.
[0086] Figure 9D(C) illustrates the desired result where all child cells are indexed as a continuous sequence of numbers and Figure 9D(D) graphically shows the “links” between parent cells across borders for parent cells 1, 2, 3 and across a vertex for parent cell 4.
[0087] In comparison, Figure 9D(E) shows cells indexed with barycentric coordinates with an arbitrary polygon overlaid. Each of the cells that touch the polygon are listed.
[0088] Figure 9D(F) shows the same cells with sequential indexing with the arbitrary polygon overlaid.
[0089] For barycentric to sequential index conversion, for reference, Figure 9E shows representative cells with barycentric coordinates (triplet number), a sequential index (single digit number) and a base coordinate (middle number). The smaller triangle on the upper left shows a “barycentric indexing map”, which is referenced to illustrate a process to convert barycentric coordinates to sequential coordinates, as will be explained below. The barycentric indexing map is the simplest unit of a barycentric index showing 4 child cells of a basic parent cell labelled with barycentric coordinates.
[0090] That is, for the 3 sides of the basic child cell, namely sides V1, V2 and V3, each cell of the basic triangular cell has a basic barycentric coordinate namely (0,1,1), (1,1,1), (1,1,0) and (1,0,1).
[0091] Also, for reference, Figure 9F illustrates a “walking pattern” for sequential indexing across child and parent cells and how a sequential index is repeated across 4 parent cells and 16 child cells and the relative “orientation” of a sequential index across the child cells of each parent cell. That is, Figure 9F(A) shows a basic numbering or walking pattern within a basic triangular cell having child cells 0, 1, 2 and 3 where the numbering from child cells 0 to 1 is across a border, numbering from child cells 1 to 2 is across a border, and numbering from child cells 2 to 3 is across a vertex. Figure 9F(B) also shows how the numbering or walking pattern is continued between parent cells (also, see Figure 9D(F)).
[0092] A process of converting barycentric coordinates to a sequential index is described with reference to Figure 9E, Figure 9F and Figure 11A with representative examples:
[0093] For cells with barycentric coordinates, an example cell may have coordinates (5,4,5) at resolution 3. In one embodiment the process to convert these coordinates to sequential coordinates includes the following steps as shown in Figure 11a:i. Convert each number 11a of the barycentric triplet to binary numbers 11b. The number of bits corresponds to the resolution. For example, (5,4,5) becomes (101, 100, 101) as per the binary number for each ofthe numbers of the triplet. The number of bits also corresponds to the resolution.ii. Interleave the bits into groups of triplets 11c from most significant to least significant. For example, (101, 100, 101) becomes (111,000,101). That is, the first number of the binary number for each of the triplet numbers is combined as one number of a new triplet.iii. Process each triplet one at a time from left to right 11 d. In this example, 111 is processed first.iv. Each cell is then processed to determine how it maps relative to the basic barycentric index. That is, each triplet is checked to determine which of the 4 basic barycentric (“basic index”) coordinates it may be.1. Check if a triplet is 101 (i.e., cell 3 in the basic index), or 111 (ie., Cell 1 in the basic index) which are the two cells of the basic index that are “oriented” differently to the cells 0 and 2 of the basic index.a. If the triplet is 101 (11e), the system swaps the first and last bits of each subsequent processed triplet number (11f).b. If the triplet is 111 (11g), the system inverts bits and swaps first and last bits of each subsequent triplet number (11h).v. When each triplet is processed, each processed triplet 11i is converted to the corresponding cell using the basic index. That is:1. 011 maps to 02. 111 maps to 13. 110 maps to 24. 101 maps to 3.5. Optionally, the final linear number can be converted to decimal.a. For example, (111, 111, 101) is converted to 113 vi. Interpret in base 41. For example 1134A2 * (1) + 4A1 * (1) + 4A0 * (3) = 23
[0094] Intuitively, if the triplet is 111 , the triangle is upside down so the coordinates are changed for subsequent triplets (which are the hierarchical or contained next triangles).
