CN111446968A - Multilevel Compression Method for Vector Spatial Data - Google Patents

Abstract

The disclosure belongs to the technical field of spatial information, and particularly relates to a method for multistage compression of vector spatial data. The spatial vector data compression method provided by the disclosure takes visualization of basic geographic spatial data as an application scene, sets the minimum resolution distance meeting the requirement of multiple scales as a precision requirement, realizes multilevel compression processing of vector spatial data, and further improves the data storage efficiency by combining grid filtering and binary offset storage.

Description

Multilevel Compression Method for Vector Spatial Data

technical field

The present disclosure belongs to the technical field of spatial information, and in particular relates to a method for multi-level compression of vector spatial data.

Background technique

With the development of earth observation, navigation and positioning equipment and Internet technology, the production capacity of various vector geospatial data has gradually increased, which has led to a rapid increase in the amount of data storage. The efficiency of functions such as storage, query, network transmission, and visualization presents new challenges.

As we all know, data compression can reduce the amount of data, improve the efficiency of all aspects of data processing, and better support spatial data application scenarios including mobile Internet geographic information systems.

Vector data model is one of the most basic entity representation models of geographic elements in geographic information systems. It abstractly represents geographic entities as geometric entities such as points, lines, and surfaces under a given geographic spatial coordinate system, and realizes the expression of geographic entities by recording the spatial coordinates of feature points and setting expression rules for point sets. The target description is complete and the spatial relationship is easy to obtain.

SUMMARY OF THE INVENTION

The purpose of the present disclosure is to provide a multi-level compression method for vector space data. The specific process of this method is as follows:

Divide the space expressed by geographic coordinates step by step in the horizontal and vertical directions to obtain a multi-level grid that divides the geographic space;

Determine the position of the coordinate point of the data to be compressed in the grid, and continue to divide the grid to replace the coordinate value of the coordinate point of the data to be compressed with the coordinates of the center point of the divided grid; wherein The coordinate deviation between the center point of the divided grid and the coordinate point of the data to be compressed conforms to the preset value of the multi-level scale precision;

Taking a single data file to be compressed as a unit, set a local coordinate reference system, and calculate the binary coordinate deviation of the coordinate point of the data to be compressed and the center point of the grid after the division to obtain a binary offset;

The geographic coordinates of the vector spatial data are stored at the binary offset.

Further, when the division is performed in the horizontal and vertical directions, the left and right or upper and lower subspaces after division are identified by binary.

Further, the preset value is the minimum distinguishable distance expressed under the condition of multi-scale visual lossless.

Further, the position of the coordinate point of the data to be compressed in the grid is determined, and the grid is continuously divided, so that the coordinates of the center point of the divided grid are used to replace the coordinates of the coordinate point of the data to be compressed. Value steps include:

Obtain the coordinate data of the coordinate points of the data to be compressed;

Obtain the minimum and maximum values of the latitude coverage of the grid containing the coordinate points of the data to be compressed in the geospatial data;

Determine whether the latitude coordinates of the coordinate points of the data to be compressed are greater than the median value of the latitude coverage;

If the latitude coordinate is greater than the corresponding median value of the latitude coverage, the grid is divided, so that the latitude coverage of the grid is reduced from the median value to the maximum value;

If the latitude coordinate is less than or equal to the middle value of the latitude coverage, dividing the grid, so that the latitude coverage of the grid is reduced from the minimum value to the middle value;

The step of recursively dividing the latitude of the grid, and similarly dividing the longitude of the grid, so that the deviation between the center point of the grid and the coordinate point of the data to be compressed conforms to the preset value

Further, the maximum division error between the center point of the divided grid and the coordinate point of the data to be compressed is calculated by the following formula:

Where Width is the width of the grid, Height is the height of the grid, and Scale is the scale.

Further, the grid width Width and the grid height Height are respectively calculated by the following formulas:

Width=width/ ²ⁿ ;

Height=height/2 ⁿ .

