CN105205126B - A kind of storage method and its storage system of map tile - Google Patents

Abstract

The invention discloses a method for storing map tiles and a storage system thereof. The method includes: sequentially dividing an electronic map into map tiles according to a predetermined size, and classifying according to a first predetermined rule; According to the number of map tiles in the map, construct c files according to the second predetermined rule; if c is greater than the first predetermined number, then construct d folders according to the third predetermined rule; if d is greater than the second predetermined number, then follow the fourth predetermined The rule constructs e lower-level folders, and so on, until the number of upper-level folders stored in the lower-level folders is less than or equal to the corresponding predetermined number. The invention centrally stores the map tiles by classifying the map tiles and constructing files and folders, which is beneficial to data storage speed and later reading speed.

Description

A storage method and storage system for map tiles

technical field

The invention relates to the technical field of data storage, in particular to a method for storing map tiles and a storage system thereof.

Background technique

Because WebGIS reads geographic data from the server, and the amount of data read each time is limited, and WebGIS needs to support multiple clients to read geographic data, so for the acquisition of server-side geographic data, the method of using tile data is adopted. method. The most common tile data are raster data such as jpg data and png data of 256*256 size, and vector tile data of corresponding size. Taking raster tile data as an example, common raster tile data such as ArcGIS uses the tile level as a large folder, and the row and column numbers are calculated to classify folders, and the final storage is 256*256 jpg or PNG data. When publishing services, the client directly obtains the server-side raster tile data files through certain rules through web server-side software such as IIS or Apache. This approach has several disadvantages:

1. Because the tile data is finally a jpg or png file with a size of 256*256, their file size is too small; and because the amount of image data to be displayed is too large, such as the global 17-level image tile of Google's standard WebMercator coordinates The amount of data can reach 4 to the 17th power of about 17179869184 image tiles. Based on the average size of each image tile of 15KB, there are 245,760GB, or 240TB of data. Google's global image data is about 20 levels, and even reaches 24 levels in some areas. If all are stored in jpg format, it will bring a lot of small file data. Due to the characteristics of computer hard disk storage, the speed of sequential access is much faster than that of random access. Generally speaking, for existing ordinary desktop mechanical hard drives, sequential access can reach a speed of about 200MB/S, while random 4K access can only reach a speed of about 3MB/S at most. Because the storage of scattered jpg files on the hard disk is often not continuous, the speed of reading and writing it is similar to the random 4K access speed of a mechanical hard disk. This feature leads to a particularly slow file transfer speed when copying and distributing cut raster tile data. It is understood that the distribution of raster tile data in this way is often not as fast as re-cutting the data directly on the target storage. This is the first drawback.

2. Still because the tile file is too small, when the client requests a data file, the server-side software will directly read the data of the small files one by one. This kind of reading is still the random 4K access speed of the mechanical hard disk, so using such small files as image tile data will affect the speed at which the server provides image tile services. In most cases, the image tile server deployed in this way cannot provide a service speed that reaches the server egress bandwidth. In this case, it is equivalent to the waste of CPU and memory resources of the server, which does not play any role in improving the service speed.

3. This small file tile data storage method cannot effectively provide a notification to the browser client when the tile data is updated. At present, the industry generally adopts the method of updating the tile data of such small files to release the updated tile data as a new address, so as to force refresh the user's cookie to refresh the tile data. This approach is not suitable for scenarios where data is frequently updated, because for users, after browsing to a region, they will continuously update to new tiles when the server is updated, and this tile is likely to be unavailable. Nothing has changed. This wastes valuable bandwidth resources and increases the burden on the server.

Contents of the invention

The technical problem to be solved by the present invention is to provide a method for storing map tiles and a storage system thereof, which can solve the shortcomings and deficiencies in the prior art.

The technical scheme that the present invention solves the problems of the technologies described above is as follows:

Based on one aspect of the present invention, a method for storing map tiles is provided, the method comprising:

S1. Sequentially divide the electronic map into map tiles according to a predetermined size, and classify the map tiles according to a first predetermined rule to form a b-level data file, wherein b is a positive integer, and the value of b is 1, 2, 3…;

S2. According to the number of map tiles in the i-th level data file, c files are constructed according to the second predetermined rule, and each of the constructed files stores a corresponding number of map tiles, wherein i is an integer, and i is taken as The value is 0, 1...(b-1), c is a positive integer, and the value of c is 1, 2, 3...;

S3. If c is greater than the first predetermined number, construct d folders according to the third predetermined rule, and store corresponding files in each of the constructed folders, where d is an integer, and the value of d is 1 , 2, 3...;

S4. If d is greater than the second predetermined number, then construct e lower-level folders according to the fourth predetermined rule, and each of the constructed lower-level folders stores a corresponding number of folders, and so on until the next level The number of upper-level folders stored in the folder is less than or equal to the corresponding predetermined number.

