A vector map compression system and method based on delta encoding
By using an incremental coding-based vector map compression system, differentiated coding strategies are adopted for different data types, which solves the problems of low compression ratio, poor access performance and difficulty in incremental updates in existing technologies. This achieves efficient data compression and improved access performance, while reducing storage costs.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- FEIDU AEROSPACE TECH CO LTD
- Filing Date
- 2026-03-06
- Publication Date
- 2026-06-12
AI Technical Summary
Existing technologies for vector map data compression suffer from problems such as low compression ratio, low coding efficiency, poor access performance, data accuracy risks, and difficulty in incremental updates. They cannot fully utilize the characteristics of vector map data and lack differentiated coding strategies for different data types.
An incremental coding-based vector map compression system is adopted. The data preprocessing module analyzes the characteristics of vector map data, and the coordinate incremental coding module, the identifier incremental coding module, the weight incremental coding module, and the intersection ID coding module are used to perform incremental coding on map coordinates, feature IDs, road network weights, and intersection IDs respectively. Combined with Varint, Gamma, Delta+ZigZag and bitstream coding strategies, adaptive optimal coding is achieved.
It improves compression ratio by 50-70%, enhances access performance by 3-5 times, supports streaming and random access, maintains data accuracy, improves incremental update efficiency by 5-10 times, and reduces storage costs by 50-70%.
Smart Images
Figure CN122199694A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of vector map technology, and in particular to a vector map compression system and method based on incremental coding. Background Technology
[0002] Vector map data is core data in geographic information systems, containing coordinate information, feature identifiers, topological relationships, attribute data, etc. With the continuous expansion of map data scale (road network data for a single city can reach GB levels, and national road network data can reach TB levels), data compression has become a key technological challenge.
[0003] In the field of data compression, there are several general-purpose compression algorithms: LZ series algorithms (LZ77, LZ78, LZW, etc.): dictionary-based compression algorithms suitable for text and general data. Huffman coding: frequency-based variable-length coding suitable for data with known probability distributions. Arithmetic coding: coding based on probability models, with high compression ratios but complex computation. Zlib / Gzip: general-purpose compression based on the DEFLATE algorithm, widely used in various scenarios. In the field of vector data compression, existing technologies mainly include: direct application of general-purpose compression algorithms: serializing vector data and then compressing it using general-purpose algorithms such as Zlib; simple differential coding: performing simple differential coding on the coordinate sequence and then using general-purpose compression; geometric simplification: reducing the number of coordinate points to reduce the data volume, but at the cost of precision.
[0004] The existing technology has the following technical disadvantages: Low compression ratio. General compression algorithms cannot fully utilize the characteristics of vector map data, and the compression ratio is usually only 30-50%; simple differential coding uses fixed-length encoding, which cannot fully utilize the advantages of small differences; and there is a lack of differentiated encoding strategies for different data types.
[0005] The encoding efficiency is low. It requires secondary compression (differential encoding + general compression), which increases computational overhead; the encoding / decoding process is complex and has high time complexity; it cannot support streaming processing and must be completely decompressed before it can be accessed.
[0006] Poor access performance. Compressed data cannot be accessed randomly and must be completely decompressed; streaming decoding is not supported, resulting in high data access latency; and it cannot meet the needs of real-time navigation and dynamic loading.
[0007] Data accuracy risks. Some compression algorithms may quantize the data, leading to a loss of accuracy; geometric simplification methods can directly result in a loss of coordinate accuracy.
[0008] Incremental updates are difficult. Each update requires recompressing the entire dataset; update efficiency is low, network transmission volume is large; and partial data updates cannot be supported.
[0009] The encoding strategy is limited. It cannot adopt the optimal encoding strategy for different types of data such as coordinate data, ID data, and weighted data; nor can it adaptively select the encoding method based on data characteristics (monotonicity, correlation, repetition). Summary of the Invention
[0010] The present invention aims to propose a vector map compression system and method based on incremental coding to solve the above-mentioned problems existing in the prior art.
