Geographic space tile compression method and device

Abstract

The invention provides a compression method and device for geographic space tiles, and relates to the technical field of compression, and the method comprises the steps: carrying out unified preprocessing and type-adaptive differential preprocessing on multiple types of geographic space tiles; compressing the tile data subjected to differential preprocessing by adopting corresponding compression strategies and compression depths according to types, and marking compression feature information corresponding to each type; extracting common data from the compressed tile data based on the compression feature information, and constructing a global shared dictionary and a plurality of grouped shared dictionaries; based on a QuadKey prefix in the global shared dictionary, converting a QuadKey coding sequence of the compressed tile data into a residual sequence, and compressing the residual sequence by using the grouping shared dictionary to generate residual coding data; and writing the residual error coding data and the compressed tile data into a storage structure comprising an index area and a data area. Based on the scheme, the compression redundancy of the geographic space tiles can be reduced, and the compatibility is improved.

Description

Technical Field

[0001] This disclosure relates to the field of compression technology, and more particularly to a method and apparatus for compressing geospatial tiles. Background Technology

[0002] Geospatial tiles typically include raster and vector types. Raster types include digital orthophotos and digital elevation models, while vector types include roads, buildings, and water bodies. As the application of geospatial tiles becomes more widespread, the requirements for their storage and processing are also increasing.

[0003] Currently, the storage and processing of geospatial tiles follow a standardized technical approach based on the pyramid multi-resolution model, employing either the complete XYZ Tile Coordinate System or Quadkey encoding. Storage media include: local file systems, cloud object storage—S3 (Simple Storage Service) or OBS (Object Storage Service), dedicated tile packages—MBTiles (Map Tile Storage Specification) or ESRI (Environmental Systems Research Institute) File Geodatabase, and the distributed database MongoDB. The compression strategy is as follows: raster tiles rely on native lossless or lossy compression of WebP (Web Picture), PNG (Portable Network Graphics), and GeoTIFF (Geographically Tagged Image File Format), with some additional lightweight compression of gzip (GNU's Not Unix zip, a recursive compression format) / bzip2 (Burrows–Wheeler zip version 2); vector tiles (MVT (Mapbox Vector Tile) / PBF (Protocol Buffer Format)) rely solely on the format's built-in binary geometry encoding for compression.

[0004] However, although current compression methods cover raster and vector tiles, the indexing rules, compression strategies, and storage media of different types of tiles are completely independent, requiring the development of multiple processing modules for adaptation. Moreover, the storage volume after compression is relatively large, especially in scenarios with large-scale and high-precision data, where storage costs increase exponentially. Therefore, current compression processing methods have poor compatibility and high storage redundancy. Summary of the Invention

[0005] To overcome the problems existing in related technologies, this disclosure provides a method and apparatus for compressing geospatial tiles, which can reduce the compression redundancy of geospatial tiles and improve compatibility.

[0006] According to a first aspect of the present disclosure, a method for compressing geospatial tiles is provided. The method includes: performing unified preprocessing and type-adaptive differentiated preprocessing on multiple types of geospatial tiles; compressing the differentiated preprocessed tile data according to the type using a corresponding compression strategy and compression depth, and marking compression feature information corresponding to each type; extracting common data from the compressed tile data based on the compression feature information, and constructing a global shared dictionary and multiple group shared dictionaries; storing the QuadKey prefix of the quadtree key value as the basis for residual coding in the global shared dictionary; storing shared data fragments extracted from the common data for compressing residual coding in the group shared dictionaries; converting the QuadKey encoded sequence of the compressed tile data into a residual sequence based on the QuadKey prefix in the global shared dictionary, and compressing the residual sequence using the group shared dictionaries to generate residual encoded data; and writing the residual encoded data and the compressed tile data into a storage structure containing an index area and a data area; the index area is used to locate the tile data.

[0007] According to a second aspect of the present disclosure, a geospatial tile compression apparatus is provided, comprising: a preprocessing module, a compression module, a dictionary construction module, a residual coding module, and a storage module; the preprocessing module is used to perform unified preprocessing and type-adaptive differentiated preprocessing on multiple types of geospatial tiles; the compression module is used to compress the differentiated preprocessed tile data according to the type using corresponding strategies and compression depths, and to mark the compression feature information corresponding to each type; the dictionary construction module is used to extract common data from the compressed tile data based on the compression feature information, and construct a globally shared dictionary and multiple sub-dictionaries. The system consists of a group shared dictionary; a global shared dictionary storing the QuadKey prefix, a quadtree key value used as the basis for residual coding; a group shared dictionary storing shared data fragments extracted from common data for compressing residual coding; a residual coding module, used to convert the QuadKey encoded sequence of compressed tile data into a residual sequence based on the QuadKey prefix in the global shared dictionary, and to compress the residual sequence using the group shared dictionary to generate residual encoded data; a storage module, used to write the residual encoded data and compressed tile data into a storage structure containing an index area and a data area; and an index area for locating tile data.

[0008] According to a third aspect of the present disclosure, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the geospatial tile compression method described in the first aspect.

[0009] According to a fourth aspect of the present disclosure, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the geospatial tile compression method described in the first aspect.

[0010] The technical solutions provided by the embodiments of this disclosure may include the following beneficial effects: In this embodiment, firstly, all types of tiles undergo unified preprocessing, and then differentiated preprocessing is performed on the preprocessed tiles to ensure that tiles of different origins, types, and specifications can be standardized before encoding, eliminating differences caused by different origins. Secondly, customized deep compression of tile data can be performed by combining compression methods adapted to each type, and compression feature information of each type of data is marked. Then, based on the marked compression feature information and common data extracted from the compressed tile data, a two-level shared dictionary system of global and grouped components is constructed. Based on the QuadKey prefix in the global shared dictionary, the QuadKey encoding sequence of the compressed tile data is converted into a residual sequence, and the residual sequence is compressed using the grouped shared dictionary to generate residual encoded data, i.e., the index of the tile data is also compressed. Finally, the two-level shared dictionary and the index residual encoding are merged, and the residual encoded data and the compressed tile data are written into a storage structure containing an index area and a data area. This reduces storage redundancy and cost, improves the compression compatibility of various types of tiles, and has strong scalability.

[0011] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0012] The accompanying drawings, which are incorporated in and form part of this disclosure, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.

[0013] Figure 1 A flowchart illustrating a method for compressing geospatial tiles as provided in this embodiment of the disclosure.

[0014] Figure 2 This is a schematic diagram of a two-level shared dictionary system provided in an embodiment of this disclosure.

