CN101483779A - Compressing method for two-dimension vector map - Google Patents

Abstract

This invention relates to a compression method for 2-dimensional vector map, aiming to solve problems that massive computation amount and data conversion loss in existing compression method. Method according to this invention comprises vector data file compression and property data file compression. Vector data file compression comprises processing contents of vector data file contained in vector map according to layer classification, respectively encoding coordination data and property codes in layer data of same classification, compressing data amount of property codes with redundancy reduction method, compressing coordination data with irregularity coefficient forecast encoding method. Property data file compression comprises compressing property data file contained in vector map and matching with vector data file with BWT algorithm. This invention can enhance compression efficiency and reduce encode/decode complex.

Description

A kind of compression method of two-dimension vector map

Technical field

The invention belongs to WebGIS map vector transmission and the field of storage relevant, relate to a kind of lossless compression method of two-dimension vector map with mobile GIS.

Background technology

Universal day by day along with the GIS application technology, numerical map has replaced traditional paper map to become the main carrier of storage geographical spatial data gradually.Just because the high speed development of network technology and Wireless Networking technology, WebGIS technology and mobile GIS technology are arisen at the historic moment, and these The Application of Technology all to the efficient storage of a large amount of numerical maps and fast transmission requirement has been proposed, numerical map efficiently be compressed into the key that addresses this problem.

Different with the compression method of grating map; because current application is generalized information system (as: MapInfo widely; ArcGIS) graph data of map vector can be separated with attribute data usually and deposit; with respect to attribute data; the graph data (promptly putting ordered sets) that comprises figure and topological relation information accounts for exhausted vast scale data volume, so the compression method of graph data becomes the key of map vector data compression efficiency.At present developed multiple algorithm in the lossy compression method field of map vector, as Douglas-Peucker algorithm etc., but owing to adopt the method for lossy compression method will cause unpredictable mistake than higher occasion in some required precision, the research of vector map lossless compression has had certain realistic meaning, but at present also fewer to the achievement in research in this field.Existing achievement in research has proposed the compression algorithm at the longitude and latitude coordinate system map vector in the map vector, but owing to different periods, different zones, different purposes make the coordinate system of various maps have very big difference, there is defective in the application of this algorithm.

As previously mentioned, existing achievement is not supported various map vector, and only supports the compression of longitude and latitude coordinate system map vector.Adopt this algorithm, must be converted into the longitude and latitude coordinate system such as national coordinate system, local coordinate system, and then compress earlier with other coordinate systems before the compression.Such method will inevitably increase operand, and since in the transfer process because the restriction of precision can not guarantee that data can't harm in transfer process.

Summary of the invention

The object of the present invention is to provide a kind of lossless compression method of two-dimension vector map.

The compression method of map vector of the present invention comprises vector data file compression and attribute data File Compress.

The compression of described vector data file is that the content in the vector data file that map vector is comprised is separately handled by layer type, specifically: at first the figure layer data of file mid point class, line class, Noodles is separated according to type, coordinate data in the figure layer data of the same type that will separate is separated with attribute codes then, coordinate data is separated with attribute codes encode at last:

Adopt the method amount of compressed data of eliminating redundancy for attribute codes, exactly identical data is merged, different pieces of information is carried out superposition;

Adopt coefficient of irregularity predictive coding method to compress for coordinate data.

Described attribute data File Compress is that the attribute data file with the vector data file coupling that map vector is comprised compresses by BWT (Burrows-Wheeler-Transformation) algorithm.The BWT algorithm is ripe data compression method.

The concrete steps of described coefficient of irregularity predictive coding method are:

Step 1, the number of significant digit of the coordinate figure of all coordinate datas is unified, X coordinate and Y coordinate all with the maximum coordinate figure of number of significant digit as standard, mend 0 behind the coordinate decimal point for the number of significant digit deficiency, convert after reunification coordinate figure to the long integer;

Step 2, keep the coordinate of first coordinate data, since second coordinate data, deduct the X coordinate figure of I-1 coordinate data with the X coordinate figure of I coordinate data, the offset value that obtains replaces the X coordinate figure of I coordinate data, and to offset value with long integer representation; Simultaneously, since second coordinate data, deduct the Y coordinate figure of I-1 coordinate data with the Y coordinate figure of I coordinate data, the offset value that obtains replaces the Y coordinate figure of I coordinate data, and to offset value with long integer representation;

