Real-scene three-dimensional system geographic entity space-time coding method, system, equipment and medium

Abstract

The invention discloses a real-scene three-dimensional system geographic entity space-time coding method, system and device and a medium, belongs to the technical field of three-dimensional scenes, and aims to solve the technical problem of how to quickly and accurately carry out space-time positioning and coding on a geographic entity in a real-scene three-dimensional system and improve data processing efficiency and precision. According to the technical scheme, multi-level network division and unique identifier: multi-level and multi-scale grids on the surface of the earth are divided based on a GeoSOT model, the GB / T 40087-2021 earth space grid coding rule is followed, the grid level is determined by calculating the size of a bounding box of a geographic entity, coordinate binarization left-shift operation is carried out at a coordinate of a center point of the entity, and a unique identifier is obtained; generating a unique grid code containing south and north hemisphere marks and east and west hemisphere marks by combining a Morton code conversion technology and a quaternary conversion technology; bTC Beidou subdivision of time codes is carried out; a space-time embedding model; optimizing spatial data coding in combination with Morton codes and binary tree coding; and sorting the space-time codes through a heap sorting algorithm.

Description

Technical Field

[0001] The present invention relates to the field of three-dimensional scene technology, and in particular to a method, system, device and medium for spatiotemporal coding of geographic entities in a real-scene three-dimensional system. Background Art

[0002] 3D Real Scene (3DRealScene) is a digital virtual space that reflects and expresses human production, life and ecological space in a real, three-dimensional and time-sequential manner. It is a new type of basic surveying and mapping standardized product that provides a unified spatial foundation for economic development and informatization. 3D Real Scene is constructed by carrying structured, semantic geographic entities that support human-machine compatible understanding and real-time perception of the Internet of Things on a three-dimensional geographic scene.

[0003] With the advancement of smart city construction and real-life 3D policies, higher requirements are placed on the accurate expression and management of geographic information. Currently, commonly used geographic information coding methods such as latitude and longitude coordinates and Morton codes have problems such as large data volume and low processing efficiency. Especially for dynamically changing spatiotemporal data, there is a lack of effective coding and rapid retrieval mechanisms.

[0004] In summary, the prior art has the following disadvantages:

[0005] (1) The traditional latitude and longitude coordinate system is inefficient when dealing with large amounts of data.

[0006] (2) Although the Morton code improves the compression efficiency of spatial data, it still has limitations in processing multi-dimensional spatiotemporal data.

[0007] (3) There is a lack of coding schemes that are closely integrated with the time dimension, which makes it difficult to meet real-time dynamic management needs.

[0008] Therefore, how to quickly and accurately locate and encode geographic entities in real-life three-dimensional systems in space and time and improve the efficiency and accuracy of data processing is a technical problem that needs to be solved urgently. Summary of the invention

[0009] The technical task of the present invention is to provide a method, system, device and medium for spatiotemporal coding of geographic entities in a real-life three-dimensional system to solve the problem of how to quickly and accurately perform spatiotemporal positioning and coding of geographic entities in a real-life three-dimensional system and improve the efficiency and accuracy of data processing.

[0010] The technical task of the present invention is achieved in the following way: a method for spatiotemporal coding of geographic entities in a real-scene three-dimensional system, the method being specifically as follows:

[0011] Multi-level network division and unique identifier: Based on the GeoSOT (Geographic Spatial Object Tree) model, the earth's surface is divided into multi-level and multi-scale grids. Following the GB / T 40087-2021 Earth Space Grid Coding Rules, the grid level is determined by calculating the bounding box size of the geographic entity, and the coordinate binary left shift operation is performed at the coordinates of the entity center point. Combined with Morton code conversion and quaternary conversion technology, a unique grid code containing northern and southern hemispheres and eastern and western hemispheres is generated, namely the GeoSOT grid position vector;

[0012] BTC BeiDou time code: The BeiDou Time Code (BTC) of the BeiDou Navigation Satellite System (BDS) time service is used to generate time code, which divides time into 42 levels. The MTC fixed-length time period integer code is converted through the STC time single-granularity code conversion, and the final time code, namely the BTC time vector, is obtained through the calculation of STC0 and MTC0, so as to achieve accurate expression and unique coding of time information.