[0015] Figure 3 This is a schematic diagram of residual coding provided in an embodiment of the present disclosure.

[0016] Figure 4 This is a schematic diagram of residual decoding provided in an embodiment of the present disclosure.

[0017] Figure 5 This is a hardware structure diagram of a computer device containing a geospatial tile compression device according to an embodiment of this disclosure.

[0018] Figure 6 This is a schematic diagram of a compression device for geospatial tiles provided in an embodiment of the present disclosure.

[0019] Figure 7A schematic diagram of another geospatial tile compression device provided in an embodiment of this disclosure. Detailed Implementation

[0020] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0021] The terminology used in this disclosure is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. The singular forms “a,” “the,” and “the” as used in this disclosure and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

[0022] It should be understood that although the terms first, second, third, etc., may be used in this disclosure to describe various information, such information should not be limited to these terms. These terms are used only to distinguish information of the same type from one another. For example, without departing from the scope of this disclosure, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0023] The embodiments of this disclosure will now be described in detail.

[0024] like Figure 1 As shown, Figure 1 A flowchart of a method for compressing geospatial tiles provided in this disclosure includes the following steps S101 to S105.

[0025] S101. Perform unified preprocessing and type-adaptive differentiated preprocessing on various types of geospatial tiles.

[0026] The types of tiles include: raster and vector; the raster types include: image raster and numerical raster.

[0027] In this embodiment of the disclosure, the unified preprocessing includes: unifying the coordinate system of various types of geospatial tiles, unifying the size of various types of geospatial tiles, and unifying the metadata of various types of geospatial tiles.

[0028] The metadata for a tile includes: the tile type, the QuadKey prefix (quadtree key prefix), and the preprocessing precision. The metadata is in binary format and consists of 32 bytes.

[0029] It should be noted that differentiated preprocessing allows for a unified and reusable architecture to prepare data for subsequent intelligent tile compression and to establish a foundation for structured analysis, ensuring that all tiles can be processed using the same framework and maximizing compatibility with tiles of various types and sources.

[0030] S102. For the preprocessed tile data, compress it according to the type using the corresponding compression strategy and compression depth, and mark the compression feature information corresponding to each type.

[0031] S103. Based on the compressed feature information and the compressed tiles, extract common data from the compressed tile data to construct a global shared dictionary and multiple grouped shared dictionaries.

[0032] Specifically, the global shared dictionary stores the QuadKey prefix, which serves as the basis for residual coding; the group shared dictionary stores shared data fragments extracted from common data and used for compressing residual coding. In other words, a two-level shared dictionary system comprising a global shared dictionary and a group shared dictionary is constructed.

[0033] S104. Based on the QuadKey prefix in the global shared dictionary, the QuadKey encoded sequence of the compressed tile data is converted into a residual sequence, and the residual sequence is compressed using the grouped shared dictionary to generate residual encoded data.

[0034] S105. Write the residual encoded data and the compressed tile data into a storage structure containing an index area and a data area.

[0035] The index area is used to locate tile data.

[0036] This disclosure provides a method for compressing geospatial tiles. First, all types of tiles undergo unified preprocessing. Then, the preprocessed tiles are subjected to differentiated preprocessing to ensure that tiles of different origins, types, and specifications can be standardized before encoding, eliminating differences caused by different sources. Second, customized deep compression of tile data can be performed by combining compression methods suitable for each type, and compression feature information of each type of data is marked. Then, based on the marked compression feature information and common data extracted from the compressed tile data, a two-level shared dictionary system of global and grouped components is constructed. Based on the QuadKey prefix in the global shared dictionary, the QuadKey encoding sequence of the compressed tile data is converted into a residual sequence, and the residual sequence is compressed using the grouped shared dictionary to generate residual encoded data, i.e., the index of the tile data is also compressed. Finally, the two-level shared dictionary and the index residual encoding are merged, and the residual encoded data and the compressed tile data are written into a storage structure containing an index area and a data area. This reduces storage redundancy and cost, improves the compression compatibility of various types of tiles, and has strong scalability.

[0037] Optionally, in the geospatial tile compression method provided in this embodiment, the unified preprocessing of S101 may specifically include the following: S101a to S101d.

[0038] S101a. Unify the coordinate systems of tiles from different sources to the same map projection coordinate system, and classify the layers according to the original accuracy of each type of tile.

[0039] For example, all tiles of various types from different sources are forced to use the Web Mercator (Web Mercator projection coordinate system) with code EPSG:3857 (European Petroleum Survey Group).

[0040] The tiles are divided into levels according to their original precision. For example, the mapping relationship between precision and level is shown in Table 1. As shown in Table 1, the Level of Detail is divided into 17 levels. The original precision includes three dimensions: map width and height, in pixels; ground resolution, in meters per pixel; and map scale, which is 96 dpi (dots per inch).

[0041] Table 1 - Accuracy and Hierarchy Mapping Relationship Based on the mapping relationship, the geographic range of multiple types of tiles in the same geographic area can be aligned, so that the geographic alignment error between different types of tiles is ≤0.0001°, which is the sub-meter level error under Web Mercator.

[0042] It is understandable that, under the premise of unifying the map projection coordinate system, the residual coding process can be grouped by region and level.

[0043] S101b: Unify the pixel size of the grid tiles to the preset pixel specifications, and match the vector tiles to the geographic spatial range of the grid tiles at the same level.

[0044] It should be noted that the preset pixel specifications can be custom specifications, but they must be consistent globally.

[0045] For example, the pixels of the grid tiles are uniformly 256×256 pixels. The tile cutting algorithm can divide the reprojected large image into 256×256 pixel tiles.

[0046] Vector tiles match the geospatial extent of raster tiles at the same level. Specifically, the geospatial extent of a vector tile needs to strictly match that of raster tiles with the same index.

[0047] It is understandable that by standardizing tile sizes, a foundation for cross-type commonality analysis is established. For example, under the same geographic spatial range and pixel grid, the problem of being unable to extract common data of cross-type tiles when building a shared dictionary due to differences in tile sizes of different types can be avoided.

[0048] S101c generates uniform metadata for each tile.

[0049] The metadata includes at least: tile type identifier, QuadKey prefix, and preprocessing precision field.

[0050] The QuadKey prefix indicates the first N characters of the QuadKey string of the tile, where N is a positive integer. Preprocessing precision indicates the original ground precision represented by the tile data.

[0051] It is understandable that unified metadata can provide index support for subsequent shared dictionaries to filter common data by type and level.