[0011] This invention proposes a vector map compression system based on incremental coding, comprising: a vector map data source for providing vector map data; a data preprocessing module for analyzing the characteristics of the vector map data and identifying data type and distribution characteristics; a coordinate incremental coding module for incrementally coding the map coordinate sequence, using Varint to encode coordinate differences and utilizing the spatial locality of coordinates; an identifier incremental coding module for incrementally coding the feature ID and point ID sequences, using Gamma to encode ID differences and utilizing the monotonicity of ID sequences; a weight incremental coding module for incrementally coding the road network weight sequence, using Delta coding combined with ZigZag coding to encode weight differences and process signed differences; an intersection ID coding module for bitstream coding of intersection IDs, distinguishing between new and repeated intersections and utilizing the repetitiveness of intersections; a coding combination strategy module for automatically selecting the optimal coding method based on data type and data characteristics, and coordinating the work of each coding module; and a data output module for outputting the compressed data and supporting streaming processing and random access.
[0012] Furthermore, the data preprocessing module includes: a feature analyzer for analyzing the features of vector map data; a data type identifier for identifying the feature types of vector map data; and a distribution feature calculator for calculating the distribution features of vector map data.
[0013] Furthermore, the coordinate incremental encoding module includes: a coordinate difference calculator, used to calculate the difference between the current coordinate and the previous coordinate; a Varint encoder, used to encode the difference using Varint; and a reference point manager, used to set the first coordinate as a reference point and update the reference point coordinates according to the encoding of the difference.
[0014] Furthermore, the identifier incremental encoding module includes: an ID difference calculator, used to initialize an initial ID and calculate the difference between each ID and the initial ID; a monotonicity detector, used to determine whether the feature ID and point ID sequence has monotonicity characteristics; and a Gamma encoder, used to encode the difference using Gamma and update the latest ID value to the initial ID according to the encoding.
[0015] Furthermore, the weight increment encoding module includes: a weight quantizer, used to quantize the weights to a specified granularity based on the road network weights and calculate the corresponding difference; a ZigZag encoder, used to calculate the difference and map the signed number to an unsigned number; and a Delta encoder, used to compress the encoding result of the ZigZag encoder.
[0016] Furthermore, the intersection ID encoding module includes: a duplicate detector for detecting whether the intersection ID is duplicated; an intersection index manager for guiding independent intersections and duplicate intersections to use different encoding strategies for encoding; and a bitstream encoder for encoding independent intersections and duplicate intersections using different encoding strategies according to the guidance of the intersection index manager.
[0017] Furthermore, the encoding combination strategy module includes: a strategy selector for selecting different encoding combination strategies according to different data types; an encoding coordinator for coordinating the encoding of each encoder according to the encoding combination strategy selected by the strategy selector; and a performance optimizer for optimizing the encoding order and performance among the encoders.
[0018] This invention also provides a vector map compression method based on incremental coding, comprising the following steps: S1, acquiring and analyzing the characteristics of vector map data, identifying data type and distribution characteristics; S2, performing incremental coding on the map coordinate sequence, using Varint to encode coordinate differences, utilizing the spatial locality characteristics of coordinates; S3, performing incremental coding on the feature ID and point ID sequences, using Gamma to encode ID differences, utilizing the monotonicity characteristics of the ID sequence; S4, performing incremental coding on the road network weight sequence, using Delta coding combined with ZigZag coding to encode weight differences, processing signed differences; S5, performing bitstream coding on intersection IDs, distinguishing between new intersections and repeated intersections, utilizing the repetitive characteristics of intersections; S6, automatically selecting the optimal coding method according to data type and data characteristics, coordinating the work of each coding module; S7, outputting the compressed data, supporting streaming processing and random access.
[0019] The vector map compression system and method based on incremental coding of the present invention have the following advantages: Improve compression ratio. Achieve a 50-70% compression ratio improvement through incremental encoding strategies for different data types, compared to traditional methods. Optimize access performance. Supports streaming decoding and random access, with decoding speed improved by 3-5 times, and encoding / decoding complexity of O(n). Maintain data precision. Employs lossless compression, preserving data precision during compression without loss of accuracy. Support incremental updates. Supports incremental update mechanisms, eliminating the need to recompress the entire dataset, improving update efficiency by 5-10 times. Adaptive encoding strategy. Automatically selects the optimal encoding method based on data type and characteristics, achieving the best balance between compression efficiency and access performance. Reduce storage costs. Reduces storage space by 50-70% compared to the original data, and by 30-50% compared to traditional compression methods. Attached Figure Description
[0020] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 The diagram shows the system framework of the vector map compression system based on incremental coding in this invention.