[0013] Spatiotemporal embedding model: The GeoSOT grid position vector and BTC time vector are used as inputs of the deep learning spatiotemporal embedding model. The spatiotemporal correlation is learned through a multi-layer neural network structure, and the attention mechanism and HiRA-Pro fusion method are used in the hidden layer to achieve feature fusion, generate semantically rich spatiotemporal embedding vectors, and finally output fixed-length spatiotemporal codes.

[0014] Combining Morton code and binary tree coding to optimize spatial data coding: Based on GeoSOT grid coding, Morton code and binary tree coding algorithms are introduced. Quadtrees are used for two-dimensional spatial data, and octrees are used for three-dimensional data. The space is divided into four plane quadrants and eight spatial quadrants. Morton code is used to maintain spatial proximity, and binary tree coding is used to improve data retrieval and indexing efficiency, forming an efficient spatial data coding and indexing mechanism.

[0015] Sort the spatiotemporal codes through the heap sort algorithm: By building a heap structure, fast sorting and retrieval of data can be achieved, ensuring the orderliness and fast accessibility of data, and improving the overall efficiency of large-scale data set processing.

[0016] As a preferred option, the BTC Beidou split time code is as follows:

[0017] Time coding rules: According to the time granularity segmentation rules, the geographic entity tag time is divided into 42 levels, and the entire time is split into five domains of B, C, D, E, and F for single-granularity STC time coding;

[0018] Space-time association: BTC time code records time information and is combined with GeoSOT grid location code to form a complete space-time identifier, achieving an accurate description of the space-time location of geographic entities.

[0019] Preferably, the time encoding rules are as follows:

[0020] STC is converted and expressed in 43 fixed-length time segment integer codes MTC. The maximum time granularity of MTC is 32768 years, and the minimum time granularity is 1s. It includes 42 different time granularities and is recorded in binary coding, such as 1: 1s, 2s, 4s...32s, 1min, 2min, 4min...32min, 1h (hour), 2h, 8h, 16h, 1d (day), 2d, 4d...16d, 1mon (month), 2mon, 4mon, 8mon, 1yr, 4yr...32768yr, 65536yr;

[0021] Take the middle value of STC to get STC0, and calculate MTC0 corresponding to time 0 according to the time segmentation level N;

[0022] According to MTC=(STC0>>(43-N))<<(43-N)+MTC0, the time code is calculated; wherein the time code rule divides the time into a plurality of standard time periods and assigns a unique code to each time period.

[0023] Preferably, the spatiotemporal embedding model includes an input layer, a hidden layer, and an output layer;

[0024] The input layer receives the GeoSOT grid position vector and the BTC time vector, learns the intrinsic correlation between time and space through a multi-layer neural network structure, and maps high-dimensional spatiotemporal data into a low-dimensional embedding space.

[0025] The hidden layer uses attention and HiRA-Pro fusion methods to achieve high-precision alignment of multimodal data in spatiotemporal resolution through process signatures, reduce the impact of spatiotemporal noise, improve the prediction performance of machine learning models, and effectively fuse temporal and spatial features to generate semantically rich spatiotemporal embedding vectors;

[0026] The output layer converts the spatiotemporal embedding vector into a fixed-length spatiotemporal code, which contains precise time information and reflects the spatial location of geographic entities.

[0027] More preferably, the spatial data encoding is optimized by combining Morton code and binary tree coding as follows:

[0028] Data preprocessing: Apply Morton code encoding to each point in a multidimensional data set (such as a 2D point set or a 3D voxel); and sort the points according to the value of the Morton code.

[0029] Initialize the root node: The root node represents the area of ​​the entire data set;

[0030] Recursive partitioning: Select the Morton code value of the first point (or the middle point, depending on the specific implementation) from the sorted list of data points. The Morton code value is used to determine how to partition the area represented by the current node; specifically:

[0031] For a quadtree, the decision of whether to partition along the x-axis or the y-axis is based on the bit at a specific position in the Morton code (e.g., the least significant bit);

[0032] For the octree, consider multiple bits to decide along which axis of x, y, z and the direction of division;

[0033] Recursively repeat the division process in each sub-region until the stopping condition is met (such as each sub-region contains only one point, or the maximum depth is reached);

[0034] Query: For a given query point, first calculate the Morton code of the query point, and then search in the quadtree according to the Morton code; since the Morton code maintains the locality of the data, the search can usually locate the relevant data area faster;

[0035] Retrieval: Once the relevant sub-region is located, search for data points that meet the query conditions in the corresponding sub-region.