[0052] Based on this scheme, spatial alignment, structural alignment, and descriptive alignment are performed using a unified preprocessing framework, which eliminates the uncertainty of the data source. Spatial alignment ensures that different types of tiles can completely cover the same real-world area at the same geographical location and scaling level. Structural alignment ensures that the data is comparable and correlated at the pixel / vector level. Descriptive alignment ensures that all tiles carry unified and concise metadata, facilitating systematic management and retrieval. This establishes a highly consistent and standardized data foundation layer, providing support for subsequent complex cross-type analysis to extract common information, residual coding, and compression.

[0053] Optionally, in S101 above, the tile data of each type can be preprocessed differently according to the tile type. Specifically, after uniform preprocessing, it includes S101d to S101f below.

[0054] S101d performs color space format conversion on image-type raster tiles.

[0055] For example, the RGB (Red Green Blue) color space format will be used as an example for illustration.

[0056] Specifically, RGB format conversion is performed on image raster tiles. Unifying the format of image raster tiles can eliminate complex issues such as color management and high bit depth processing, thereby simplifying the process. Standardization through RGB format conversion not only ensures that images from any source can be processed uniformly in subsequent workflows, but also reduces interference from irrelevant color space and channel information on compression, thus improving subsequent compression efficiency.

[0057] S101e: Perform numerical index conversion on numerical raster tiles.

[0058] An elevation value to 16-bit index matrix conversion is performed on numerical raster tiles. Firstly, this maps the tile values ​​to integer space, creating conditions for lossless compression of the numerical raster tiles. Since general compression algorithms are far more efficient at compressing integer data than floating-point numbers, the elevation value to 16-bit index matrix conversion is linear, without losing any precision. The information is encapsulated in the 16-bit index matrix, maintaining its integrity and thus ensuring lossless compression. Secondly, numerical raster tiles are typically high-bit-depth; for example, a 32-bit floating-point number generally occupies 4 bytes, while the converted 16-bit integer only occupies 2 bytes. This halving of storage volume controls data expansion, providing a well-organized data foundation for subsequent interval feature analysis.

[0059] S101f: Filter the attributes of vector tiles.

[0060] Perform core attribute filtering on vector tiles, where the core attributes include: type identifier and priority.

[0061] It's important to note three points: First, since the volume expansion of vector tile data primarily stems from redundant attributes, core attribute filtering can remove over 80% of useless attributes, forcibly retaining the core attributes of the tile data. This significantly reduces the input burden for subsequent encoding, thereby minimizing subsequent compression redundancy. Second, by mandating type identifiers and priorities, we inject clear semantics and rendering intent into the data, enabling targeted processing for subsequent MVT structure optimization and geometric instruction feature analysis. Third, it provides a unified and clean compression input template for vector tiles from different sources or with chaotic attribute structures.

[0062] Based on this scheme, a unified, streamlined, and scalable preprocessing framework was constructed. Advanced analysis functions were stripped away, leaving only the most basic and universal format conversion tasks that must be completed before the three types of data enter the encoding process for standardized encoding input. This provides high-quality, redundant raw materials for subsequent analysis and intelligent compression.

[0063] Optionally, in the method for compressing geospatial tiles provided in this embodiment of the present disclosure, the above-mentioned S103 may specifically include S103a to S103c.

[0064] S103a. For image-type raster tiles, the first compression is performed based on encoder encoding adapted to the image-type scene, and the second compression is performed based on the first compression depth of the general lossless compression method. After the second compression, color redundancy features are marked.

[0065] Image-based scenes include both lossy and lossless scenes. Lossy scenes typically target natural landscapes and remote sensing imagery, requiring the highest compression ratio while maintaining controllable visual loss. Lossless scenes are typically designed for materials, medical images, and precise maps, requiring pixel-level fidelity.

[0066] The following explanation uses the general lossless compression method, the compression algorithm LZMA (Lempel–Ziv–Markov chain Algorithm), with a first compression depth of any one of compression levels from 6 to 8.

[0067] For example, for image-type raster tiles, WebP (Web Picture) image codec is first used for encoding in a lossless or lossy encoding method, i.e., the first compression. Then, the WebP-encoded tiles are compressed a second time based on LZMA compression levels 6 to 8. After WebP combined with LZMA compression, the color redundancy within the image is identified and quantified, and the color redundancy features of the compressed image-type data are marked.

[0068] For example, color redundancy features are defined as the percentage of consecutive pixel blocks. The percentage of consecutive pixel blocks indicates the total area of ​​consecutive adjacent pixel regions with identical color values ​​(or within the tolerance range of lossy compression) in an image, representing the proportion of the entire image area.

[0069] It is understandable that compression processing for image-type raster tiles, through a combination of lossy / lossless compression, feature marking, and template extraction, can not only minimize the storage of repetitive image data but also prepare for subsequent global deduplication during storage. Compression based on WebP and LZMA at the single-file level, followed by deduplication at the cross-file library level through color redundancy and duplicate template extraction, ultimately reconstructs a pixel warehouse with a large amount of duplicate data into a high-efficiency database composed of unique templates and reference relationships, achieving an order-of-magnitude improvement in storage and transmission efficiency.

[0070] S103b: For numerical raster tiles, perform initial compression based on an encoder adapted to numerical scenarios, perform secondary compression based on a second compression depth using a general lossless compression method, and mark numerical interval features after secondary compression.

[0071] In the numerical scene, the encoder's initial compression is lossless, with a first compression depth higher than the second compression depth. Numerical interval characteristics indicate the regular intervals in the distribution of data values.

[0072] For example, the second compression depth is compression level 5 of the LZMA compression algorithm.

[0073] It should be noted that in different fields, such as engineering data or scientific experimental data, the values ​​are not continuously and randomly distributed, but appear at fixed intervals around some benchmark values.

[0074] For example, for numerical raster tiles, a lossless initial compression can be performed using a JXR (JPEG Extended Range) encoder. Then, a second compression is performed on the numerical raster tiles processed by the JXR encoder using LZMA compression level 5. After the second compression, the numerical interval characteristics of the numerical raster tiles are marked. For example, an elevation interval of 1 cm or 10 cm.

[0075] It should be noted that the original numerical matrix can be transformed into a set of indices and parameters using numerical interval features, and all pixels in the entire dataset that share the same numerical interval features can be grouped together.

[0076] Specifically, by using JXR lossless encapsulation, LZMA compression, and the extraction of numerical interval features from the same interval index template, the stored bytes are transformed into stored knowledge. Instead of treating raster data as a black box set of pixels, raster data is regarded as a signal carrying a clear physical meaning and generation rules. By automatically discovering and utilizing the rules of intervals, intelligent compression can be achieved.