Step 3, with the value of the side-play amount of long integer as conversion coefficient, storing the required byte number of this conversion coefficient is n _x

$n_x=\left\{\begin{array}{c}1,x_i<2^7-1\\ 2,x_i<2^{14}-1\\ 3,x_i<2^{21}-1\\ 4,\mathrm{otherwise}\end{array}\right.$

x _iThe numerical value of expression conversion coefficient, otherwise are represented under other all situations except that above three kinds of situations;

Step 4, before the conversion coefficient of each coordinate data, insert check code, the length of each check code is 1 byte, comprise eight, wherein first is that X coordinate flag bit, second and the 3rd are that X coordinate data section, the 4th are that the sign flag bit, the 5th of X coordinate is that Y coordinate flag bit, the 6th and the 7th are that Y coordinate data section, the 8th are the sign flag bits of Y coordinate;

Data section DataValue carries out assignment according to following formula

$\mathrm{DataValue}=\left\{\begin{array}{c}\&prime;\&prime;00\&prime;\&prime;,n_x=1\\ \&prime;\&prime;01\&prime;\&prime;,n_x=2\\ \&prime;\&prime;10\&prime;\&prime;,n_x=3\\ \&prime;\&prime;11\&prime;\&prime;,n_x=4\end{array}\right.$

Step 5, the conversion coefficient and the corresponding check sign indicating number of each coordinate data are stored.

During decoding, at first the attribute data file is carried out the inverse transformation of BWT algorithm, reduction attribute data file.Vector data file then earlier by the inverse process reduction coordinate data of BWT algorithm and the conversion coefficient of attribute codes, obtains each property value of conversion coefficient according to check code then, by the inverse process reduction coordinate data of " based on the predictive coding of coefficient of irregularity "; According to the map vector form and annex the inverse process reduction file header of algorithm and attribute codes such as paintbrush paintbrush, finally obtain complete vector data file.

The present invention is on the basis of using for reference traditional text or Image Compression characteristics, account for the feature of most sign indicating number amounts at vector data in the map vector, lay particular emphasis at coding side and increase information clustering processing module, and employing " based on the predictive coding of coefficient of irregularity " transform method, can remove information unnecessary in the vector data and simplify data to be encoded, thereby can reduce the coding and decoding complexity,, improve the map vector compression ratio at last by ripe BWT compression algorithm conversion coefficient.The inventive method can improve compression efficiency, and reduces the coding and decoding complexity.

Embodiment

The lossless compression method of n dimensional vector n map adopts following steps:

The first step: the vector data file content in the two-dimension vector map is pressed layer type classification, and the coordinate data in the phase diagram layer data that will separate is then separated with attribute codes and is encoded;

Second step: adopt the method amount of compressed data of eliminating redundancy for attribute codes, exactly identical data is merged, different pieces of information is carried out superposition;

The 3rd step: adopt coefficient of irregularity predictive coding method to compress for coordinate data, the concrete steps of described coefficient of irregularity predictive coding method are:

Step 1, the number of significant digit of the coordinate figure of all coordinate datas is unified, X coordinate and Y coordinate all with the maximum coordinate figure of number of significant digit as standard, mend 0 behind the coordinate decimal point for the number of significant digit deficiency, convert after reunification coordinate figure to the long integer;

Step 2, keep the coordinate of first coordinate data, since second coordinate data, deduct the X coordinate figure of I-1 coordinate data with the X coordinate figure of I coordinate data, the offset value that obtains replaces the X coordinate figure of I coordinate data, and to offset value with long integer representation; Simultaneously, since second coordinate data, deduct the Y coordinate figure of I-1 coordinate data with the Y coordinate figure of I coordinate data, the offset value that obtains replaces the Y coordinate figure of I coordinate data, and to offset value with long integer representation;

Step 3, with the value of the side-play amount of long integer as conversion coefficient, storing the required byte number of this conversion coefficient is n _x

$n_x=\left\{\begin{array}{c}1,x_i<2^7-1\\ 2,x_i<2^{14}-1\\ 3,x_i<2^{21}-1\\ 4,\mathrm{otherwise}\end{array}\right.$

x _iThe numerical value of expression conversion coefficient;