[0036] Preferably, the spatiotemporal codes are sorted by a heap sort algorithm as follows:

[0037] Reading data: When the data set exceeds the preset size, divide the data set into multiple data blocks to ensure that the size of each data block can be placed in the memory for processing, implement data segmentation, and load each block into the memory in turn;

[0038] In-memory heap sorting: For each data block in memory, create a maximum heap or minimum heap, and use the heap sorting method to sort the data blocks in memory, that is, repeatedly remove the maximum or minimum element from the heap, put the maximum or minimum element at the end of the result array or list, and then rearrange the remaining elements to restore the heap properties;

[0039] Merge sort results: For the sorted data blocks in memory, use the external merge sort method to merge the sorted data blocks, that is, manage the merging process of data blocks on external storage (such as disk), while minimizing the number of read and write operations to improve efficiency; and create a minimum heap to maintain the minimum element of all currently unmerged data blocks, ensuring that the currently available minimum element is always taken out and merged, thereby ensuring the merge efficiency;

[0040] Detail optimization: Use multi-way merging to improve merging efficiency, optimize reading and writing, reduce the number of read and write times to external storage, and reduce disk I / O overhead through caching mechanisms; when available resources allow, parallelize data reading, in-memory heap sorting or merge sorting results, use parallel processing of multiple data blocks, or perform merging operations in parallel on multi-core processors.

[0041] A real-life 3D system geographic entity spatiotemporal coding system, based on the back-end spring framework and the front-end VUE framework, uses the pytorch deep learning library to build a spatiotemporal feature fusion model, cleans the collected data set to remove noise, trains the model, and performs spatiotemporal coding fusion, and finally integrates the above-mentioned real-life 3D system geographic entity spatiotemporal coding method; at the same time, it supports real-time dynamic geographic information management and analysis, including data collection, processing, storage, retrieval and visualization functions;

[0042] The system also uses VUE to develop front-end pages, removing redundant functions as much as possible, retaining core pages and functional points, so that users can easily operate the software platform to query, analyze and display geographic information.

[0043] Preferably, the system comprises:

[0044] The multi-level network partitioning module is used to divide the earth's surface into multi-level and multi-scale grids based on the GeoSOT (Geographic Spatial Object Tree) model. It follows the GB / T 40087-2021 Earth Space Grid Coding Rules, determines the grid level by calculating the bounding box size of the geographic entity, and performs a binary left shift operation on the coordinates of the entity center point. Combined with Morton code conversion and quaternary conversion technology, it generates a unique grid code containing northern and southern hemispheres and eastern and western hemispheres, namely the GeoSOT grid position vector;

[0045] BTC Beidou time code generation module is used to generate time code using BTC (BeiDou Time Code) of Beidou Navigation Satellite System (BDS) time service, divide time into 42 levels, convert STC time single granularity code to form MTC fixed-length time period integer code, and calculate STC0 and MTC0 to obtain the final time code, namely BTC time vector, to achieve accurate expression and unique coding of time information;

[0046] The spatiotemporal embedding model building module is used to use the GeoSOT grid position vector and BTC time vector as the input of the deep learning spatiotemporal embedding model, learn the spatiotemporal correlation through a multi-layer neural network structure, and use the attention mechanism and HiRA-Pro fusion method in the hidden layer to achieve feature fusion, generate semantically rich spatiotemporal embedding vectors, and finally output fixed-length spatiotemporal codes;

[0047] Spatial data coding optimization module is used to introduce Morton Code and binary tree coding algorithm based on GeoSOT grid coding. Quadtree is used for two-dimensional spatial data, and octree is used for three-dimensional data, dividing the space into four plane quadrants and eight spatial quadrants. Morton code is used to maintain spatial proximity, and binary tree coding is used to improve data retrieval and indexing efficiency, forming an efficient spatial data coding and indexing mechanism.