Based on another aspect of the present invention, the present invention provides a storage system for map tiles, the system comprising:

The tile grading module is used to sequentially divide the electronic map into map tiles according to a predetermined size, and classify the map tiles according to a first predetermined rule to form a b-level data file, wherein b is an integer, and the value of b is 1, 2, 3...;

The file construction module is used to construct c files according to the second predetermined rule according to the number of map tiles in the i-th level data file, and store corresponding several map tiles in each of the constructed files, where i is an integer , the value of i is 0, 1...(b-1), c is a positive integer, and the value of c is 1, 2, 3...;

A folder construction module, used to construct d folders according to the third predetermined rule if c is greater than the first predetermined number, and store corresponding several files in each of the constructed folders, where d is a positive integer, and d The values are 1, 2, 3...;

The next-level folder construction module is used to construct e lower-level folders according to the fourth predetermined rule if d is greater than the second predetermined number, and each of the constructed lower-level folders stores a corresponding number of folders , and so on, until the number of upper-level folders stored in the lower-level folder is less than or equal to the corresponding predetermined number.

In the method and system for storing map tiles provided by the present invention, the map tiles are first graded, and then for each level of data files, files, folders, and next files are constructed according to the number of map tiles in the data files. Level folders are used to classify map tiles layer by layer, which makes the management of map tiles easier, and is conducive to data storage speed and later reading speed.

Description of drawings

FIG. 1 is a flowchart of a method for storing map tiles in Embodiment 1 of the present invention;

FIG. 2 is a schematic diagram of a storage system for map tiles according to Embodiment 2 of the present invention.

Detailed ways

The principles and features of the present invention are described below in conjunction with the accompanying drawings, and the examples given are only used to explain the present invention, and are not intended to limit the scope of the present invention.

Embodiment 1. A method for storing map tiles. The method provided in this embodiment will be described in detail below with reference to FIG. 1 .

Referring to Fig. 1, S1, sequentially divide the electronic map into map tiles according to the predetermined size, and classify the map tiles according to the first predetermined rule to form a b-level data file, wherein b is an integer, and the value of b is 1, 2, 3….

Specifically, first, the electronic map is sequentially divided into a map tiles according to a predetermined size, and the a map tiles are classified to form a b-level data file, wherein the map tile in the i+1th level data file The number is a predetermined multiple corresponding to the number of map tiles in the i-level data file, and the predetermined multiple can be an integer multiple or a decimal multiple, and the ratio of the number of map tiles in each adjacent two-level data file (that is, the corresponding The predetermined multiple) can be a fixed multiple or a dynamic multiple, where i is an integer, and the value of i is 0, 1...(b-1), and the map tiles in the i-th level data file form an m*n tile matrix. Among them, m and n are both positive integers, m and n can be equal, then the formed tile matrix is a square tile matrix (for example, the tile map of the Google coordinate system), m and n can also be unequal, then the formed The tile matrix is a rectangular tile matrix (for example, a tile map in a non-Google coordinate system).

Take the tile map of the commonly used Google coordinate system as an example. The tile map of the Google coordinate system is the WebMercator coordinate system. The total range indicated by the tile map is longitude -180 degrees to 180 degrees, latitude -85 degrees to 85 degrees, according to The coordinate rules represented by the electronic map segment the electronic map, and classify the segmented map tiles. In this embodiment, the 0th level is represented by a map tile. Starting from the 1st level, the horizontal and vertical directions of each level of map tiles are increased to twice the original, that is, each level of map tiles The number of tiles is four times the number of tiles in the upper level map. Since the number of map tiles in each level of data files is four times the number of map tiles in the previous level of data files, after reaching a certain level, the total number of map tiles in this level of data files can always be divisible by 65536 . In this way, in this embodiment, a file system can be designed so that the file system can store a certain amount of map tiles and meet the performance and storage capacity requirements.

S2. According to the number of map tiles in the i-th level data file, c files are constructed according to the second predetermined rule, and each of the constructed files stores a corresponding number of map tiles, wherein i is an integer, and i is taken as The value is 0, 1...(b-1), c is a positive integer, and the value of c is 1, 2, 3....

Specifically, according to the value of the number of rows m and the number of columns n of the tile matrix in the i-th level data file, c files are constructed in the i-th level data file, where c=ceiling(m/k1)*ceiling (n/k2) Among them, 16≤k1≤m, 16≤k2≤n; Among them, c files form a g*h file matrix in the i-level data file, c=g*h, g and h are both integer.