Further, when the data to be compressed is a point vector element, the calculation steps of the binary offset of the point vector element are as follows:

Rearrange point vector elements in order according to binary code;

Record the first point as the original binary code, and store the subsequent points as the number of grids that deviate from the previous point, and convert them into binary storage;

The length of the offset binary code will be greater than 1 and less than the length of the original binary code, and the length of the offset binary code will be the maximum length of all offset binary codes;

For records with insufficient offset bits, the front end of the binary sequence is padded with 0.

Further, when the data to be compressed is a line vector element or an area vector element, the calculation steps of the binary offset of the line vector element or the area vector element are as follows:

Record the number of nodes of the feature object in each vector feature and the coordinates of the starting point, and perform the offset calculation in the unit of the feature object;

Set up the direction bit, calculate the offset direction between each coordinate point and the previous coordinate point, and store the eight possible offset directions that each point may generate with a two-bit binary code;

Calculate the offset, record the offset of the row and column along the horizontal and vertical directions, convert it to binary code, and then alternately store it as the final offset;

If the length of the offset binary code is greater than 1 and less than the length of the original binary code, the length of the offset binary code will be the maximum length of all offset binary codes;

For records with insufficient offset bits, padding is performed at the front with 0.

The space vector data compression method provided by the present disclosure takes the visualization of basic geospatial data as the application scenario, and sets the minimum resolution distance that meets the visual lossless screen display as the precision requirement, and realizes the vector space data compression processing. Mobile storage further improves data storage efficiency. The compression ratio is also higher compared to the prior art.

Description of drawings

Fig. 1 shows the flow chart of the vector space data compression algorithm according to the present invention;

Figure 2 is a schematic diagram of geographic space subdivision;

3 is a schematic diagram of a grid filtering process;

Fig. 4 is the maximum division error evaluation diagram under different division times;

Figure 5 is a schematic diagram of the storage format of the binary offset of point elements;

Fig. 6 is a schematic diagram of the binary offset storage format of line and area elements;

Figure 7a original geographic information image;

Fig. 7b, Fig. 7c and Fig. 7d are the result graphs of the compression ratio with the number of bits being 16, 18 and 20 respectively under the same number of divisions;

FIG. 8 is a schematic diagram of compression results under different bit numbers.

Detailed ways

It can be seen from the above description that the vector data model is one of the most basic geographic element entity expression models in the geographic information system. It realizes the expression of geographic element entities by recording the spatial coordinates of feature points and setting point set expression rules, and has the characteristics of complete target description and easy access to spatial relationships.

Because vector geospatial data uses high-precision coordinate values to express aggregate information, it has higher precision than raster data structure, and is suitable for expressing dynamic and multi-type complex geographic phenomena. The expression mostly adopts the vector data model. For the compression of vector data structures, many methods have also been proposed. For example, the patent document with the application number of 2010101806110 and the invention titled "Vector Data University Transmission Method Based on Pixel Lossless Compression of Ordered Point Sets" proposes a compression method based on ordered point sets on the premise that pixels are lossless; there are also applications No. 2014100243164, the patent document titled "A method for compressing vector data" discloses an offset-based vector data compression method; another application No. 2015107723716, the title of the invention is "Compression of Scalable Vector Graphics" , Drawing method and device" proposes an optimized storage method for SVG (Scalable Vector Graphics, scalable vector image) format. The solutions disclosed in the above-mentioned patent documents propose corresponding solutions for the compression of space vector data, but there are still problems such as limited compression ratio, complex storage format, and inability to achieve multi-resolution compression.