[0048] The spatiotemporal coding sorting module is used to achieve fast sorting and retrieval of data by building a heap structure, ensuring the orderliness and fast accessibility of data and improving the overall efficiency of large-scale data set processing.

[0049] An electronic device comprising: a memory and at least one processor;

[0050] Wherein, the memory stores a computer program;

[0051] The at least one processor executes the computer program stored in the memory, so that the at least one processor executes the above-mentioned method for spatiotemporal encoding of geographic entities in a real-scene three-dimensional system.

[0052] A computer-readable storage medium stores a computer program, which can be executed by a processor to implement the above-mentioned real-scene three-dimensional system geographic entity spatiotemporal encoding method.

[0053] The real-scene three-dimensional system geographic entity spatiotemporal coding method, system, device and medium of the present invention have the following advantages:

[0054] (1) The present invention introduces the GeoSOT earth segmentation model and the Beidou segmentation time code rule (BTC), combines the Morton code encoding algorithm and the multi-tree encoding algorithm, introduces the spatiotemporal feature fusion deep learning model, and fuses the time and space coding features, thereby solving the problems of low efficiency and poor accuracy of existing spatial data coordinate extraction and encoding algorithms when processing large-scale real-life three-dimensional data. It can quickly and accurately perform spatiotemporal positioning and encoding of geographic entities in the real-life three-dimensional system, thereby improving the efficiency and accuracy of data processing;

[0055] (ii) The present invention also uses a heap sorting algorithm to optimize the encoding results, further improving the operating speed and stability of the system, and is applicable to the field of geospatial grid encoding, especially having significant advantages in processing large-scale real-scene three-dimensional data;

[0056] (iii) When processing large-scale data sets, the present invention uses a heap sort algorithm to sort the spatiotemporal codes to ensure the orderliness and rapid accessibility of the data;

[0057] (iv) The present invention uses encryption technology and rights management mechanism to protect user data and ensure the data security of the software platform; at the same time, an extensible architecture is designed to support the integration and upgrade of more functional modules in the future; the user rights management mechanism is used to protect the security and privacy of user data;

[0058] (V) The present invention converts the GeoSOT grid code into a new format that integrates Morton code and binary tree code, and constructs a corresponding index structure to support fast data retrieval and update operations; wherein, the GeoSOT model designs multiple levels, from coarse-grained grids at the global level to fine-grained grids at the city level; the grid size of each level is set according to the needs of the application scenario to ensure that it can cover the global scope and meet the fine management of local areas; and the basic coding algorithm is designed according to the GB / T40087-2021 Earth Grid Coding Rules standard file to ensure that each grid unit can be located by a unique identifier; wherein, the coding rules should take into account the grid level, longitude and latitude range, and possible expansion requirements to ensure the scalability and compatibility of the coding;

[0059] (VI) The present invention selects a universal time recording type BTC time code according to the requirements of the application scenario; if there are projects in special demand fields, for fields that require high-precision time recording, a precise time recording type BTC can be used; at the same time, according to the universal time recording type BTC described in the GBT / 4257-2023 standard document, a time encoding algorithm is designed to convert time information into a coding form that is easy to store and transmit to ensure the accuracy and reliability of the coding;

[0060] (VII) The present invention collects geographic entity datasets containing temporal information and spatial information for training the spatiotemporal embedding model; wherein the datasets should cover different time points and geographical locations to ensure the generalization ability of the model;

[0061] (VIII) The present invention uses a deep learning framework (PyTorch) to train the model. By adjusting parameters such as the model structure, optimization algorithm, and loss function, the training efficiency and performance of the model are improved; at the same time, the trained model is evaluated using a test data set to verify the accuracy and efficiency of its spatiotemporal coding, and the model is further optimized and adjusted based on the evaluation results;

[0062] (IX) The present invention formulates an algorithm fusion strategy to effectively combine the Morton code and the multitree coding algorithm. The Morton code is used to maintain spatial proximity. The four quadrants of the two-dimensional spatial coding correspond to the quadtree index sorting, and the eight quadrants of the three-dimensional spatial coding correspond to the octree index sorting. The multitree coding is used to improve the efficiency of data retrieval and indexing, and supports fast data retrieval and update operations to meet the needs of real-time geographic information management.