[0021] Figure 2 The diagram shows the flowchart of coordinate incremental encoding in this invention.
[0022] Figure 3 The diagram shows a flowchart of ID incremental encoding in the identifier incremental encoding of this invention.
[0023] Figure 4 The diagram shows the flowchart of the weight increment encoding process in this invention.
[0024] Figure 5 The diagram shows the flowchart framework of the encoding combination strategy in this invention. Detailed Implementation
[0025] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.
[0026] like Figures 1 to 5As shown, this invention provides a vector map compression system based on incremental coding, comprising: a vector map data source for providing vector map data; a data preprocessing module for analyzing the characteristics of the vector map data and identifying data type and distribution characteristics; a coordinate incremental coding module for incrementally coding the map coordinate sequence, using Varint to encode coordinate differences and utilizing the spatial locality characteristics of coordinates; an identifier incremental coding module for incrementally coding the feature ID and point ID sequences, using Gamma to encode ID differences and utilizing the monotonicity characteristics of ID sequences; a weight incremental coding module for incrementally coding the road network weight sequence, using Delta coding combined with ZigZag coding to encode weight differences and process signed differences; an intersection ID coding module for bitstream coding of intersection IDs, distinguishing between new and repeated intersections and utilizing the repetitive characteristics of intersections; a coding combination strategy module for automatically selecting the optimal coding method based on data type and data characteristics, and coordinating the work of each coding module; and a data output module for outputting compressed data and supporting streaming processing and random access. The compression method used in this system includes the following steps: S1. Acquire and analyze the characteristics of vector map data, and identify data type and distribution characteristics; S2. Perform incremental encoding on the map coordinate sequence, using Varint encoding for coordinate differences, and utilizing the spatial locality characteristics of coordinates; S3. Perform incremental encoding on the feature ID and point ID sequences, using Gamma encoding for ID differences, and utilizing the monotonicity characteristics of the ID sequences; S4. Perform incremental encoding on the road network weight sequence, using Delta encoding combined with ZigZag encoding for weight differences, and processing signed differences; S5. Perform bitstream encoding on intersection IDs, distinguishing between new and repeated intersections, and utilizing the repetitive characteristics of intersections; S6. Automatically select the optimal encoding method based on data type and data characteristics, and coordinate the work of each encoding module; S7. Output the compressed data, and support streaming processing and random access.
[0027] In a preferred embodiment of the present invention, the data preprocessing module includes: a feature analyzer for analyzing the features of vector map data; a data type identifier for identifying the feature types of the vector map data; and a distribution feature calculator for calculating the distribution features of the vector map data. After processing by each unit in the data preprocessing module, data suitable for subsequent processing is generated.
[0028] In a preferred embodiment of the present invention, the coordinate incremental encoding module includes: a coordinate difference calculator for calculating the difference between the current coordinate and the previous coordinate; a Varint encoder for encoding the difference using Varint; and a reference point manager for setting the first coordinate as a reference point and updating the reference point coordinates according to the encoding of the difference. The processing flow of coordinate incremental encoding is as follows: Figure 2As shown, the technical principle is that road network coordinates have spatial locality characteristics, and the difference between adjacent coordinates is usually small. Varint encoding (variable-length integer encoding) can fully utilize the advantage of small differences. Finally, the highest bit of each byte indicates whether there are subsequent bytes; small values use one byte, and large values use multiple bytes. The specific steps are: first, read the coordinate sequence; then, set the first coordinate as the reference point and write it to the output; finally, perform the following calculations for each subsequent coordinate: Calculate the difference between the previous coordinate and the previous coordinate: ; Use Varint encoding for dx and dy; Update prev = curr.
[0029] It can achieve a coordinate data compression rate of 50-70%, saving 30-50% of space compared to fixed-length encoding.