In response to the above analysis, the inventor has proposed a space vector data compression method through long-term creative work, which can effectively solve the technical problems raised in the above analysis. In the space vector data compression method provided by the present disclosure, firstly, according to the coordinate positions of the vector feature points, the geographic space is divided into two horizontal and vertical directions for multiple times, and the binary structure is sequentially constructed according to the positions of the coordinates of the feature points after each division. Integer encoding, the number of subdivisions is determined by the set compression precision. The more layers are subdivided, the higher the precision of expression and the longer the corresponding encoding, which can achieve the purpose of multi-resolution vector data compression. The low-precision encoding result of a given region is the prefix of the high-precision encoding result of the region, so the longer the encoded prefixes have the same length, the closer the distance between the two grid spaces is. The space vector data compression method provided by the present disclosure takes the visualization of basic geospatial data as the application scenario, and sets the minimum resolution distance that meets the visual lossless screen display as the precision requirement, and realizes the vector space data compression processing. Mobile storage further improves data storage efficiency. The compression ratio is also higher compared to the prior art.

The space-based data compression method proposed by the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. The advantages and features of the present invention will become apparent from the following description and claims. It should be noted that, the accompanying drawings are all in a very simplified form and in inaccurate scales, and are only used to facilitate and clearly assist the purpose of explaining the embodiments of the present invention.

Spatial terms such as "under", "below", "below", "above" and the like are used in the detailed description of the present disclosure for the purpose of easily describing one element and another shown in the figures positional relationship of components, but these are only examples and are not intended to limit the present invention. In addition to the orientation shown in the figures, spatially relative terms will encompass various orientations of the device in use or operation. The device may be otherwise positioned, such as rotated 90 degrees or at other orientations, and interpreted accordingly by the spatially relative descriptors used herein.

The steps of the multi-level compression method for vector space data provided by the present disclosure are as follows:

Step 1: According to the display accuracy of the device, the map scale and the display scale information, determine multiple Minimum Discernible Distances (MDD) expressed by the vector space data under the condition of multi-scale visual loss. In this process, it mainly includes the following: Implementation steps:

S11: For a display terminal with a display precision of DPI (Dot Per lnch, DPI), there are DPI pixels in a length of one inch. Under the International System of Units, 1 meter represents 39.3701 inches, which can be expressed in inches per meter (Inches PerMeter, IPM). Then the calculation formula of the SI unit length (Merer Per Dots, MPD) that a pixel point on the display terminal can represent is:

MPD=1/(DPI×IPM)

(1)

This distance is used to represent the minimum distinguishable distance that ensures visual loss during computer display.

S12: Each map has its own map scale (Scale) information, and the electronic map can be enlarged or reduced according to the display requirements during the display process in the computer. At this time, the display scale (Display Scale, Dscale) of the map will be changed to calculate the map. The calculation method of the magnification (Ratio) in the display process is:

Ratio=Dscale/Scale

(2)

S13 From S11, we have already learned the minimum distance that the computer pixel can distinguish, and for the specific picture displayed on the computer screen, the minimum distinguishable distance of the actual ground distance expressed by it conforms to the calculation formula:

MDD=MPD×Dscale

(3)

=MPD×Scale/Ratio

(4)

Step 2: Divide the space expressed by geographic coordinates step by step along the horizontal and vertical directions, and divide the geographic space into multi-level regular grids, as shown in Figure 2.

S21: Divide the two-dimensional space staggeredly along the horizontal and vertical directions respectively, and use the binary number "0" or "1" to identify the left and right or upper and lower subspaces after the division. After the two-dimensional space is divided in the horizontal direction, the left subspace is marked as "0", and the right subspace is marked as "1"; after the two-dimensional space is divided in the vertical direction, the lower subspace is marked as "0", The upper subspace is marked "1". Based on Morton coding, the binary identifiers "0" or "1" are alternately stored in the order of first, left, and then up and down, and the unique identifiers "00", "01", "10", "11" of each grid are generated, and Divide geographic space into a regular grid of quadrants.

S22: For the quadratic grid generated by S21, the division process of S21 is carried out with each grid as a unit. The grid generated by each division is one layer, and the new binary identification generated during the division process is stored alternately. It is then connected to the binary identification of the previous layer to form a binary string corresponding to the grid position one-to-one. The deeper the subdivision level is, the longer the binary string accumulates, the smaller the range represented by the grid, and the more accurate the representation of the position information.