[0063] (10) The present invention designs a heap sorting strategy for time-space coding to ensure that the continuity of the time-space relationship of data is maintained during the sorting process; the sorting strategy takes into account the distribution characteristics and access patterns of the data to improve the efficiency and accuracy of the sorting;

[0064] (XI) The present invention optimizes the performance of the heap sorting algorithm, reduces unnecessary comparison and exchange operations, and optimizes the memory access mode; at the same time, parallel computing technology (GPU acceleration) is used to further improve the sorting speed;

[0065] (XII) The present invention adopts a modular design concept and is divided into multiple independent modules to facilitate system maintenance and upgrading; and designs clear interface specifications to ensure that data exchange and communication between modules can proceed smoothly; at the same time, provides an open API interface so that third-party applications can easily access and use the system; and also implements a user authority management mechanism to ensure that different users can only access data within their authority scope, and protects the security and privacy of user data through encryption technology and security protocols. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] The present invention is further described below in conjunction with the accompanying drawings.

[0067] Attached Figure 1 The figure is a flowchart of the spatiotemporal coding method of geographic entities in a real-life three-dimensional system. DETAILED DESCRIPTION

[0068] The real-scene three-dimensional system geographic entity spatiotemporal encoding method, system, device and medium of the present invention are described in detail below with reference to the accompanying drawings and specific embodiments of the specification.

[0069] Embodiment 1:

[0070] As attached Figure 1 As shown, this embodiment provides a method for spatiotemporal coding of geographic entities in a real-scene 3D system, and the method is specifically as follows:

[0071] S1. Multi-level network division and unique identifier: Based on the GeoSOT (Geographic Spatial Object Tree) model, the earth's surface is divided into multi-level and multi-scale grids, following the GB / T 40087-2021 Earth Space Grid Coding Rules, the grid level is determined by calculating the bounding box size of the geographic entity, and the coordinate binary left shift operation is performed at the coordinates of the entity center point. Combined with Morton code conversion and quaternary conversion technology, a unique grid code containing northern and southern hemispheres and eastern and western hemispheres is generated, namely the GeoSOT grid position vector;

[0072] S2, BTC BeiDou time code: The BTC (BeiDou Time Code) of the BeiDou Navigation Satellite System (BDS) time service is used to generate the time code, which divides the time into 42 levels. The MTC fixed-length time period integer code is converted through the STC time single-granularity code conversion, and the final time code, namely the BTC time vector, is obtained through the calculation of STC0 and MTC0, so as to achieve the accurate expression and unique coding of time information.

[0073] S3, spatiotemporal embedding model: The GeoSOT grid position vector and BTC time vector are used as the input of the deep learning spatiotemporal embedding model. The spatiotemporal correlation is learned through a multi-layer neural network structure, and the attention mechanism and HiRA-Pro fusion method are used in the hidden layer to achieve feature fusion, generate semantically rich spatiotemporal embedding vectors, and finally output fixed-length spatiotemporal codes.

[0074] S4. Combining Morton Code and Binary Tree Coding to Optimize Spatial Data Coding: Based on GeoSOT grid coding, Morton Code and Binary Tree Coding Algorithm are introduced. Quadtree is used for two-dimensional spatial data, and octree is used for three-dimensional data. The space is divided into four plane quadrants and eight spatial quadrants. Morton Code is used to maintain spatial proximity, and binary tree coding is used to improve data retrieval and indexing efficiency, forming an efficient spatial data coding and indexing mechanism.

[0075] S5. Sort the spatiotemporal codes using a heap sort algorithm: By building a heap structure, fast sorting and retrieval of data can be achieved, ensuring the orderliness and fast accessibility of data, and improving the overall efficiency of large-scale data set processing.

[0076] The BTC Beidou segmentation time code in step S2 of this embodiment is as follows:

[0077] S201, time coding rule: according to the time granularity segmentation rule, the geographic entity tag time is divided into 42 levels, and the whole time is split into five domains of B, C, D, E, and F for STC time single granularity coding;

[0078] S202, time-space association: The BTC time code records time information and is combined with the GeoSOT grid location code to form a complete time-space identifier, achieving an accurate description of the time-space location of geographic entities.