[0077] This method enables high-fidelity storage and efficient compression of high dynamic range and high bit depth numerical raster tiles. While ensuring absolute loss of numerical accuracy, it leverages the inherent mathematical laws of the data to reorganize pixels into a parametric model of intervals and indices. It is suitable for numerically intensive fields with stringent accuracy requirements, strong data regularity, and massive data volumes.

[0078] S103c: For vector tiles, perform structured processing based on the encoder adapted to the vector scene, use a second compression depth based on a general lossless compression method to compress the structured data, and mark the geometric instruction features after compression.

[0079] For example, the structure of vector data can be optimized using the MVT (Mapbox Vector Tile) structure. MVT optimization includes coordinate deduplication and attribute filtering. The MVT-optimized tiles are then compressed against the numerical raster tiles processed by the JXR encoder using the LZMA compression algorithm at compression level 5. After compression, geometric instruction features are labeled. These geometric instruction features are essentially drawing instructions, typically including rectangles, road lines, polygons, etc.

[0080] Specifically, for coordinate deduplication, a global coordinate pool is established when encoding a single tile. The geometry no longer stores the original coordinate sequence, but instead stores an index sequence pointing to the coordinate pool. Coordinate deduplication eliminates the redundant storage of shared boundaries and common vertices within vector tiles. This method is highly effective for data such as administrative divisions and building grids.

[0081] Specifically, for attribute filtering, based on visualization requirements, filtering attributes can remove redundant analytical and descriptive attributes from the source data, retaining only the key-value pairs required for rendering, which can significantly reduce the size and complexity of the attribute dictionary.

[0082] After compression, instruction sequence patterns can be marked, such as the combination rules of instructions indicating specific geometric patterns; parameter statistical characteristics can be marked, such as common incremental values ​​of instructions (e.g., standard road width, standard building size); and instruction frequency can be marked, such as the frequency of different instruction patterns within a single tile and across tiles.

[0083] It is understandable that, in compressing vector tiles, this solution establishes a new paradigm for the storage, transmission, and distribution of large-scale vector tiles by combining MVT structure optimization, LZMA high-efficiency compression (level 5), and an architecture from geometric feature marking to high-frequency segment extraction. This addresses the problems of massive tile data volume, numerous internal repetitive patterns, and low transmission efficiency in vector map services. Through encoding optimization, intelligent compression, and feature-based reorganization, this solution lays the foundation for global data redundancy reduction and streaming transmission optimization. It no longer treats data as a passive stream of bytes to be compressed, but rather as structured information that can be deeply understood, reconstructed, and reused. This represents a leap from general compression to semantic (geometric) level compression, solving the problem of distributing large-scale geospatial data.

[0084] Based on this scheme, a compression framework was constructed that integrates original tile format adaptation, LZMA deep compression, and compression feature marking. This framework allows for differentiated compression based on tile type. The categorized and phased compression strategy enables each tile to achieve a good balance between its data characteristics and compression efficiency, resulting in better overall compression performance and reduced redundancy. This provides adaptability support for subsequent shared dictionary grouping and residual coding management files.

[0085] Optionally, in a method for compressing geospatial tiles provided in this embodiment of the present disclosure, the above-mentioned S103 may specifically include S103a to S103d.

[0086] S103a: Extract globally shared repeating pixel templates based on color redundancy features, extract group-shared same-interval index value templates based on numerical interval features, and extract group-shared high-frequency instruction fragments based on geometric instruction features.

[0087] In the subsequent construction of a globally shared dictionary, for image-type raster tiles, globally shared WebP repeating pixel templates can be extracted from the compressed tiles as common data based on the color redundancy features marked after compression. WebP repeating pixel templates indicate visually repetitive local areas in the image where the WebP-encoded data blocks are also highly similar or even identical. Specifically, highly redundant blocks marked in the image, such as solid color matrices and gradient regions, are subjected to feature comparison and clustering to identify identical pixel blocks. These repeating pixel blocks are extracted into independent, globally shared template files, and the region is replaced with a lightweight reference pointer in the original image. This approach is particularly suitable for processing ultra-large-scale, highly repetitive image asset libraries.

[0088] For numerical raster tiles, based on the numerical interval features of the compressed labels, the same interval index value templates of the group-shared dictionary can be extracted from the compressed tiles as common data during the construction of the group-shared dictionary. This is a deep structure optimization and encoding optimization based on regularity.

[0089] For vector tiles, frequently occurring geometric command patterns are identified and extracted into reusable macro commands and geometric templates. High-frequency vector command fragments are predefined, compact binary blocks of vector drawing commands. Specifically, based on geometric command features, geometric elements in massive amounts of tiles can be clustered to abstract general patterns. These high-frequency patterns are encoded into highly optimized, independent command fragments, which can be considered as miniature, dedicated MVT subsets. A global fragment library is established, and in the original tile data, the original command sequences matching high-frequency patterns are replaced with a simple fragment ID reference, thus enabling services for ultra-large-scale, highly repetitive vector maps.

[0090] Understandably, the patterns in the tile data discovered during compression are transformed into standardized data blocks that can be directly used in subsequent encoding, significantly improving the efficiency of tile data reuse and the compression rate of encoding.

[0091] S103b. Construct a globally shared dictionary to store common data across types and levels.

[0092] The common data includes: format identifier and QuadKey prefix information.

[0093] Specifically, the globally shared dictionary adopts a key-value pair structure, where the key corresponds to a common data identifier and the value corresponds to a binary encoded fragment. The core data included in the global encoded dictionary includes: a fixed-format identifier and a QuadKey prefix.

[0094] For example, the format fixed identifier may include: the WebP file header (89504E47) of the image raster tiles and the protocol identifier (tile information) of the vector MVT.

[0095] It should be noted that for image tiles encoded with WebP, the fixed byte sequence at the beginning of the file is usually completely repeated, so only one copy of the global dictionary needs to be stored, and each tile only needs to store a short reference to that segment.

[0096] The QuadKey prefix can include the number of prefixes provided by the user, which can be used as a baseline reference value for residual coding.

[0097] It should be noted that the QuadKey prefix can serve as the spatial index base for the entire dataset and be stored in the global dictionary. During subsequent QuadKey residual encoding, the complete QuadKey tile can be encoded as a differential code of the prefix ID and the remaining portion from the global shared dictionary, thus compressing the spatial index itself. The global shared dictionary is updated very infrequently, typically only when there are fundamental changes to the data format or the underlying spatial framework.

[0098] S103c: Based on tile type, tile level, and compressed feature information, a grouped shared dictionary is constructed to store common data within a group based on the same compressed feature information.