S23: After the geographic space is recursively divided, each area has a unique code corresponding to it, and has obvious hierarchical characteristics in space. The code prefixes of different levels in the same area are the same, and the proximity of different areas is higher. , the prefix matching degree is also higher.

Step 3: Determine the position of the coordinate point of the data to be compressed in the multi-level regular grid, and replace the original double-precision type coordinate value with the grid binary code. The space vector data coordinate points in the grid are replaced by the regular grid center points generated by the subdivision.

S31: For the coordinate point of the data to be compressed (117.67198438°E, 42.1855639°N), the steps for determining the position in the regular grid are as follows:

S311: Obtain the latitude coverage of the geospatial data from a minimum value of 40°N to a maximum value of 44°N.

S312: Find the median value of the latitude coverage, and determine whether the latitude coordinates are greater than the median value.

S313: If it is greater than the value, output the character "1", and the coverage range is reduced from the middle value to the maximum value; otherwise, the character "0" is output, and the coverage range is reduced from the minimum value to the middle value.

It is easy to understand that the two steps of S312 and S313 are the steps of dividing the grid latitude.

S314: The two steps of S312 and S313 are recursive, so that the latitude division is gradually approached to the precise coordinate value, and the output characters are spliced into a string. When the number of divisions is n=9, 42.1855639°N is converted into a binary code such as 100010111 to represent.

S315: Of course, the principle of division of longitude is the same as that of latitude. Therefore, when the number of divisions is n=9, 117.67198438°E is converted into binary code 100111101

S316: alternately store the binary codes of the latitude and longitude coordinates to obtain the binary code 110000011101111011 of the coordinate point (117.67198438°E, 42.1855639°N), and also determine the position of the coordinate point in the regular grid.

S32: After the geographic space is divided into regular grids, each coordinate point located in the study area will fall into the unique grid of the classification, and the grid filtering method is used to simplify the vector space data: for the same element object falling on the grid All coordinate points in the same grid are replaced by the center point of the grid; for coordinate points of different feature objects that fall in the same grid, the grid center points need to be replaced and stored to ensure the integrity of the feature objects . As shown in Figure 3, such a grid filtering method realizes the compression of space vector data by reducing the number of coordinate points.

S33: Vector coordinate points are usually stored in double-precision data type, and one coordinate value storage space occupies 16 bytes, 8 bits per byte, occupying 96 bits of memory space. After replacing the vector coordinates with the grid center point in S32, the grid center point position information is stored in the grid binary code, and the storage number can be determined according to the requirements, which also greatly saves the storage space.

Step 4: Calculate the error generated by the regularized subdivision, and determine the subdivision level and data storage accuracy on the premise that the maximum error generated by the subdivision is less than the minimum distinguishable distance, that is, the visual loss is satisfied.

S41: For a rectangular grid, the maximum distance from all points in the grid to the center point is half the diagonal. Therefore, the calculation formula of the maximum division error (Maximal Error, ME) that can be generated by replacing other points in the grid with the grid center point is:

Where Width is the width of the grid, Height is the height of the grid, and Scale is the scale.

S42: grid width Width, which is related to grid height Height and subdivision times n:

Width=width/2 ⁿ

(6)

Height=height/2 ⁿ (7)

Where width is the width of the frame, and height is the height of the frame.

Bring (6) (7) into (5) to get the final formula for maximizing the score error:

It can be known from formula (8) that the larger the value of n is, the smaller the maximum division error is.

S43: Calculate the maximum division error under different n values, ME<MDD is a necessary prerequisite for the visual lossless display of the space vector data. To display spatial vector data in a geographic information system, it is necessary to meet the needs of enlarged display. The formula for magnifying magnification (Magnify) under the premise of visual loss is:

S44: On the premise of satisfying the data display requirements, determine the minimum number of subdivisions n, and then determine the data simplification degree of the space vector data, the number of bits in coordinate storage, the maximum division error under different n values, and the magnification of the map As shown in Figure 4.