[0079] The time encoding rules in step S201 of this embodiment are as follows:

[0080] S20101, STC is converted and expressed in 43 fixed-length time periods with integer codes MTC. The maximum time granularity of MTC is 32768 years, and the minimum time granularity is 1s. It includes 42 different time granularities and is recorded in binary coding, such as 1: 1s, 2s, 4s...32s, 1min, 2min, 4min...32min, 1h (hour), 2h, 8h, 16h, 1d (day), 2d, 4d...16d, 1mon (month), 2mon, 4mon, 8mon, 1yr, 4yr...32 768yr, 65536yr;

[0081] S20102, take the middle value of STC to obtain STC0, calculate MTC0 corresponding to time 0 according to the time segmentation level N;

[0082] S20103. Calculate the time code according to MTC=(STC0>>(43-N))<<(43-N)+MTC0; wherein the time code rule divides the time into a plurality of standard time periods and assigns a unique code to each time period.

[0083] The spatiotemporal embedding model in step S3 of this embodiment includes an input layer, a hidden layer and an output layer;

[0084] The input layer receives the GeoSOT grid position vector and the BTC time vector, learns the intrinsic correlation between time and space through a multi-layer neural network structure, and maps high-dimensional spatiotemporal data into a low-dimensional embedding space.

[0085] The hidden layer uses attention and HiRA-Pro fusion methods to achieve high-precision alignment of multimodal data in spatiotemporal resolution through process signatures, reduce the impact of spatiotemporal noise, improve the prediction performance of machine learning models, and effectively fuse temporal and spatial features to generate semantically rich spatiotemporal embedding vectors;

[0086] The output layer converts the spatiotemporal embedding vector into a fixed-length spatiotemporal code, which contains precise time information and reflects the spatial location of geographic entities.

[0087] The optimization of spatial data coding by combining Morton code and binary tree coding in step S4 of this embodiment is as follows:

[0088] S401, data preprocessing: applying Morton code encoding to each point in a multidimensional data set (such as a two-dimensional point set or a three-dimensional voxel); and sorting the points according to the value of the Morton code.

[0089] S402, initializing the root node: the root node represents the area of ​​the entire data set;

[0090] S403, recursive partitioning: select the Morton code value of the first point (or the middle point, depending on the specific implementation) from the sorted data point list. The Morton code value is used to determine how to partition the area represented by the current node; specifically:

[0091] For a quadtree, the decision of whether to partition along the x-axis or the y-axis is based on the bit at a specific position in the Morton code (e.g., the least significant bit);

[0092] For the octree, consider multiple bits to decide along which axis of x, y, z and the direction of division;

[0093] Recursively repeat the division process in each sub-region until the stopping condition is met (such as each sub-region contains only one point, or the maximum depth is reached);

[0094] S404, query: for a given query point, first calculate the Morton code of the query point, and then search in the quadtree according to the Morton code; wherein, since the Morton code maintains the locality of the data, the search can usually locate the relevant data area more quickly;

[0095] S405, search: once the relevant sub-region is located, search for data points that meet the query conditions in the corresponding sub-region.

[0096] In step S5 of this embodiment, the spatiotemporal codes are sorted by the heap sorting algorithm as follows:

[0097] S501, reading data: when the data set exceeds the preset size, the data set is divided into multiple data blocks to ensure that the size of each data block can be put into the memory for processing, so as to realize data segmentation, and each block is loaded into the memory in turn;

[0098] S502, in-memory heap sorting: for each data block in the memory, a maximum heap or a minimum heap is established, and the data blocks in the memory are sorted using the heap sorting method, that is, the maximum or minimum element is repeatedly removed from the heap, the maximum or minimum element is placed at the end of the result array or list, and the remaining elements are rearranged to restore the heap property;

[0099] S503, merging sort results: for the sorted data blocks in the memory, using an external merge sort method to merge the sorted data blocks, that is, managing the merging process of the data blocks on the external storage (such as disk), while minimizing the number of read and write operations to improve efficiency; and creating a minimum heap to maintain the minimum element in all the currently unmerged data blocks, ensuring that the currently available minimum element is always taken out and merged, thereby ensuring the merging efficiency;

[0100] S504, detail optimization: Use multi-way merging to improve merging efficiency, optimize reading and writing, reduce the number of read and write times to external storage, and reduce disk I / O overhead through caching mechanism; when available resources allow, parallelize data reading, in-memory heap sorting or merge sorting results, use parallel processing of multiple data blocks or perform merging operations in parallel on multi-core processors.