[0095] If the triplet is 101, the triangle is flipped left to right and we change coordinates for subsequent triplets (which are the hierarchical or contained next triangles).
[0096] Three examples are as follows:i. Example 11. (5,4,5) at Resolution 32. (101, 100, 101)3. (111, 000, 101)4.a. Iteration 1i. Input: 111, 000, 101ii. Output: 111, 111, 010b. Iteration 2i. Input: 111, 111, 010ii. Output: 111, 111, 101c. Iteration 3i. Input: 111, 111, 101ii. Output: 111, 111, 101 5.i. Input: 111, 111, 101 ii. Output 1 , 1 , 3 6. 4A2 * (1) + 4A1 * (1) + 4A0 * (3) = 23
[0097] Example 2i. (2,3,1) at Resolution 2ii. (10, 11, 01)iii. (110, 011)iv.1. Iteration 1a. Input: 110, 011b. Output: 110, 0112. Iteration 2a. Input: 110, 011b. Output: 110, 011v.1. Input: 110, 0112. Output 2, 0vi.1. 4A1 * (2) + 4A0 * (0) = 8vii. Example 31. (15,3,13) at Resolution 42. (1111 , 0011, 1101)3. (101, 101, 110, 111)viii.1. Iteration 1a. Input: 101, 101, 110, 111b. Output: 101 , 101, 011, 1112. Iteration 2a. Input: 101 , 101, 011, 111b. Output: 101 , 101, 110, 1113. Iteration 3a. Input: 101 , 101, 110, 111b. Output: 101 , 101, 110, 1114. Iteration 4a. Input: 101 , 101, 110, 111b. Output: 101 , 101, 110, 1115.a. Input: 101 , 101 , 110, 111 b. Output 3, 3, 2, 1 6. 4A3 * (3) + 4A2 * (3) + 4A1 * (2) + 4A0 * (1) = 249
[0098] Similarly, the inverse steps can be followed to convert a sequential index to barycentric coordinates.
[0099] Figure 11 B shows an example of determining if a boundary overlaps a series of indexed cells or not whether in barycentric or indexed coordinates. For a representative case with a list of barycentric cells that intersect with polygon boundary 12, the index conversion is applied to obtain a corresponding list of sequential index cells 12a. The list is sorted in ascending order 12b. Starting with the lowest cell number, 12c, that cell is appended as the first interval in the list (e.g., [(1,?)]. Each successive cell is then tested to determine if it is sequential with its preceding number, in order to either create a list of successive numbers or to determine that a successive number is not successive. If a number is not successive, a current list of successive numbers is closed, and the process is repeated to determine a next list of successive numbers until all numbers of the list have been processed.
[0100] Figure 11C shows an example of processing lists at different resolutions. In this example, a list of cells can be easily processed at a different coarser resolution, simply by dividing the start and end numbers of each group of numbers by 4 (see Figures 11C and 13a) and rounding each number down to the nearest whole number.Operations
[0101] The foregoing processes of index conversion (Figure 11a), boundary determination (Figure 11 B) and resolution adjustment (Figure 11C) allow for rapid operations to be determined.
[0102] Figures 10A-10B show how union, intersection and difference operations can be conducted using the sequential index. Figure 10A shows two overlapping polygons within a grid. The polygon PG2 includes cells 4-11 and 13-15 and the polygon PG1 includes cells 1-5 and 7. A linear comparison of the two groups is shown in Figure 10A. Figure 10B shows how the operations of union, intersection and difference can be calculated from the linear comparison.
[0103] Union is all cells within either polygon and thus includes 1-11 and 13-15.
[0104] Intersection is cells common to both polygons and thus includes 4-5 and 7.
[0105] Difference is all in the polygon PG2 but not in the polygon PG1 and thus includes 6, 8-11 and 13-15.