[0099] Specifically, the shared dictionary can be grouped according to tile type and level.

[0100] For example, numerical rasters construct a dedicated index dictionary for 1cm interval data based on the numerical interval features marked during the compression process. A dedicated index dictionary is also constructed for 10cm interval data.

[0101] For example, for elevation data groups marked with a numerical interval of 1 cm, the elevation range of all tiles within the group is determined, and a global "value-index" dictionary covering all 1 cm interval elevations is constructed for that group. Any elevation value in any tile within the group is no longer stored as a floating-point number, but rather as an integer index within this group's dictionary, eliminating the storage of duplicate values ​​between tiles.

[0102] For image groups, high-frequency color palettes, JPEG quantization tables, WebP encoding parameters, etc., are generated as high-frequency parameter templates; for vector groups, high-frequency road style templates (line width, color), building fill patterns, and vector path data of complex symbols are generated as high-frequency parameter templates.

[0103] For repetitive geographic features, taking cross-type examples, a bridge might have a fixed texture in imagery, a fixed outline in vector graphics, and a fixed height in a digital elevation model. A unique geographic feature ID can then be generated and stored in a grouping dictionary. Each type of tile only needs to reference this ID and store its own independent residual information, thus enabling the encoding of repetitive geographic features.

[0104] It is understandable that the grouping dictionary can be incrementally expanded as new regional data is added and local data is updated.

[0105] S103d: Perform target processing on the global shared dictionary and the group shared dictionary.

[0106] The target processing includes: removing duplicate entries from the dictionary through deduplication, compressing the dictionary itself, and reserving space for dictionary expansion.

[0107] It is understandable that after the global shared dictionary is constructed according to S102a above, and after the group shared dictionary is constructed according to S102b above, duplicate entries in the global shared dictionary and the group shared dictionary can be removed by hash deduplication until the dictionary size is less than or equal to 5% of the average size of a single tile.

[0108] After deduplication, the globally shared dictionary and the grouped shared dictionary can be compressed using LZMA compression. The mapping relationship between the dictionary before and after compression is recorded.

[0109] During decompression, the original data can be quickly located using dictionary IDs and offsets.

[0110] It should be noted that when building the dictionary, 20% of the entries can be reserved to facilitate the access of common data for new tile types, such as oblique photography tiles. When expanding, the group dictionary only needs to be updated without affecting the global structure.

[0111] Figure 2 This is a schematic diagram illustrating the structure of a two-level shared dictionary system provided in an embodiment of this disclosure. For example... Figure 2 As shown, it includes a global shared dictionary and group shared dictionaries. The global shared dictionary centrally stores the format identifiers and encoding reference values ​​shared by various types of tiles, while each group dictionary stores its corresponding data template according to the tile type and its level. This ensures that the most suitable reference data for the current tile can be found quickly during encoding, improving the dictionary query efficiency and encoding accuracy.

[0112] Based on this scheme, a two-level linkage framework combining a globally shared dictionary and a grouped shared dictionary is constructed on the basis of customized deep compression of multi-type tiles. By extracting common redundant data across tiles and types, such as fixed format identifiers, high-frequency parameter templates, and repeated geographic feature codes, the unchanging and high-frequency parts of the data are precipitated into a reusable dictionary, so that the changing and specific data exist in the form of residuals. Thus, while ensuring information integrity, tiles can be stored once and reused multiple times. After compression, the overall storage redundancy can be further reduced. At the same time, the dictionary can dynamically adapt to tile type expansion and data updates, and can seamlessly connect with subsequent residual coding and scheduling adaptation.

[0113] Optionally, in the method for compressing geospatial tiles provided in this embodiment of the present disclosure, the above-mentioned S104 may specifically include S104a to S104c.

[0114] S104a. The QuadKey sequence of the tile is encoded into a numerical sequence, and the numerical difference between adjacent tiles is calculated as the residual.

[0115] For example, Figure 3 This is a schematic diagram of residual coding provided in an embodiment of the present disclosure, combined with... Figure 3 For all tiles that need to be encoded within a region, sort them lexicographically according to the QuadKey. Convert the original QuadKey sequence in string form after sorting into decimal values. Starting from the second tile, calculate the difference between its QuadKey integer value and the QuadKey integer value of the previous tile. Based on the QuadKey hierarchy, redundancy can be reduced by calculating the difference (i.e., residual) between adjacent tiles.

[0116] S105b: Residuals are pre-processed by encoding based on a shared dictionary system.

[0117] The encoding preprocessing includes replacing recurring residual patterns with index references of the corresponding group shared dictionary, and using the QuadKey prefix in the global shared dictionary as a reference value for residual calculation.

[0118] It's important to note that for a set of tiles within the same region (e.g., a province), the QuadKey shares a common prefix. This common prefix is ​​stored in a globally shared dictionary and assigned a very short ID, such as 1 byte. During subsequent encoding, for the first tile within the region, the QuadKey is no longer stored in its entirety; instead, it's stored as a dictionary ID and a suffix residual. Subsequent tiles then use this residual to calculate their residuals, avoiding the duplication of long prefixes and achieving the first level of optimization.

[0119] Specifically, in highly regular data grids or specific regions, the QuadKey differences between adjacent tiles frequently exhibit fixed patterns, such as +1, +256, -255, etc. These frequently occurring residual values ​​are stored in a grouping dictionary. When encountering these high-frequency residuals, instead of storing the original 3-byte value, a 2-byte dictionary index is stored, such as a 1-byte dictionary ID and a 1-byte entry index, achieving a second level of optimization. Due to the high pattern repetition rate, the spatial index flow can be significantly compressed.

[0120] S105c: Based on the tile type identification information and the residual after encoding preprocessing, residual encoding is performed to generate residual encoded data.

[0121] Specifically, residuals exceeding a preset range are split to ensure that a single encoding operation does not exceed a preset number of bytes. It should be noted that in the hierarchical encoding provided in this disclosure, for non-contiguous tiles or tiles spanning multiple levels, under ideal conditions, the difference is within the range of ±8388607 (a 3-byte signed integer). Therefore, a 3-byte signed integer (the highest bit is the sign bit, and the remaining 23 bits represent the value) can be used to store this difference. However, in cases of large jumps, such as... Figure 3 As shown, if the difference exceeds the aforementioned range, one or more virtual intermediate QuadKeys are inserted. By adding intermediate-level QuadKeys (e.g., jumping to the common ancestor first and then descending step by step), the process is broken down into multiple steps, ensuring that the difference in each step is within the 3-byte range. During decoding, these intermediate QuadKeys are reconstructed as virtual nodes but do not correspond to the actual tile data.