[0101] In this embodiment, a quadtree or other spatial indexing technology is used to replace the GeoSOT model.

[0102] In this embodiment, a quadtree or other spatial indexing technology is used to replace the GeoSOT model.

[0103] In this embodiment, different data structures and algorithms are used to optimize performance.

[0104] Embodiment 2:

[0105] This embodiment provides a real-scene three-dimensional system geographic entity spatiotemporal coding system, which is based on the back-end spring framework and the front-end VUE framework, uses the pytorch deep learning library to build a spatiotemporal feature fusion model, cleans the collected data set to remove noise, performs model training, and performs spatiotemporal coding fusion, and finally integrates the real-scene three-dimensional system geographic entity spatiotemporal coding method in Example 1; at the same time, it supports real-time dynamic geographic information management and analysis, including data collection, processing, storage, retrieval and visualization functions;

[0106] The system also uses VUE to develop front-end pages, removing redundant functions as much as possible, retaining core pages and functional points, so that users can easily operate the software platform to query, analyze and display geographic information.

[0107] The system includes:

[0108] The multi-level network partitioning module is used to divide the earth's surface into multi-level and multi-scale grids based on the GeoSOT (Geographic Spatial Object Tree) model. It follows the GB / T 40087-2021 Earth Space Grid Coding Rules, determines the grid level by calculating the bounding box size of the geographic entity, and performs a binary left shift operation on the coordinates of the entity center point. Combined with Morton code conversion and quaternary conversion technology, it generates a unique grid code containing northern and southern hemispheres and eastern and western hemispheres, namely the GeoSOT grid position vector;

[0109] BTC Beidou time code generation module is used to generate time code using BTC (BeiDou Time Code) of Beidou Navigation Satellite System (BDS) time service, divide time into 42 levels, convert STC time single granularity code to form MTC fixed-length time period integer code, and calculate STC0 and MTC0 to obtain the final time code, namely BTC time vector, to achieve accurate expression and unique coding of time information;

[0110] The spatiotemporal embedding model building module is used to use the GeoSOT grid position vector and BTC time vector as the input of the deep learning spatiotemporal embedding model, learn the spatiotemporal correlation through a multi-layer neural network structure, and use the attention mechanism and HiRA-Pro fusion method in the hidden layer to achieve feature fusion, generate semantically rich spatiotemporal embedding vectors, and finally output fixed-length spatiotemporal codes;

[0111] Spatial data coding optimization module is used to introduce Morton Code and binary tree coding algorithm based on GeoSOT grid coding. Quadtree is used for two-dimensional spatial data, and octree is used for three-dimensional data, dividing the space into four plane quadrants and eight spatial quadrants. Morton code is used to maintain spatial proximity, and binary tree coding is used to improve data retrieval and indexing efficiency, forming an efficient spatial data coding and indexing mechanism.

[0112] The spatiotemporal coding sorting module is used to achieve fast sorting and retrieval of data by building a heap structure, ensuring the orderliness and fast accessibility of data and improving the overall efficiency of large-scale data set processing.

[0113] Embodiment 3:

[0114] This embodiment also provides an electronic device, including: a memory and a processor;

[0115] Wherein, the memory stores computer-executable instructions;

[0116] The processor executes the computer-executable instructions stored in the memory, so that the processor executes the spatiotemporal encoding method for geographic entities in a real-scene three-dimensional system in any embodiment of the present invention.

[0117] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The processor may be a microprocessor or any conventional processor, etc.

[0118] The memory can be used to store computer programs and / or modules. The processor realizes various functions of the electronic device by running or executing the computer programs and / or modules stored in the memory, and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required for a function, etc.; the data storage area can store data created according to the use of the terminal, etc. In addition, the memory can also include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (SMC), a secure digital (SD) card, a flash memory card, at least one disk storage period, a flash memory device, or other volatile solid-state storage devices.