Step 5: Using a single data file to be compressed as a unit, set a local coordinate reference system, perform binary offset calculation, and finally store the geographic coordinates of the vector space data with the binary offset. The data to be compressed can be further divided into point quantity elements, line vector elements and area vector elements.

S51: For point vector elements, the spatial distribution is random, but each point is an independent object, and the drawing sequence does not have any effect on the drawing result. The specific steps are as follows.

S511: Rearrange the point vector elements in order according to binary coding.

S512: Record the first point as the original binary code, and store the subsequent points as the number of grids that deviate from the previous point, and convert them into binary storage.

S513: The length of the offset binary code will be greater than 1 and less than the length of the original binary code, and the length of the offset binary code will be unified as the maximum length of all offset binary code lengths.

S514: For records with insufficient offset digits, fill in the front end with "0" to ensure the uniformity of offset digits

The specific storage format is shown in Figure 5.

S52: The coordinates of line elements or area elements (boundaries) are related in units of element objects, which cannot be reordered and drawn like point elements, but need to be organized in units of each object. The specific steps are as follows:

S521 : Record the number of nodes of element objects in each vector element and the coordinates of the starting point, and perform offset calculation in units of vector objects.

S522: Set up a direction bit, calculate the offset direction between each coordinate point and the previous coordinate point, and store the four possible offset directions generated by each point with a two-bit binary code.

S523: Calculate the offset, record the row and column offsets in the horizontal and vertical directions respectively, convert them into binary codes, and then alternately store them as the final offset.

S524: The length of the offset binary code is greater than 1 and less than the length of the original binary code, and the length of the offset binary code is unified as the maximum length of all offset binary code lengths.

S525: For records with insufficient number of offset bits, fill in the front end with "0" to ensure the unification of the number of offset bits.

The specific storage format is shown in Figure 6.

In another aspect of the present disclosure, a device is also provided, that is, the steps of the multi-level compression method for vector spatial data disclosed in the present disclosure can be executed on the device.

The multi-level compression method for vector space data provided by the present disclosure can be implemented on a Dell XPS8930 desktop computer configured with Intel i7-9700K@3.6GHz, using the Java programming language as an implementation tool. Taking the road vector data in the 1:1 million thematic map of administrative divisions in Beijing and its surrounding areas as an example, according to the method implementation process shown in Figure 1, the vector data compression calculation is carried out; then, the compression efficiency is evaluated according to different binary storage lengths .

Fig. 7 Different Bit (bit) lengths generated under different subdivision times, and the realized vector space data compression ratio. Figure 8 is a display of the compression effect under different Bit lengths. It can be seen from the figure that the redrawn compression results have obvious aliasing when the number of Bits reaches 16 and 18, which fails to meet the requirement of visual lossless. When the number of Bits reaches 16 and 18 At 20 bits, it can already meet the display requirements under the display scale of 1:1 million. At this time, the compression ratio of the data can reach 97.18%. It can be seen from the above verification that the vector space data compression algorithm proposed in this patent has very considerable compression efficiency.

To sum up, the space vector data compression method provided by the present disclosure takes the visualization of basic geospatial data as the application scenario, and sets the minimum resolution distance that meets the visual lossless of the screen display as the precision requirement, and realizes the vector space data compression processing. Net filtering and binary offset storage further improve data storage efficiency. The compression ratio is also higher compared to the prior art.

The above description is only a description of the preferred embodiments of the present invention, and is not intended to limit the scope of the present invention. Any changes and modifications made by those of ordinary skill in the field of the present invention based on the above disclosure all belong to the protection scope of the claims.