[0122] It should be noted that the group-shared dictionary in this embodiment includes a hierarchical residual identifier. When encoding the residual, an additional 1-2 bits are used to identify the hierarchical relationship (same level, child level, parent level) between the current tile and the preceding tile. This facilitates rapid location and verification during decoding.

[0123] It should be noted that bit-level multiplexing is used for cross-type tiles. When data packets contain a mixture of image, vector, and numerical tiles, the QuadKey sequences for each type of tile are interleaved. Therefore, the highest bit (or two bits) of the 3 bytes (24 bits) storing the residual is borrowed as the type identifier. This allows the decoder to determine the current tile type instantly upon reading the difference without needing to look up a table, thus enabling streaming, mixed parsing.

[0124] Based on this scheme, a QuadKey residual encoding method is disclosed, which can compress the QuadKey index, which requires an average of 1 to 10 bytes per tile, to an average of 1-2 bytes. The decoded sequence is naturally ordered, which facilitates fast querying and batch loading based on spatial range. The built-in verification and error correction mechanism can tolerate a certain degree of data error, thereby improving the reliability of the entire data distribution pipeline in the geospatial domain.

[0125] Optionally, in a geospatial tile compression method provided in this embodiment, the storage structure includes a header area, an index area, and a data area. The header area stores version information, dictionary identifiers, total number of tiles, and compression type markers. The index area stores the mapping relationship between residual encoding results and data location information in hierarchical chunking. The data area uses a block storage mechanism to store the compressed tile data in QuadKey order, adding data verification information to each data block during storage.

[0126] For example, the compression type is marked as 0x01=LZMA, 0x02=Webp.

[0127] All index entries are stored strictly in the order of the QuadKey residual encoding values. Since the residual encoding itself is ordered, this makes the entire index area also ordered. During a search, to find the tile corresponding to a target QuadKey: the global dictionary is located using the dictionary identifier in the header; the base prefix of the target QuadKey and the calculated residual value are decoded; a binary search is performed on this residual value in the index area to find the corresponding index entry.

[0128] In other words, the data area uses block-based storage. During a lookup, the data offset and data length are read from the data location information of the entry. The data offset is the offset relative to the starting point of a data block in the data area.

[0129] For example, block storage of 64KB / block can be used, with each 64KB block tightly packing compressed data from multiple tiles. A 4-byte CRC32 checksum is reserved between every two 64KB data blocks to ensure data integrity.

[0130] It's important to note that hard drives and operating systems currently read and write data in units of "pages" (typically 4KB). The 64KB block size is aligned with the I / O system, reducing the number of I / O operations when randomly reading small files. This facilitates caching on the client or server side, allowing for pre-reading or eviction in blocks. CRC32 checksums are performed on a block-by-block basis. If a block is corrupted due to network transmission or disk damage, a failed checksum can pinpoint the corrupted 64KB block without rendering the entire file unusable. Combined with an index, it's possible to identify which tiles were affected by the bad block, allowing for individual re-requests of those tiles.

[0131] Figure 4 This is a schematic diagram of residual decoding provided in an embodiment of the present disclosure. During decoding, as follows: Figure 4 As shown, the current QuadKey can be calculated backwards from the preceding QuadKey value and the residual. Simultaneously, the decoder deduces the actual layer of the tile based on the layer difference flag and the calculated QuadKey, checking if the deduced layer matches the layer indicated by the flag. If they don't match, it indicates a possible error during encoding or transmission, triggering an error correction mechanism. This mechanism recalculates the residuals of adjacent tiles and attempts to recalculate the residuals of other successfully decoded tiles spatially adjacent to the current tile. By inferring the most likely correct QuadKey using context information, this mechanism ensures that the error between the final recovered tile geographic range and the correct location is less than or equal to 1 pixel within the same layer when scheduling tiles.

[0132] The header area can be understood as the file's "identity card" and "master directory," with a fixed size for easy reading by the parser. The index area is the file's "navigation system," enabling instantaneous mapping from logical tile IDs (QuadKeys) to physical data locations. The data area is the file's "warehouse," actually storing the compressed tile binary data.

[0133] Based on this scheme, a three-tiered storage structure of header, index area, and data area is used to store tile data. This allows for rapid identification of file attributes through metadata stored in the header area. Random access to tiles and rapid tile location are achieved through the index mapping relationships stored in the index area, enabling the reading of compressed tile data stored in the data area.

[0134] Optionally, in a method for compressing geospatial tiles provided in this embodiment of the present disclosure, after step S105, at least one of steps S106 to S108 may be included.

[0135] S106. Upon receiving a scheduling request, locate the tile data based on the memory caching algorithm and the hierarchical fragmentation information of the index area.

[0136] Among them, the memory caching algorithm is LRU (Least Recently Used).

[0137] It is understandable that low-level indexes reside in memory to ensure immediate response at global overview and medium zoom levels, while caching recently accessed index shards conforms to the spatiotemporal locality of user browsing. Global lookup is decomposed into shard locations through hierarchical mapping, and then the search is performed within each shard. In other words, multi-level indexes can accelerate tile location, making the tile location time less than or equal to 10ms.

[0138] S107. Based on the tile type of the scheduling request, decompress on demand using a group dictionary that matches dictionary identifiers and compression tags, and a decompression algorithm.

[0139] It's understandable that full decompression usually wastes CPU resources, increases memory usage, and prolongs response time.

[0140] It is understandable that by decompressing on demand, problems such as excessively long loading times associated with fully decompressed models can be avoided.

[0141] S108. Select the encoding strategy and compression level according to network conditions.

[0142] Specifically, in scenarios with low bandwidth transmission networks, residual coding and lightweight compression are selected, while in scenarios with high bandwidth transmission networks, full coding and deep compression are selected.

[0143] For example, when the bandwidth is less than the first bandwidth, choosing residual coding and LZMA level 3 compression reduces the data volume by more than 30%. When the bandwidth is greater than or equal to the first bandwidth, choosing full coding and LZMA level 8 compression balances transmission speed and storage usage.

[0144] Based on this scheme, during the compression of geospatial tile data, the problem of slow queries can be solved by using multi-level index caching to trade space for time. On-demand decompression reduces data processing, and network scenario adaptation allows the system to automatically select data versions with different compression levels for transmission and loading based on user requests and current network conditions. This dynamic selection mechanism enables the data service to adapt to different network environments. When bandwidth is limited, highly compressed data is prioritized to ensure response speed; when bandwidth is sufficient, lower compression ratio but faster decompression versions can be selected to improve the user experience.