[0030] The specific coding implementation is as follows:
[0031] The specific decoding implementation is as follows:
[0032] In a preferred embodiment of the present invention, the identifier incremental encoding module includes: an ID difference calculator, used to initialize an initial ID and calculate the difference between each ID and the initial ID; a monotonicity detector, used to determine whether the feature ID and point ID sequences have monotonicity characteristics; and a Gamma encoder, used to encode the differences using Gamma and update the latest ID value to the initial ID according to the encoding. The processing flow of ID incremental encoding in identifier incremental encoding is as follows: Figure 3 As shown, the technical principle is to first determine that the sequence of element IDs and point IDs usually has monotonicity (increasing or decreasing), then determine that the difference between adjacent IDs is usually small and positive, and then use Gamma encoding (Elias Gamma encoding) to efficiently compress small integers. The Gamma encoding principle is as follows: for a value v, first encode its binary digits n (using n zeros + 1 one), then encode the lower n bits of v. Small values use fewer bits, and large values use more bits. For example, value 1 uses 1 bit, values 2-3 use 3 bits, and values 4-7 use 5 bits.
[0033] The specific processing steps are as follows: Read the ID sequence; initialization (Utilizing the modular arithmetic characteristics of uint32_t); For each ID: Calculate the difference: Using Gamma-encoded delta, update .
[0034] It can achieve an ID data compression rate of 60-80%, saving 50-70% of space compared to fixed-length encoding.
[0035] The specific coding implementation is as follows:
[0036] In a preferred embodiment of the present invention, the weight increment encoding module includes: a weight quantizer, used to quantize the weights to a specified granularity based on the road network weights and calculate the corresponding difference; a ZigZag encoder, used to calculate the difference and map the signed number to an unsigned number; and a Delta encoder, used to compress the encoding result of the ZigZag encoder. The processing flow of weight increment encoding is as follows: Figure 4 As shown, the technical principle is that the road network weight data is correlated, and the difference between adjacent weights is usually small, and the difference may be negative. Signed numbers need to be processed. ZigZag encoding is used to convert the signed difference into an unsigned number, and then Delta encoding is used. The ZigZag encoding principle is: mapping signed integers to unsigned integers, positive numbers to even numbers: 0→0, 1→2, 2→4, ...; negative numbers to odd numbers: -1→1, -2→3, -3→5, ...; formula: Delta encoding principle: First, use Gamma encoding to encode the binary number of bits n+1 of the value, then encode the lower n bits of the value. It is suitable for compressing small integers.
[0037] The specific processing steps are as follows: Read the weight sequence; Initialize prev_weight = 1; For each weight: if there is no path marker, write the marker bit 0; otherwise, write the marker bit 1. Quantize the weights to a specified granularity (round up), and calculate the difference: ZigZag encoding difference (mapping signed numbers to unsigned numbers), Delta encoding; renew .
[0038] It can achieve a weighted data compression rate of 40-60%, saving 30-50% of space compared to fixed-length encoding.
[0039] The specific coding implementation is as follows:
[0040] In a preferred embodiment of the present invention, the intersection ID encoding module includes: a duplicate detector for detecting whether the intersection ID is duplicated; an intersection index manager for guiding independent intersections and duplicate intersections to use different encoding strategies; and a bitstream encoder for encoding independent intersections and duplicate intersections using different encoding strategies according to the guidance of the intersection index manager. The technical principle is as follows: because intersection IDs have a repeatability characteristic, the same intersection may be shared by multiple roads, so first, it is determined whether the intersection ID is duplicated; then, new intersections and duplicate intersections are guided to use different encoding strategies; wherein, for new intersections: a flag bit 1 is written, followed by the feature ID and point ID; for duplicate intersections: a flag bit 0 is written, followed by the existing intersection index. This achieves an intersection ID data compression rate of 50-70%, with better compression effect for higher duplication rates.
[0041] The specific coding implementation is as follows:
[0042] In a preferred embodiment of the present invention, the encoding combination strategy module includes: a strategy selector for selecting different encoding combination strategies according to different data types; an encoding coordinator for coordinating the encoding of each encoder according to the encoding combination strategy selected by the strategy selector; and a performance optimizer for optimizing the encoding order and performance among the encoders. The strategy selection rules are as follows: Coordinate data: Always use Varint incremental encoding; ID data: If monotonicity is satisfied and the average difference is <100, use Gamma incremental encoding; otherwise, use general compression; Weight data: If correlation is satisfied and the average absolute difference is <50, use Delta+ZigZag incremental encoding; otherwise, use general compression; Intersection ID: If the repetition rate is >30%, use bitstream encoding + repetition detection; otherwise, use general compression.