According to the example in the aforementioned step S1: for levels with a total number of map tiles less than 65536, all map tiles in each level are stored in one file, that is, from level 0 to level 7, all are stored in one file All map tiles in this level; starting from level 8, the total number will be divided by 65536 files for storage, and each file consists of 256×256 small tile images. That is, level 8 is one file, level 9 is four files, and so on. There are 16777216 files at level 20.

S3. If c is greater than the first predetermined number, construct d folders according to the third predetermined rule, and store corresponding files in each of the constructed folders, where d is an integer, and the value of d is 1 , 2, 3... .

Specifically, according to the value of the number of rows g and the number of columns h of the file in the i-level data file, d folders are constructed in the i-level data file, and each of the constructed folders stores a number of corresponding files , wherein d=ceiling(g/k3)*ceiling(h/k4), wherein, k3, k4 are positive integers; wherein, d folders form a l*w folder matrix in the i-level data file, d=l*w, wherein l and w are both integers.

S4. If d is greater than the third predetermined number, construct e lower-level folders according to the fourth predetermined rule, and each of the constructed lower-level folders stores corresponding folders, and so on until the next level The number of upper-level folders stored in the folder is less than or equal to the corresponding predetermined number.

Specifically, according to the number of rows l and the number of columns w of folders in the i-level data file, e lower-level folders are constructed in the i-level data file, and each of the constructed lower-level folders stores Corresponding folders, where e=ceiling(l/k5)*ceiling(w/k6), where k5 and k6 are both positive integers.

For step S3 and step S4, according to the aforementioned example: due to the file system mechanism of the operating system, for example, there are at most 65534 files or folders under the next folder of the FAT32 file format system under the Windows operating system (NTFS is limited to about 4294967294, But it is better not to put too many files, which will affect the speed of the operating system), so these files should not be placed in the same folder. Therefore, in this embodiment, a folder is generated for every six levels down to include files that could not be loaded before. For example, the 15th level data files were previously stored in the root folder with the level as the root folder name, and 4096 files were stored in the 14th level root folder. Starting from level 15, there are four subfolders, and each subfolder stores 4096 files. There are 4096 subfolders in the 20th level, and 4096 files are stored in each subfolder. And when it reaches the 24th level, there are 256 subfolders, 4096 grandchildren folders in each subfolder, and 4096 files in each grandchildren folder. So on and so forth. At present, the world's highest resolution map tile is only up to the 24th level, and has not been extended further. Therefore, for any tile, the file at the corresponding location can be found in the file system, so that the corresponding tile data can be read from the file.

Using the map tile storage method provided in this embodiment, even if the projection coordinate system is different, or the map resolution of the hierarchical data does not increase evenly according to the double relationship, resulting in a certain level of map tile total Quantity The number is not a number divisible by 65536, it can be applied. For tiles that exceed 256 horizontally or vertically, a large file composed of 256×256 small tile pictures is still used to store them. The number of rows and columns of horizontal or vertical files is rounded up to the value of the maximum number of horizontal or vertical tiles/256, which may be set to N. Then within the range of row or column numbers from 0 to N-1, the storage value of the row or column of the large file is 256, and the row or column at the edge is the total value minus 256×(N-1 ). The rules for setting up folders are similar to the above-mentioned folder value rules. Take the larger value of the total number of rows or columns in the file and divide it by 64. If it is greater than 1, continue to divide by 64 until it is less than 1. In addition to M times, M-1 layer subfolders are required for management. For example, if the total number of rows and columns of the file is 65 rows x 63 columns, and the serial number of the file is 8, 8, then a layer of subfolders is required. The final file path should be L/0/520.dat (8/64=0, 8×64+8=520).

To sum up, the internal structure of a specific map tile storage system provided by this embodiment is as follows: for the total number of rows and columns of a map tile (whichever is larger) is less than 2 to the power of N (N<=8) but For levels greater than 2 to the N-1 power, only one file is generated, and the number of tile rows and columns in this file are both 2 to the N power. For the level where the total number of rows and columns (whichever is larger) of tiles is greater than 256, the total number of generated files is ([maximum row number/256]+1)×([maximum column number/256]+1) , the number of rows and columns of each generated map tile is 256. The rules for generating subfolders refer to the above rules for generating folders. Take the seventh level of map tiles in the Google coordinate system as an example. There are 128×128 slices in the world at level 7. Then a file is generated. The number of rows and columns of map tiles in this file are both 128. For each file, the corresponding tiles to be stored form an N×N tile matrix, where N is the number of rows and columns of tiles. For map tiles in a non-Google coordinate system, the number of map tiles actually stored in each file may be N1×N2, where N1 and N2 are not greater than N. For any map tile (i, j) of a certain level, its logical position in the file storing it is i%N*N+j%N. Each map tile can find the location of the file where it is stored and its logical location in this file, and each logical location of each file in the file system can correspond to the location in the map tile system A map tile. In this way, the map tile and the logical position of the map tile in the file form a one-to-one correspondence.