[0145] Corresponding to the embodiments of the foregoing methods, this disclosure also provides embodiments of the apparatus and the terminal to which it is applied.

[0146] Embodiments of the document processing apparatus disclosed herein can be applied to computer devices, such as servers or terminal devices. The apparatus embodiments can be implemented through software, hardware, or a combination of both. Taking software implementation as an example, as a logically defined apparatus, it is formed by the processor in which it processes the file, reading the corresponding computer program instructions from non-volatile memory into memory for execution. From a hardware perspective, such as... Figure 5 The diagram shown is a hardware structure diagram of a computer device housing a geospatial tile compression device according to an embodiment of this disclosure, except... Figure 5 In addition to the processor 510, memory 530, network interface 520, and non-volatile memory 540 shown, the server or electronic device where the geospatial tile compression method is located in the embodiment may also include other hardware depending on the actual function of the computer device, which will not be described in detail here.

[0147] Figure 6This is a schematic diagram of a geospatial tile compression device provided in an embodiment of the present disclosure. The geospatial tile compression device 600 includes: a preprocessing module 601, a compression module 602, a dictionary construction module 603, a residual coding module 604, and a storage module 605. The preprocessing module 601 is used to perform unified preprocessing and type-adaptive differentiated preprocessing on multiple types of geospatial tiles. The compression module 602 is used to compress the differentiated preprocessed tile data according to the type using corresponding strategies and compression depths, and to mark the compression feature information corresponding to each type. The dictionary construction module 603 is used to extract common data from the compressed tile data based on the compression feature information. The system constructs a global shared dictionary and multiple group shared dictionaries. The global shared dictionary stores the QuadKey prefix, which serves as the basis for residual coding. The group shared dictionaries store shared data fragments extracted from common data for compressing residual coding. The residual coding module 604 converts the QuadKey encoded sequence of the compressed tile data into a residual sequence based on the QuadKey prefix in the global shared dictionary, and compresses the residual sequence using the group shared dictionaries to generate residual coded data. The storage module 605 writes the residual coded data and the compressed tile data into a storage structure containing an index area and a data area. The index area is used to locate the tile data.

[0148] Optionally, the preprocessing module 601 is specifically used for: unifying the coordinate systems of tiles from different sources to the same map projection coordinate system, and dividing them into layers according to the original precision of each type of tile; unifying the pixel size of raster tiles to a preset pixel specification, and matching vector tiles to the geographic spatial range of raster tiles at the same layer; generating unified metadata for each tile, including: tile type identifier, QuadKey prefix, and preprocessing precision field; performing color space format conversion on image raster tiles after unified preprocessing; performing numerical index conversion on numerical raster tiles; and performing attribute filtering on vector tiles.

[0149] Optionally, the compression module 602 is specifically used for: for image-type raster tiles, performing initial compression based on encoder encoding adapted to the image scene, performing secondary compression based on a first compression depth using a general lossless compression method, and marking color redundancy features after secondary compression; for numerical raster tiles, performing initial compression based on encoder encoding adapted to the numerical scene, performing secondary compression based on a second compression depth using a general lossless compression method, and marking numerical interval features after secondary compression; for vector tiles, performing structured processing based on encoder encoding adapted to the vector scene, and compressing the structured data using a second compression depth using a general lossless compression method, and marking geometric instruction features after compression; wherein, the first compression depth is higher than the second compression depth.

[0150] Optionally, the dictionary construction module 603 is specifically used for: extracting globally shared repeating pixel templates based on color redundancy features, extracting group-shared same-interval index value templates based on numerical interval features, and extracting group-shared high-frequency instruction fragments based on geometric instruction features; constructing a globally shared dictionary to store common data across types and levels, including format identifiers and QuadKey prefix information; constructing a group-shared dictionary based on tile type, tile level, and compression feature information to store common data within a group based on the same compression feature information; and performing target processing on the globally shared dictionary and the group-shared dictionary; wherein, the target processing includes: removing duplicate entries in the dictionary through deduplication, compressing the dictionary itself, and reserving expansion space for the dictionary.

[0151] Optionally, the residual encoding module 604 is specifically used to: encode the QuadKey sequence of the tile into a numerical sequence, calculate the numerical difference between adjacent tiles as the residual; perform encoding preprocessing on the residual based on the shared dictionary system; the encoding preprocessing includes: replacing the repeated residual patterns with the index reference of the corresponding group shared dictionary, using the QuadKey prefix in the global shared dictionary as the reference value for residual calculation; and performing residual encoding based on the tile type identification information and the encoded preprocessed residual to generate residual encoded data.

[0152] Optionally, the storage structure includes a header area, an index area, and a data area; the header area stores version information, dictionary identifier, total number of tiles, and compression type marker; the index area stores the mapping relationship between residual coding results and data location information according to hierarchical fragmentation; the data area uses a block storage mechanism to store compressed tile data in QuadKey order, and each data block is added with data verification information during storage.

[0153] Optionally, such as Figure 7 As shown, the geospatial tile compression device 600 further includes: a scheduling adaptive module 607; the scheduling adaptive module 607 is used to locate the tile based on the memory caching algorithm and the hierarchical tile information of the index area when a scheduling request is received; to perform on-demand decompression based on the tile type of the scheduling request, the group dictionary and decompression algorithm based on the matching dictionary identifier and compression mark; and to select the encoding strategy and compression level according to the network conditions.

[0154] This disclosure provides a geospatial tile compression device. First, it performs unified preprocessing on tiles of various types. Then, it performs differentiated preprocessing on the preprocessed tiles to ensure that tiles of different sources, types, and specifications can be standardized before encoding, eliminating differences caused by different sources. Second, it can perform customized deep compression on tile data by combining compression methods suitable for each type and mark the compression feature information of each type of data. Then, based on the marked compression feature information and common data extracted from the compressed tile data, a global and grouped two-level shared dictionary system is constructed. Based on the QuadKey prefix in the global shared dictionary, the QuadKey encoding sequence of the compressed tile data is converted into a residual sequence, and the residual sequence is compressed using the grouped shared dictionary to generate residual encoded data, that is, the index of the tile data is also compressed. Finally, the two-level shared dictionary and the index residual encoding are merged, and the residual encoded data and the compressed tile data are written into a storage structure containing an index area and a data area. This reduces storage redundancy and cost, improves the compression compatibility of various types of tiles, and has strong scalability.