[0043] Its encoding implementation is as follows:
[0044] In addition, such as Figure 1 As shown, it also includes modules for data streaming processing and random access. The data streaming processing method supports streaming output during the encoding process, eliminating the need to wait for all data to be encoded, and streaming input during the decoding process, allowing for simultaneous reading and decoding, suitable for network transmission and real-time processing scenarios. The random access processing method uses an index table to record the offset and size of each data segment, enabling quick location of specific data segments without complete decompression, supporting on-demand loading, and reducing memory usage.
[0045] Its encoding implementation is as follows:
[0046] The beneficial effects of the incremental coding-based vector map compression system of the present invention are as follows: Multi-level incremental coding system: Different incremental coding strategies are adopted for different types of data such as coordinates, ID, weights, and intersections. It is not a simple single coding method, but rather the optimal coding is selected based on the data characteristics.
[0047] Application of Varint encoding in coordinate compression: Taking advantage of the spatial locality of coordinates, where the difference between adjacent coordinates is small, Varint encoding makes full use of the advantage of small values, saving 30-50% of space compared to fixed-length encoding.
[0048] Application of Gamma coding in ID sequence compression: By utilizing the monotonicity of ID sequences, Gamma coding efficiently compresses small integers and is suitable for ID difference coding.
[0049] Application of Delta+ZigZag combined encoding in weighted compression: ZigZag encoding processes signed differences, and Delta encoding further compresses them, achieving a dual compression effect.
[0050] Application of bitstream coding + duplicate detection in intersection ID compression: By utilizing the repeatability characteristics of intersections, new intersections and repeated intersections are encoded differently, and the higher the repeatability rate, the better the compression effect.
[0051] Adaptive encoding strategy selection: Automatically selects the encoding method based on data characteristics (monotonicity, correlation, repetition rate) to achieve the optimal balance between compression efficiency and access performance.
[0052] This invention also provides a vector map compression method based on incremental coding, comprising the following steps: S1, acquiring and analyzing the characteristics of vector map data, identifying data type and distribution characteristics; S2, performing incremental coding on the map coordinate sequence, using Varint to encode coordinate differences, utilizing the spatial locality characteristics of coordinates; S3, performing incremental coding on the feature ID and point ID sequences, using Gamma to encode ID differences, utilizing the monotonicity characteristics of the ID sequence; S4, performing incremental coding on the road network weight sequence, using Delta coding combined with ZigZag coding to encode weight differences, processing signed differences; S5, performing bitstream coding on intersection IDs, distinguishing between new and repeated intersections, utilizing the repetitive characteristics of intersections; S6, automatically selecting the optimal coding method according to data type and data characteristics, coordinating the work of each coding module; S7, outputting the compressed data, supporting streaming processing and random access. The system used is the aforementioned vector map compression system based on incremental coding, whose functional modules and working methods are the same and correspond one-to-one with the above system, and will not be described again here.
[0053] Therefore, the vector map compression system and method based on incremental encoding provided by this invention improves the compression ratio. Through incremental encoding strategies targeting different data types, a 50-70% improvement in compression ratio is achieved, compared to traditional methods. Optimized access performance is also enhanced. It supports streaming decoding and random access, with decoding speed improved by 3-5 times and encoding / decoding complexity of O(n). Data accuracy is maintained. Lossless compression is employed, preserving data accuracy during compression without loss of precision. Incremental updates are supported. An incremental update mechanism is supported, eliminating the need to recompress the entire dataset, improving update efficiency by 5-10 times. An adaptive encoding strategy automatically selects the optimal encoding method based on data type and characteristics, achieving an optimal balance between compression efficiency and access performance. Finally, storage costs are reduced. Storage space is reduced by 50-70% compared to the original data, and by 30-50% compared to traditional compression methods. This paper proposes a multi-level incremental coding compression method for vector map data. By analyzing the characteristics of different data types (spatial locality, monotonicity, correlation, and repetition), it employs differentiated incremental coding strategies (Varint, Gamma, Delta+ZigZag, and bitstream coding) to achieve unified optimization of high compression ratio, fast access, and accuracy preservation. This is the first time a multi-level incremental coding system has been applied to vector map data compression. Instead of a single universal compression, it uses optimal coding strategies for different data types, supports streaming processing and random access, overcomes the access limitations of traditional compression, achieves lossless compression, and maintains data accuracy.