Claims (8)

1. a multistage compression method for vector space data, the method comprising the steps: Divide the space expressed by geographic coordinates step by step in the horizontal and vertical directions to obtain a multi-level grid that divides the geographic space; Determine the position of the coordinate point of the data to be compressed in the grid, and continue to divide the grid to replace the coordinate value of the coordinate point of the data to be compressed with the coordinates of the center point of the divided grid; wherein The coordinate deviation between the center point of the divided grid and the coordinate point of the data to be compressed conforms to a preset value; Taking a single data file to be compressed as a unit, set a local coordinate reference system, and calculate the binary coordinate deviation of the coordinate point of the data to be compressed and the center point of the grid after the division to obtain a binary offset; The geographic coordinates of the vector spatial data are stored at the binary offset. 2 . The multi-level compression method for vector space data according to claim 1 , wherein when dividing the horizontal and vertical directions, the left and right or upper and lower subspaces after the division are identified by binary. 3 . 3 . The multi-stage compression method for vector space data according to claim 1 , wherein the plurality of preset values are the minimum distinguishable distances expressed under the condition of visual lossless. 4 . 4. The multi-stage compression method for vector space data according to claim 1, characterized in that, the position of the coordinate points of the data to be compressed located in the grid is determined, and the grid is continuously divided to utilize the The step of substituting the coordinates of the center point of the divided grid for the coordinates of the coordinate points of the data to be compressed includes: Obtain the coordinate data of the coordinate points of the data to be compressed; Obtain the minimum and maximum values of the latitude coverage of the grid containing the coordinate points of the data to be compressed in the geospatial data; Determine whether the latitude coordinates of the coordinate points of the data to be compressed are greater than the median value of the latitude coverage; If the latitude coordinate is greater than the corresponding median value of the latitude coverage, the grid is divided, so that the latitude coverage of the grid is reduced from the median value to the maximum value; If the latitude coordinate is less than or equal to the middle value of the latitude coverage, dividing the grid, so that the latitude coverage of the grid is reduced from the minimum value to the middle value; The step of recursively dividing the latitude of the grid, and similarly dividing the longitude of the grid, so that the deviation between the center point of the grid and the coordinate point of the data to be compressed conforms to the preset value. 5. The multistage compression method of vector space data as claimed in claim 1, wherein the maximum division error between the center point of the divided grid and the coordinate point of the data to be compressed is calculated by the following formula :

Where Width is the width of the grid, Height is the height of the grid, and Scale is the scale. 6. The multi-level compression method for vector space data as claimed in claim 5, wherein the grid width Width and the grid height Height are calculated by the following formula respectively: Width=width/ ²ⁿ ; Height=height/2 ⁿ . 7. The multistage compression method for vector space data as claimed in claim 1, wherein when the data to be compressed is a point vector element, the binary offset calculation steps of the point vector element are as follows: Rearrange the point vector elements in order according to the binary code; Record the first point as the original binary code, and store the subsequent points as the number of grids that deviate from the previous point, and convert them into binary storage; The length of the offset binary code will be greater than 1 and less than the length of the original binary code, and the length of the offset binary code will be uniformly taken as the maximum length of all offset binary codes; For records with insufficient offset bits, the front end of the binary code sequence is filled with 0. 8. The multi-stage compression method for vector space data according to claim 1, wherein when the data to be compressed is a line vector element or an area vector element, the binary offset of the line vector element or the area vector element The calculation steps are as follows: Record the number of nodes of the feature object in each vector feature and the coordinates of the starting point, and perform the offset calculation in the unit of feature object; Set up the direction bit, calculate the offset direction of each coordinate point and the previous coordinate point, and store the eight possible offset directions of each point with two-bit binary code; Calculate the offset, record the offset of the row and column along the horizontal and vertical directions respectively, convert it to binary code, and then alternately store it as the final offset; If the length of the offset binary code is greater than 1 and less than the length of the original binary code, the length of the offset binary code will be the maximum length of all offset binary codes; For records with insufficient offset bits, padding is performed at the front with 0.

Images