[0155] Accordingly, this disclosure also provides a computer device, which includes a processor; a memory for storing processor-executable instructions; wherein the processor is configured to perform the steps in the above-described embodiments of the geospatial tile compression method.

[0156] Accordingly, this disclosure also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of each of the above-described geospatial tile compression method embodiments.

[0157] The specific implementation process of the functions and roles of each module in the above device can be found in the implementation process of the corresponding steps in the above method, and will not be repeated here.

[0158] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this disclosure according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0159] The foregoing has described specific embodiments of this disclosure. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired results. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0160] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention applied herein. This disclosure is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not claimed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the following claims.

[0161] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.

[0162] The above description is merely a preferred embodiment of this disclosure and is not intended to limit this disclosure. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for compressing geographic spatial tiles, characterized in that, The method includes: Perform unified preprocessing and type-adaptive differentiated preprocessing on various types of geospatial tiles; For the preprocessed tile data with differential characteristics, compression is performed according to the type using the corresponding compression strategy and compression depth, and the compression feature information corresponding to each type is marked. Based on the compression feature information, common data is extracted from the compressed tile data to construct a global shared dictionary and multiple group shared dictionaries; the global shared dictionary stores the QuadKey prefix, which serves as the basis for residual coding; the group shared dictionaries store shared data fragments extracted from the common data for compressing residual coding. Based on the QuadKey prefix in the global shared dictionary, the QuadKey encoded sequence of the compressed tile data is converted into a residual sequence, and the residual sequence is compressed using the grouped shared dictionary to generate residual encoded data. The residual encoded data and the compressed tile data are written into a storage structure containing an index area and a data area; the index area is used to locate the tile data.

2. The method according to claim 1, characterized in that, The unified preprocessing and type-adaptive differentiated preprocessing for multiple types of geospatial tiles include: Unify the coordinate systems of tiles from different sources to the same map projection coordinate system, and divide them into layers according to the original precision of each type of tile; unify the pixel size of raster tiles to a preset pixel specification, and match vector tiles to the geographic spatial range of raster tiles at the same layer; generate unified metadata for each tile, which includes: tile type identifier, QuadKey prefix and preprocessing precision field. After unified preprocessing, color space format conversion is performed on image-type raster tiles; numerical index conversion is performed on numerical raster tiles; and attribute filtering is performed on vector tiles.

3. The method according to claim 2, characterized in that, The preprocessed tile data is compressed according to its type using a corresponding compression strategy and compression depth, and the compression feature information corresponding to each type is marked, including: For image-type raster tiles, the first compression is performed based on encoder encoding adapted to the image scene, and the second compression is performed based on the first compression depth of a general lossless compression method. After the second compression, color redundancy features are marked. For numerical raster tiles, the first compression is performed based on an encoder adapted to numerical scenarios, and the second compression is performed based on a second compression depth using a general lossless compression method. After the second compression, numerical interval features are marked. For vector tiles, structured processing based on an encoder adapted to the vector scene is performed, and the structured data is compressed using the second compression depth based on the general lossless compression method, and geometric instruction features are marked after compression. Wherein, the first compression depth is higher than the second compression depth.

4. The method according to claim 3, characterized in that, Based on the compression feature information, common data is extracted from the compressed tile data to construct a globally shared dictionary and multiple grouped shared dictionaries, including: Based on the color redundancy feature, a globally shared repeating pixel template is extracted; based on the numerical interval feature, a group-shared same-interval index value template is extracted; and based on the geometric instruction feature, a group-shared high-frequency instruction fragment is extracted. Construct a globally shared dictionary to store common data across types and levels. The common data includes: format identifiers and QuadKey prefix information. Based on tile type, tile level, and compressed feature information, a grouped shared dictionary is constructed to store common data within a group based on the same compressed feature information. Perform target processing on the globally shared dictionary and the grouped shared dictionary; The target processing includes: removing duplicate entries from the dictionary through deduplication, compressing the dictionary itself, and reserving space for dictionary expansion.

5. The method according to claim 4, characterized in that, The process of converting the QuadKey encoded sequence of the compressed tile data into a residual sequence based on the QuadKey prefix in the globally shared dictionary, and then compressing the residual sequence using the grouped shared dictionary to generate residual encoded data includes: The QuadKey sequence of the tile is encoded into a numerical sequence, and the numerical difference between adjacent tiles is calculated as the residual. The residuals are pre-processed by encoding based on the shared dictionary system; the encoding pre-processing includes: replacing the repeated residual patterns with the index references of the corresponding group shared dictionary, and using the QuadKey prefix in the global shared dictionary as the reference value for residual calculation; Residual encoding is performed based on the tile type identification information and the residual after the encoding preprocessing to generate residual encoded data.

6. The method according to any one of claims 1 to 5, characterized in that, The storage structure includes a header area, an index area, and a data area; The header region stores version information, dictionary identifier, total number of tiles, and compression type marker; The index region stores the mapping relationship between residual coding results and data location information in hierarchical fragmentation. The data area uses a block storage mechanism to store compressed tile data in QuadKey order, and each data block is equipped with data verification information during storage.

7. The method according to claim 6, characterized in that, The method further includes: Upon receiving a scheduling request, the tile is located based on the memory caching algorithm and the hierarchical fragmentation information of the index area; Based on the tile type of the scheduling request, decompression is performed on demand using a group dictionary that matches dictionary identifiers and compression tags, and a decompression algorithm. Choose the encoding strategy and compression level based on network conditions.

8. A compression device for geographic spatial tiles, characterized in that, The compression device includes: a preprocessing module, a compression module, a dictionary construction module, a residual encoding module, and a storage module; The preprocessing module is used to perform unified preprocessing and type-adaptive differentiated preprocessing on multiple types of geospatial tiles; The compression module is used to compress the preprocessed tile data according to the type using the corresponding strategy and compression depth, and to mark the compression feature information corresponding to each type. The dictionary construction module is used to extract common data from the compressed tile data based on the compression feature information, and construct a global shared dictionary and multiple group shared dictionaries; the global shared dictionary stores the quadtree key value QuadKey prefix as the basis for residual coding; the group shared dictionaries store shared data fragments extracted from the common data for compressing residual coding; The residual coding module is used to convert the QuadKey encoded sequence of the compressed tile data into a residual sequence based on the QuadKey prefix in the global shared dictionary, and to compress the residual sequence using the group shared dictionary to generate residual encoded data. The storage module is used to write the residual encoded data and the compressed tile data into a storage structure containing an index area and a data area; the index area is used to locate the tile data.

9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the geospatial tile compression method according to any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the geospatial tile compression method according to any one of claims 1-7.

Images