After the map tiles are classified according to the above method, each file is composed of a file header, an index table and data content during storage, wherein the file header stores the basic information and metadata description information of the map tiles in the file, The index table stores the map tile index value in the file, and the data content includes the map tile and metadata information, and the metadata information includes at least the MD5 code of the map tile. Specifically: the file header refers to the TIFF format specification and is stored in the front part of the file. In order to simplify the description, this embodiment only describes the part that plays a key role in the file format. The file header mainly includes: tile width and height; tile row number and column number (not the actual row number and column number, but the row number and column number calculated by the above calculation method); tile position index position; tile file length Index location; whether the tile storage is complete; whether the tile storage location has been tuned; whether to use a substitute image to represent a tile without data; the location of the substitute picture; the length of the substitute picture; the storage method of metadata information; the number of metadata information types ; Metadata information type; Metadata information position index position; Metadata information length index position.

Take the map tile file system with a tile width and height of 256 as an example: the tile position index position and the tile file length index position indicate the tile position index table and the tile file length index table in this file Offset. Whether the tile storage location is optimized indicates whether the location where the tile data is actually stored in the file has been optimized. Whether to use an alternative picture to represent a tile without data, which means that if an alternative picture is used to indicate no data, the value is true; otherwise, the index position of a tile without data is directly represented by 0. The location of the alternate image is the offset of the alternate image in file storage. If a substitute picture is used to indicate no data, when the tile data is missing, the corresponding tile position points to the location where the substitute picture is stored in the file, and the corresponding tile file length is the length of the substitute picture. The metadata information storage method indicates how the metadata information is stored in the file. There are four ways to store: the first is that each set of metadata information is directly stored at the end of each tile data; the second is that each set of metadata information is stored at the end of all tile data, but each Group metadata information is stored together; the third is that the storage location of metadata information is completely free to adjust, that is, the location of each individual piece of metadata information can be adjusted arbitrarily. The fourth is that part of the metadata information is stored in the first way, part of the metadata information is stored in the second way, and part of the metadata information is stored in the third way. The number of metadata information types, which indicates how many types of metadata information each tile image in this file has; the metadata information type, immediately after the value of the number of metadata type information, takes up (number of metadata type information × 4 words Section) space size, each int value represents a type of metadata information. The metadata information position index position and the metadata information length index position indicate the offsets of the metadata information position index table and the metadata information length index table in the file.

After the file header, the index table is stored. In this embodiment, the arrangement of the index table is as follows: tile position index; tile file length index; metadata information location type; metadata information length index. The tile position index is a matrix of N×N×4 bytes (int value), and it has a one-to-one correspondence with the tiles to be stored in this file. Each int value indicates the storage location of the corresponding tile data in this file. The tile file length index is a matrix of N×N×4 bytes (int value). As mentioned above, it also has a one-to-one correspondence with tiles. Each int value indicates the length of the corresponding tile data in this file. Whether the tile storage is complete, it means that the tile storage is complete if all the tiles that need to be stored in this file are in place. If what needs to be stored is N1×N2, then all the tiles of N1×N2 are in place, which means the storage is complete.

The size of the metadata information location index is related to the storage method of the metadata information and the number of metadata information types. If the metadata information is stored in the first way, the index size of the metadata information location is 0, because the location of the metadata information can be calculated using the tile data location, and the metadata information can be read directly; if the metadata information is stored in the second method In the first way, the metadata information location index size is (N×N×4 bytes) space size, and each int value represents the offset of each group of metadata information records; and if the metadata information is stored as the third mode, the size of the metadata information location index is (M×N×N×4 bytes) space size, M is the number of metadata information types, and each M int value represents the respective offset of each group of metadata information records amount; the fourth storage method, the metadata information location index table, is a little more complicated. First use an M-length int array to indicate the storage type of metadata information. If the value is 0, it means the metadata information of the i-th category is directly stored after each piece of tile data; if the value is a positive integer, such as 1, it means the metadata information of the first aggregated category, and all values of 1 Metadata information is stored together according to the second method, which only needs to occupy one index bit; if the value is -1, it means that the i-th type of metadata information is stored according to the third method, and any position is possible. An index bit is also occupied, therefore, the final required index space size is (M`×N×N×4 bytes). M` = number of aggregated classes + number of independent classes. An int array with a length of M`, which sequentially represents the offset of the jth aggregation class and the kth independent class of each set of tile metadata information records. So far, the storage of the metadata information location index table is completed.