[0106] As can be appreciated, the ability to simply compare groups to conduct these operations is substantially simpler than prior methods where all cells in the geocoded region would have to be tested individually (see Figure 6A and 6B) in a barycentric coordinate system.Benchmarking
[0107] A comparison of hierarchal indexing with interval based indexing is compared in Table 1 for geocoding and Boolean operations.Table 1-Comparison of Hierarchal Indexing with Interval Based Indexing for Geocoding and Boolean OperationsHierarchal Indexing Interval IndexingTime TimeGeocoding (1000 random 8.99 ms / polygon 0.48 ms / polygon polygons)Boolean Operations (1000 477-518 ps / operation 0.13-0.6 ps / operationrandom pairs)
[0108] In the first test, 1000 random polygons were geocoded using both the hierarchical method and our new interval method. The reported values are the average time to geocode a single polygon for each method. In the second test, 1000 random pairs of polygons were geocoded, and a Boolean operation was performed on the pair. The reported values are the average time to perform the Boolean operation. In this test, Boolean intersection was the fastest for both methods and Boolean union was the slowest so the lower end of the range is intersections and the higher end unions.
[0109] The results indicate that geocoding using the interval indexing is approximately 20 times faster than hierarchal indexing and that Boolean operations are 500-1000 times faster. Interval indexing in the typical case results in compression of approximately 100x, i.e., 1% of the total polygon cells are required to store the polygon as a set of index groups.
[0110] In summary, interval-based polygon geocoding enables:• Rapid geocoding• Compression• Multi-resolution• Fast Boolean operations.
[0111] Although the present disclosure has been described and illustrated with respect to preferred embodiments and preferred uses thereof, it is not to be so limited since modifications and changes can be made therein which are within the full, intended scope of the disclosure as understood by those skilled in the art.
Claims
CLAIMS1. A method of converting a barycentric coordinate to a sequential index coordinate comprising the steps of:for a DGGS cell having barycentric coordinates and 1st, 2ndand 3rdbarycentric coordinate numbers (BCNs):a) for each of the 1st, 2ndand 3rdBCNs, convert to binary number bits;b) interleave the binary number bits from step a) into groups of triplets;c) process each triplet to map a triplet to an index map, the index map having 4 designators, where each triplet is mapped to one of the 4 designators to form a converted triplet; and,d) interpret each number from step c) in base 4 to form a sequential index coordinate.
2. The method as in claim 1 where the index map maps barycentric coordinates (0,1,1) to O, (1,1,1) to 1, (1,1,0) to 2 and (1,0,1) to 2.
3. The method as in either one of claim 1 or 2 where the steps are reversed to convert sequential index coordinates to barycentric coordinates.
4. The method as in either one of claim 1 or 2, wherein for a list of sequentially indexed coordinates, determining if cells within the list of indexed cells are sequential or not, further comprising the steps of:sorting the list of indexed cells into an ascending or descending order; and, starting with the lowest or highest cell number, testing each successive cell to determine if it is sequential with its preceding cell to either create a list of successive numbers or to determine that a successive number is not successive.
5. The method as in claim 4 wherein if a number is not successive, further comprising the step of closing a current list of successive numbers.
6. The method as in claim 5 further comprising the step of repeating the process to determine a next list of successive numbers until all numbers of the list have been processed.
7. A method of indexing cells of a discrete global grid system (DGGS) and identifying cells of interest within the DGGS, comprising the steps of:within a DGGS model having a plurality of triangular parent and child cells, wherein a parent cell contains the child cells:indexing each child cell within the parent cell as a numerical sequence of successive child cells sharing a common boundary (CB) or common child vertex (CCV) wherein the numerical sequence is a progression of numbers starting at a parent cell vertex wherein indexing includes:(a) numbering successively first CC1, second CC2 and third CC3 child cells wherein CC1 and CC2 share a common boundary and CC2 and CC3 share a common boundary;(b) numbering successively a fourth CC4 child cell sharing a common child vertex CCV with CC3 and a common child cell boundary with CC2 and wherein CC1-CC4 is a child numbering set;(c) repeating steps a)-c) n times to successively number each child cell of the parent cell wherein numbering of a successive child number set is initiated across a common child cell boundary or common vertex, and,(d) overlaying a first border on the DGGS wherein the first border overlays one or more child cells and determining child cells overlapping and within the first border.