[0054] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.
Claims
1. A vector map compression system based on incremental coding, characterized in that, include: Vector map data source, used to provide vector map data; The data preprocessing module is used to analyze the characteristics of vector map data and identify data types and distribution features; The coordinate incremental encoding module is used to incrementally encode the map coordinate sequence, using Varint to encode the coordinate difference and taking advantage of the spatial locality of coordinates. The identifier incremental encoding module is used to incrementally encode the feature ID and point ID sequences, using Gamma encoding of ID differences and taking advantage of the monotonicity of the ID sequences. The weight increment encoding module is used to incrementally encode the road network weight sequence. It uses Delta encoding combined with ZigZag encoding to process signed differences. The intersection ID encoding module is used to perform bitstream encoding on the intersection ID, distinguishing between new intersections and duplicate intersections, and utilizing the repetitive characteristics of intersections; The encoding combination strategy module is used to automatically select the optimal encoding method based on data type and data characteristics, and coordinate the work of each encoding module. The data output module is used to output compressed data and supports streaming processing and random access.
2. The vector map compression system based on incremental coding according to claim 1, characterized in that, The data preprocessing module includes: Feature analyzer, used to analyze the features of vector map data; Data type identifier, used to identify the feature type of vector map data; Distribution feature calculator, used to calculate the distribution features of vector map data.
3. The vector map compression system based on incremental coding according to claim 1, characterized in that, The coordinate increment encoding module includes: Coordinate difference calculator, used to calculate the difference between the current coordinate and the previous coordinate; A Varint encoder is used to encode the difference using Varint; A reference point manager is used to set the first coordinate as a reference point and update the reference point coordinates according to the encoding of the difference.
4. The vector map compression system based on incremental coding according to claim 1, characterized in that, The identifier incremental encoding module includes: The ID difference calculator is used to initialize initial IDs and calculate the difference between each ID and the initial ID. A monotonicity detector is used to determine whether the sequence of feature IDs and point IDs has a monotonicity characteristic; A Gamma encoder is used to encode the difference using Gamma and update the latest ID value to the initial ID based on the encoding.
5. The vector map compression system based on incremental coding according to claim 1, characterized in that, The weight increment encoding module includes: The weight quantizer is used to quantize the weights to a specified granularity based on the number of road network weights and calculate the corresponding difference. The ZigZag encoder is used to calculate the difference and map signed numbers to unsigned numbers; A Delta encoder is used to compress the encoding results of the ZigZag encoder.
6. The vector map compression system based on incremental coding according to claim 1, characterized in that, The intersection ID encoding module includes: A duplicate detector is used to detect whether the intersection ID is repeated. The intersection index manager is used to guide the encoding of independent and repeated intersections using different encoding strategies. A bitstream encoder is used to encode independent intersections and repeated intersections using different encoding strategies, guided by the intersection index manager.
7. The vector map compression system based on incremental coding according to claim 1, characterized in that, The encoding combination strategy module includes: A strategy selector is used to select different encoding combination strategies based on different data types; An encoding coordinator is used to coordinate the encoding of each encoder according to the encoding combination strategy selected by the strategy selector; The performance optimizer optimizes the encoding order and performance between the various encoders.
8. A vector map compression method based on incremental coding, characterized in that, Including the following steps: S1. Acquire and analyze the characteristics of vector map data, and identify data types and distribution characteristics; S2. Incrementally encode the map coordinate sequence, using Varint to encode the coordinate difference, and utilize the spatial locality of coordinates. S3. Incrementally encode the feature ID and point ID sequences, use Gamma encoding for ID differences, and take advantage of the monotonicity of the ID sequences. S4. Incrementally encode the road network weight sequence, using Delta encoding combined with ZigZag encoding to process signed differences; S5. Perform bitstream encoding on the intersection ID to distinguish between new intersections and repeated intersections, utilizing the repetitive characteristics of intersections; S6. Automatically select the optimal encoding method based on data type and data characteristics, and coordinate the work of each encoding module; S7 outputs compressed data and supports streaming and random access.