The size of the metadata information length index is related to the number of metadata information types and the metadata information types. Metadata information types can be divided into variable-length metadata information and fixed-length metadata information. For fixed-length metadata information, you only need to know the fixed-length size. For variable-length metadata information, it is necessary to know the information length of each variable-length metadata record. Therefore, the metadata information length index table is firstly the fourth metadata information storage method similar to the metadata information location index table. An M-length int array is used to indicate whether each type of metadata information is of fixed length or variable length. . If the value is a positive integer, it means that this type of metadata information is of fixed length. If the value is 0, it means that the metadata information of this type is of variable length. Immediately following it is an index space of (M`×N×N×4). M` is the number of variable-length metadata information categories, and each M`-length int array sequentially represents the length of the i-th variable-length information of each group of tile metadata information records.

After the index table, the data content is stored, including tile data and metadata information. In this embodiment, metadata information with a fixed size and no change is stored in front of the data area, and map tiles and metadata information with change are stored in the back of the data area. When storing map tiles in each file, if performance optimization is not considered, these data can be stored in the following manner: first, if there is an alternative tile image, store the alternative tile image first. Then store the tile data one by one according to the row and column numbers. After the tile data is stored, each group of tile metadata information is stored one by one, and the index table is updated at the same time. Or write the currently read tile data and metadata information in real time at the end of the file. The read tile may be a tile at any coordinate in this file, and update the index table at the same time. However, this direct storage method does not bring optimal read performance. In view of the fact that the tile data is displayed on the user terminal to fill the entire user window, it can be known that adjacent tile data are downloaded and displayed together in most terminal applications. From the characteristics of mechanical hard disk storage, we can know that the hard disk reads data in sequential storage much faster than random storage. The deep-seated reason is that the sequential storage and reading of the mechanical hard disk reduces the number of hard disk probe seeks, unlike random reads that may occur many seeks. Therefore, one of the key points of optimizing tile data reading in the present invention is how to obtain adjacent tile data at one time or by reducing the number of times of reading. Therefore, in this embodiment, the Morton code encoding rule is used to reorder the storage order of the map tiles in each file, so as to optimize the storage order of the map tiles. Taking a 4×4 file as an example, the actual storage order of map tile files numbered 0 to 15 in the file is 0, 1, 4, 5, 2, 3, 6, 7, 8, 9, 12, 13 , 10, 11, 14, 15. In this way, if any map tile is requested, its logically adjacent map tiles can be read with the least number of reads. For example, if you request a map tile with logical number 5, you can see from the above sequence that its physical storage order in the file is the third bit, and bits 0, 1, and 2 are map tiles with logical numbers 0, 1, and 4, respectively. The map tiles with logical numbers 0, 1, 4, and 5 are also adjacent in terms of spatial relationship. However, in the actual application of this principle in this embodiment, the 16×16 range tile corresponding to the tile requested by the user is directly read. From this principle, we can see that these 256 tiles are in the adjacent range, and since the general resolution of tiles in the Google coordinate system is 256×256, the coverage of tiles in this range can reach 4096×4096 pixels. As for the resolution of the user window, the most common one is 1920×1080. And even the MacbookProRetina display only has a resolution of 2880×1800. That is to say, if the 16×16 range tiles of the first request cover the entire user screen, this reading can meet the requirements of the user request. In the worst case, it may take four reads to meet the requirements of covering the entire screen. However, under the coverage ratio of 4096×4096, it is mostly possible to read and cover the user's screen at one time. After reading, there will be some redundant map tile data stored in memory, but this does not affect performance. On the contrary, because most of the user's operations are panning, the redundant data read can be returned to the client immediately, which also improves performance. Here, in order to further improve the response speed, you can also send instructions to the server to prefetch the data according to the user's operation (such as dragging, scrolling, etc.), so that the user can operate the map more smoothly. When you browse to a certain place, the data is returned directly from the memory, and the performance will be maximized.

In addition, when map tiles and metadata information are updated, if the size of the updated map tiles and metadata information is larger than the size of the map tiles and metadata information before the update If the size of the map tiles and metadata information is small, the updated map tiles and metadata information will be directly overwritten on the original storage location and the corresponding index table will be updated; if the size of the updated map tiles and metadata information is larger than that of the map tiles If the size of tiles and metadata information is large (or the original record does not exist and points to a replacement tile), write the updated map tiles and metadata information into the discarded data area (the discarded data area refers to the data area Areas that do not store actual data. For example, when the data is updated and added to the end of the file, the original data location becomes an abandoned area, which can be calculated based on the index information). If the abandoned data area cannot be found, the updated map will be The tile and metadata information is added to the end of the file, and the corresponding index table is updated at the same time.

After the update is completed, the server-side software will start the optimization task to optimize the data that is not fully optimized. The reference basis for optimization is that for each aggregated data record, Mortan code is used for sequential storage. The basic idea adopted by the update is to read the corresponding data records in the data and store them in a new file according to the storage optimization strategy. When the update is complete, the original file is deleted and a new file is renamed to replace it.