8. The method as in claim 7 wherein step d) includes: identifying a first cell intersecting the first border; and, checking each successively numbered cell and determining if each successively numbered cell is overlapping or within the first border.
9. The method as in claim 8 further comprising the step of: for each successively numbered cell determined as overlapping or within the first border, marking each successively numbered cell as a group of cells.
10. The method as in claim 9 further comprising the step of: tabulating each group of cells as single data points and wherein the single data points collectively define cells overlapping or within the first border.
11. The method as in claim 10 further comprising the step of e) overlaying a second border wherein the second border overlays one or more child cells and determining cells overlapping and within the second border.
12. The method as in claim 11 wherein step e) includes identifying a first cell intersecting the second border; and, checking each successively numbered cell and determining if a successively numbered cell is overlapping or within the second border.
13. The method as in claim 12 further comprising the step of: for each successively numbered cell determined as overlapping or within the second border, marking each successively numbered cell overlapping or within the second border as a second border group of cells.
14. The method as in claim 13 further comprising the step of: tabulating each group of cells as single data points and wherein the single data points collectively define cells overlapping or within the second border.
15. The method as in claim 14 further comprising the step of conducting a Boolean union operation on the cells defined by the first and second borders utilizing the single data points.
16. The method as in claim 14 further comprising the step of conducting a Boolean intersection operation on the cells defined by the first and second borders utilizing the single data points.
17. The method as in claim 14 further comprising the step of conducting a Boolean difference operation on the cells defined by the first and second borders utilizing the single data points.
18. A digital global grid system (DGGS) comprising:a DGGS model defined by triangle parent and child cells at a plurality of resolutions wherein each resolution is defined as a degree of resolution and wherein a degree of resolution of a child cell is defined by a triangular cell within the parent cell and having a border length 50% of a border length of the parent cell;wherein a parent cell contains four child cells; and,wherein each child cell within a parent cell is indexed as a numerical sequence of successive child cells sharing a common boundary (CB) or common child vertex (CCV) wherein the numerical sequence is a progression of numbers starting at a parent cell vertex and wherein child cells within a parent cell are numbered successively as first, second and third child cells wherein the first and second child cells share a common boundary and the second and third child cells share a common boundary; and a fourth child cell shares a CCV with the third child cell and a common child cell boundary with the second child cell and wherein the first, second, third and fourth child cells is a child numbering set.
19. The system of claim 18 further comprising a plurality of adjacent parent cells and wherein sequential indexing of child cells within each parent cell progresses between adjacent parent cells.
20. A method of marking and compressing discrete global grid system (DGGS) data comprising the steps of:in a DGGS system having n triangular child cells sharing common borders and vertexes and being numbered sequentially through the common boundary and vertexes and where a border is overlaid at least one child cell;a) starting a cell n, checking if cell n has an overlaid border, andi) if yes marking cell n and setting a border condition to open; and,ii) if no, discarding cell n and set border condition to nil;(1) increment n by 1 and check if cell n+1 has an overlaid border;(a) if yes, performing step a) i)(b) if no, repeating step a) ii);b) if step a) is yes, incrementing n by 1 and checking if cell n+1 has an overlaid border, andi) if yes, setting a border condition to closed; and(1) marking cell n+1 and change border condition to closed; and(2) tabulate closed cells as a group;ii) if no, repeating step a) i).
21. The method as in claim 20 further comprising the step of confirming if all cells have been checked and if no, increment n by 1 and repeat step a) and, if no, stop.