Step 4, before the conversion coefficient of each coordinate data, insert check code, the length of each check code is 1 byte, comprise eight, wherein first is that X coordinate flag bit, second and the 3rd are that X coordinate data section, the 4th are that the sign flag bit, the 5th of X coordinate is that Y coordinate flag bit, the 6th and the 7th are that Y coordinate data section, the 8th are the sign flag bits of Y coordinate;

Data section DataValue carries out assignment according to following formula

$\mathrm{DataValue}=\left\{\begin{array}{c}\&prime;\&prime;00\&prime;\&prime;,n_x=1\\ \&prime;\&prime;01\&prime;\&prime;,n_x=2\\ \&prime;\&prime;10\&prime;\&prime;,n_x=3\\ \&prime;\&prime;11\&prime;\&prime;,n_x=4\end{array}\right.$

Step 5, the conversion coefficient and the corresponding check sign indicating number of each coordinate data are stored.

The 4th step: attribute data file and vector data file in the two-dimension vector map mate one by one, the ratio data of this part is very little comparatively speaking, and the compression of attribute data file is that two-dimension vector map is comprised to compress with the data compression method BWT algorithm of attribute data file vector data file coupling by maturation.

Claims (1)

1, a kind of compression method of two-dimension vector map comprises vector data file compression and attribute data File Compress, it is characterized in that:

The compression of described vector data file is that the content in the vector data file that map vector is comprised is separately handled by layer type, specifically: at first the figure layer data of file mid point class, line class, Noodles is separated according to type, coordinate data in the figure layer data of the same type that will separate is separated with attribute codes then, coordinate data is separated with attribute codes encode at last:

Adopt the method amount of compressed data of eliminating redundancy for attribute codes, exactly identical data is merged, different pieces of information is carried out superposition;

Adopt coefficient of irregularity predictive coding method to compress for coordinate data;

Described attribute data File Compress is that the attribute data file with the vector data file coupling that map vector is comprised compresses by the BWT algorithm;

Wherein, the concrete steps that adopt coefficient of irregularity predictive coding method to compress for coordinate data are:

Step 1, the number of significant digit of the coordinate figure of all coordinate datas is unified, X coordinate and Y coordinate all with the maximum coordinate figure of number of significant digit as standard, mend 0 behind the coordinate decimal point for the number of significant digit deficiency, convert after reunification coordinate figure to the long integer;

Step 2, keep the coordinate of first coordinate data, since second coordinate data, deduct the X coordinate figure of I-1 coordinate data with the X coordinate figure of I coordinate data, the offset value that obtains replaces the X coordinate figure of I coordinate data, and to offset value with long integer representation; Simultaneously, since second coordinate data, deduct the Y coordinate figure of I-1 coordinate data with the Y coordinate figure of I coordinate data, the offset value that obtains replaces the Y coordinate figure of I coordinate data, and to offset value with long integer representation;

Step 3, with the value of the side-play amount of long integer as conversion coefficient, storing the required byte number of this conversion coefficient is n _x

$n_x=\left\{\begin{array}{c}1,x_i<2^7-1\\ 2,x_i<2^{14}-1\\ 3,x_i<2^{21}-1\\ 4,\mathrm{otherwise}\end{array}\right.$

x _iThe numerical value of expression conversion coefficient;

Step 4, before the conversion coefficient of each coordinate data, insert check code, the length of each check code is 1 byte, comprise eight, wherein first is that X coordinate flag bit, second and the 3rd are that X coordinate data section, the 4th are that the sign flag bit, the 5th of X coordinate is that Y coordinate flag bit, the 6th and the 7th are that Y coordinate data section, the 8th are the sign flag bits of Y coordinate;

Data section DataValue carries out assignment according to following formula

$\mathrm{DataValue}=\left\{\begin{array}{c}\&prime;\&prime;0\&prime;\&prime;,n_x=1\\ \&prime;\&prime;01\&prime;\&prime;,n_x=2\\ \&prime;\&prime;10\&prime;\&prime;,n_x=3\\ \&prime;\&prime;11\&prime;\&prime;,n_x=4\end{array}\right.$

Step 5, the conversion coefficient and the corresponding check sign indicating number of each coordinate data are stored.