It should be noted that the metadata information of the map tile includes at least the MD5 code of the map tile, and the browser can know in time after the map tile is updated through the MD5 code of the map tile. Among them, the MD5 code occupies 16 bytes of space and is metadata information of a fixed length. Generally speaking, the MD5 code is directly stored at the end of each map tile data. In this way, when the browser side requests the map tile data, the server side will read out the MD5 code of the map tile data together. After reading out, the TAG data in the browser request is equal to the MD5 code of the tile data, which proves that the requested map tile data exists in the memory cache of the browser, and can directly return the code of 304NotModified to the request. And if the TAG data in the browser request is empty or not equal to the MD5 code of the tile data, it proves that the requested map tile data does not exist in the memory cache of the browser or the map tile data has been updated. At this time, you need to download the map tile data. When the server needs to update the map tile data, the server needs to recalculate the MD5 code of the new map tile data and update them together. This method simply and efficiently solves the drawbacks in the prior art that direct storage of small file map tile data cannot be efficiently updated, and the browser side cannot know after the local map tile data is updated.

Embodiment 2, a storage system for map tiles. The system provided by this embodiment will be described in detail below with reference to FIG. 2 .

Referring to FIG. 2 , the system provided by this embodiment includes a tile classification module 201 , a file construction module 202 , a folder construction module 203 , a next-level folder construction module 204 , an optimization module 205 and a storage module 206 .

Wherein, the tile grading module 201 is used to sequentially divide the electronic map into map tiles according to a predetermined size, and classify the map tiles according to a first predetermined rule to form a b-level data file, wherein b is an integer, and b is The values are 1, 2, 3....

Specifically, the tile grading module 201 sequentially divides the electronic map into a map tiles according to a predetermined size, and classifies the a map tiles to form a b-level data file, wherein the i+1th level data file The number of map tiles in the i-level data file is a predetermined multiple corresponding to the number of map tiles in the i-level data file, i is an integer, and the value of i is 0, 1...(b-1), and the map tiles in the i-level data file Form an m*n tile matrix, where m and n are both positive integers.

The file construction module 202 is used to construct c files according to the number of map tiles in the i-level data file according to the second predetermined rule, and store corresponding several map tiles in each of the constructed files, where i is Integer, the value of i is 0, 1...(b-1), c is a positive integer, and the value of c is 1, 2, 3....

Specifically, according to the value of the number of rows m and the number of columns n of the tile matrix in the i-th level data file, the file construction module 202 constructs c files in the i-th level data file, where c=ceiling(m/ k1)*ceiling(n/k2), wherein, 16≤k1≤m, 16≤k2≤n; wherein, c files form a g*h file matrix in the i-level data file, c=g*h, Both g and h are positive integers.

A folder construction module 203, configured to construct d folders according to a third predetermined rule if c is greater than the first predetermined number, and store corresponding several files in each of the constructed folders, wherein d is a positive integer, The value of d is 1, 2, 3....

Specifically, according to the value of the number of rows g and the number of columns h of the file matrix in the i-level data file, the folder construction module 203 constructs d folders in the i-level data file, and each of the folders constructed Deposit corresponding several files in, wherein, d=ceiling(g/k3)*ceiling(h/k4), k3 and k4 are all positive integers; Wherein, d folders form a l* in the i-level data file w folder matrix, d=l*w, both l and w are integers.

The next-level folder construction module 204 is used to construct e lower-level folders according to the fourth predetermined rule if d is greater than the second predetermined number, and each of the constructed lower-level folders stores a corresponding number of files folders, and so on, until the number of upper-level folders stored in the lower-level folder is less than or equal to the corresponding predetermined number.

Specifically, according to the number of rows l and the number of columns w of the folder matrix in the i-level data file, e lower-level folders are constructed in the i-level data file, and in each of the constructed lower-level folders Several corresponding folders are stored, where e=ceiling(l/k5)*ceiling(w/k6), where k5 and k6 are both positive integers.

The optimization module 205 is configured to reorder the storage order of the map tiles in each file by using Morton code encoding rules, so as to optimize the storage order of the map tiles.

The storage module 206 is used to store each file in the form of file header, index table and data content, wherein the file header stores the basic information and metadata description information of the map tiles in the file, and the index table stores the map in the file The tile index value, the data content includes map tiles and metadata information, and the metadata information includes at least the MD5 code of the map tiles. The storage module 206 is also used for updating map tiles and metadata information, if the size of the updated map tiles and metadata information is smaller than the size of the map tiles and metadata information before updating, then Overwrite the updated map tiles and metadata information directly to the original storage location and update the corresponding index table; if the size of the updated map tiles and metadata information is larger than the map tiles and metadata information before the update is large, write the updated map tiles and metadata information into the discarded data area, if no discarded data area is found, add the updated map tiles and metadata information to the end of the file, and at the same time Update the corresponding index table.

A storage method and storage system for map tiles provided by the present invention classify map tiles and aggregate small map tile files by constructing files and folders. When accessing map tiles, Compared with accessing each small map tile (usually a jpg image or png image) one by one in the prior art, the access speed is fast; when constructing files and folders, each file and each folder Appropriate map tile data is stored in the center to achieve the best balance of map tile access speed; when storing map tiles, the metadata information of the map tile includes the MD5 code of the map tile, and the local map tile When the map tile is updated, the browser client can know whether the requested data is updated in time when requesting to read, so as to avoid repeated downloads and improve the efficiency of map tile reading; in addition, when storing map tiles , store the source data information with a fixed size and no change in the front of the data area, and store the map tiles and metadata information that does not change in the back of the data area. This storage method, when the map tile is updated , which can avoid too many times of map tile data movement and improve the efficiency of data update; in addition, for the map tiles stored in each file, the Morton code encoding rule is used to reorder the storage order of the map tiles to optimize the map The storage order of the tiles, so that when any map tile is requested, its logically adjacent map tiles can be read with the least number of reads, reducing the number of map tile reads and improving Improve the efficiency of map tile reading.

In the description of this specification, descriptions referring to the terms "embodiment one", "example", "specific examples", or "some examples" mean that the specific methods, devices or features described in conjunction with this embodiment or example are included in In at least one embodiment or example of the invention. In this specification, the schematic representations of the above terms are not necessarily directed to the same embodiment or example. Furthermore, the described specific features, methods, devices or characteristics may be combined in any suitable manner in any one or more embodiments or examples. In addition, those skilled in the art can combine and combine different embodiments or examples and features of different embodiments or examples described in this specification without conflicting with each other.

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included in the protection of the present invention. within range.

Claims (9)

1. a kind of storage method of map tile, which is characterized in that the method includes：

S1, electronic map sequence is divided into map tile according to predefined size, and to the map tile after segmentation according to first Pre-defined rule is classified, and b level data files are formed, wherein b is positive integer, and the value of b is 1,2,3 ...；

S2, according to the quantity of map tile in i-stage data file, build c file according to the second pre-defined rule, structure it is every Corresponding several map tiles are stored in one file, wherein i is integer, and the value of i is 0,1 ... (b-1), and c is just whole The value of number, c is 1,2,3 ...；

If S3, c are more than the first predetermined quantity, d file, each described text of structure are built according to third pre-defined rule Corresponding several files are stored in part folder, wherein d is positive integer, and the value of d is 1,2,3 ...；

If S4, d be more than the second predetermined quantity, according to the 4th pre-defined rule build e next stage file, structure each The next stage file stores corresponding several files, and so on, until the upper level stored in next stage file The quantity of file is less than or equal to corresponding predetermined quantity；

The step S1 is specifically included：

Electronic map sequence is divided into a map tiles according to predefined size, and pre- according to first to a map tiles Set pattern is then classified, and b level data files are formed, wherein map tile quantity is i-stage data in i+1 level data file The corresponding prearranged multiple of map tile quantity in file, i are integer, and the value of i is 0,1 ... (b-1), and i-stage data file Middle map tile forms a m*n tile matrix, wherein m, n are positive integer；

The step S2 is specifically included：

According to the line number m of tile matrix in i-stage data file and the value of columns n, c is built in the i-stage data file A file, wherein c=ceiling (m/k1) * ceiling (n/k2), wherein k1 and k2 is positive integer, 16≤k1≤m, and 16 ≤k2≤n；Wherein, c file forms a g*h document matrix, c=g*h in i-stage data file, wherein g, h are just Integer；

The step S3 is specifically included：

According to the value of the line number g and columns h of document matrix in i-stage data file, d are built in i-stage data file File stores corresponding several files in each described file of structure, wherein d=ceiling (g/k3) * Ceiling (h/k4), wherein k3, K4 are positive integer；Wherein, d file forms a l*w text in i-stage data file Part presss from both sides matrix, and d=l*w, l, w are positive integer；

The step S4 is specifically included：

According to the line number l and columns w of file in i-stage data file, e next stage text is built in i-stage data file Part presss from both sides, and stores corresponding several files in each described next stage file of structure, wherein e=ceiling (l/ K5) * ceiling (w/k6), wherein k5, k6 are positive integer.

2. the storage method of map tile as described in claim 1, which is characterized in that each file is by file header, index Table and data content composition, wherein file header stores the essential information of map tile and metadata description information, rope in this document Draw table and store map tile index value in this document, data content includes map tile and metadata information, the metadata letter The MD5 codes of map tile are included at least in breath.

3. the storage method of map tile as claimed in claim 2, which is characterized in that do not change fixed size and Metadata information is stored in before data field, after map tile and the metadata information to change are stored in data field Face.

4. the storage method of map tile as described in claim 1, which is characterized in that using Morton code coding rules to every The storage order of map tile in one file is resequenced, to optimize the storage order of map tile.

5. the storage method of map tile as claimed in claim 3, which is characterized in that when map tile and metadata information are sent out When raw update, if updated map tile and the size of metadata information are than the map tile and metadata information before update It is big slight, then updated map tile and metadata information are directly overlie in original storage location and is updated corresponding Concordance list；If updated map tile and the size of metadata information than before update map tile and metadata information it is big It is small big, then it will more if can not find discarding data area by updated map tile and metadata information write-in discarding data area The end of file where map tile and metadata information after new are added to, while updating corresponding concordance list.

6. a kind of storage system of map tile, which is characterized in that the system comprises：

Tile diversity module for electronic map sequence to be divided into map tile according to predefined size, and presses map tile It is classified according to the first pre-defined rule, forms b level data files, wherein b is positive integer, and the value of b is 1,2,3 ...；

File builds module, and for the quantity according to map tile in i-stage data file, c is built according to the second pre-defined rule A file stores corresponding several map tiles in each described file of structure, wherein and i is integer, the value of i is 0, 1 ... (b-1), c are positive integer, and the value of c is 1,2,3 ...；

File builds module, if being more than the first predetermined quantity for c, d file, structure are built according to third pre-defined rule Corresponding several files are stored in each the described file built, wherein d is integer, and the value of d is 1,2,3 ...；

Next stage file builds module, if being more than the second predetermined quantity for d, according under the 4th pre-defined rule structure e First grade file presss from both sides, each described next stage file of structure stores corresponding several files, and so on, until next The quantity for the upper level file stored in grade file is less than or equal to corresponding predetermined quantity；

The tile diversity module is specifically used for：

Electronic map sequence is divided into a map tiles according to predefined size, and pre- according to first to a map tiles Set pattern is then classified, and b level data files are formed, wherein map tile quantity is i-stage data in i+1 level data file The corresponding prearranged multiple of map tile quantity in file, i are integer, and the value of i is 0,1 ... (b-1), and i-stage data file Middle map tile forms a m*n tile matrix, wherein m, n are positive integer；

The file structure module is specifically used for：

According to the line number m of tile matrix in i-stage data file and the value of columns n, c is built in the i-stage data file A file, wherein c=ceiling (m/k1) * ceiling (n/k2), wherein k1 and k2 is positive integer, 16≤k1≤m, and 16 ≤k2≤n；Wherein, c file forms a g*h document matrix, c=g*h in i-stage data file, wherein g, h are just Integer；

The file structure module is specifically used for：

According to the value of the line number g and columns h of document matrix in i-stage data file, d are built in i-stage data file File stores corresponding several files in each described file of structure, wherein d=ceiling (g/k3) * Ceiling (h/k4), wherein k3, K4 are positive integer；Wherein, d file forms a l*w text in i-stage data file Part presss from both sides matrix, and d=l*w, l, w are positive integer；

The next stage file structure module is specifically used for：

According to the line number l and columns w of file in i-stage data file, e next stage text is built in i-stage data file Part presss from both sides, and stores corresponding several files in each described next stage file of structure, wherein e=ceiling (l/ K5) * ceiling (w/k6), wherein k5, k6 are positive integer.

7. the storage system of map tile as claimed in claim 6, which is characterized in that the system also includes：

Memory module stores each file for the form according to file header, concordance list and data content, Chinese Part head stores the essential information of map tile and metadata description information, concordance list in this document and stores map tile in this document Index value, data content include map tile and metadata information, and the MD5 of map tile is included at least in the metadata information Code.

8. the storage system of map tile as claimed in claim 6, which is characterized in that the system also includes：

Optimization module, for carrying out weight to the storage order of the map tile in each file using Morton codes coding rule New sort, to optimize the storage order of map tile.

9. the storage system of map tile as claimed in claim 7, which is characterized in that the memory module is additionally operable to：

When map tile and metadata information update, if updated map tile and the size of metadata information ratio are more Map tile and metadata information before new it is big slight, then updated map tile and metadata information are directly overlie In storage location originally and update corresponding concordance list；If updated map tile and the update of the size of metadata information ratio The size of preceding map tile and metadata information is big, then discarding data is written in updated map tile and metadata information Area, if can not find discarding data area, the end of file where updated map tile and metadata information are added to, together The corresponding concordance lists of Shi Gengxin.