[0119] Embodiment 4:

[0120] This embodiment also provides a computer-readable storage medium, in which a plurality of instructions are stored, and the instructions are loaded by a processor, so that the processor executes the real-scene three-dimensional system geographic entity spatiotemporal encoding method in any embodiment of the present invention. Specifically, a system or device equipped with a storage medium can be provided, on which a software program code that implements the functions of any of the above embodiments is stored, and a computer (or CPU or MPU) of the system or device reads and executes the program code stored in the storage medium.

[0121] In this case, the program code itself read from the storage medium can realize the function of any one of the above-mentioned embodiments, and thus the program code and the storage medium storing the program code constitute a part of the present invention.

[0122] The storage medium embodiments for providing the program code include a floppy disk, a hard disk, a magneto-optical disk, an optical disk (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RYM, DVD-RW, DVD+RW), a magnetic tape, a non-volatile memory card, and a ROM. Alternatively, the program code can be downloaded from a server computer via a communication network.

[0123] In addition, it should be clear that the functions of any of the above embodiments can be implemented not only by executing the program code read by the computer, but also by enabling an operating system operating on the computer to complete part or all of the actual operations based on instructions from the program code.

[0124] In addition, it can be understood that the program code read from the storage medium is written to a memory provided in an expansion board inserted into the computer or written to a memory provided in an expansion unit connected to the computer, and then based on the instructions of the program code, a CPU installed on the expansion board or the expansion unit is enabled to perform part or all of the actual operations, thereby realizing the functions of any of the above-mentioned embodiments.

[0125] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for spatiotemporal coding of geographic entities in a real-scene three-dimensional system, characterized in that: The method is as follows: Multi-level network division and unique identifier: Based on the GeoSOT model, the earth's surface is divided into multi-level and multi-scale grids. The GB / T 40087-2021 Earth Space Grid Coding Rules are followed. The grid level is determined by calculating the bounding box size of the geographic entity, and the coordinate binary left shift operation is performed at the coordinates of the entity center point. Combined with Morton code conversion and quaternary conversion technology, a unique grid code containing northern and southern hemispheres and eastern and western hemispheres is generated, that is, the GeoSOT grid position vector; BTC Beidou segmented time code: BTC of Beidou satellite navigation system time service is used to generate segmented time code, which divides time into 42 levels. The MTC fixed-length time period integer code is converted through STC time single-granularity coding, and the final time code, namely BTC time vector, is obtained through calculation of STC0 and MTC0, so as to realize accurate expression and unique coding of time information. Spatiotemporal embedding model: The GeoSOT grid position vector and BTC time vector are used as inputs of the deep learning spatiotemporal embedding model. The spatiotemporal correlation is learned through a multi-layer neural network structure, and the attention mechanism and HiRA-Pro fusion method are used in the hidden layer to achieve feature fusion, generate semantically rich spatiotemporal embedding vectors, and finally output fixed-length spatiotemporal codes. Combining Morton code and binary tree coding to optimize spatial data coding: Based on GeoSOT grid coding, Morton code and binary tree coding algorithms are introduced. Quadtrees are used for two-dimensional spatial data, and octrees are used for three-dimensional data. The space is divided into four plane quadrants and eight spatial quadrants. Morton code is used to maintain spatial proximity, and binary tree coding is used to improve data retrieval and indexing efficiency, forming an efficient spatial data coding and indexing mechanism. Sort the spatiotemporal codes through the heap sort algorithm: By building a heap structure, fast sorting and retrieval of data can be achieved, ensuring the orderliness and fast accessibility of data, and improving the overall efficiency of large-scale data set processing.

2. The method for spatiotemporal coding of geographic entities in a real-scene three-dimensional system according to claim 1, characterized in that: The BTC Beidou split time code is as follows: Time coding rules: According to the time granularity segmentation rules, the geographic entity tag time is divided into 42 levels, and the entire time is split into five domains of B, C, D, E, and F for single-granularity STC time coding; Space-time association: BTC time code records time information and is combined with GeoSOT grid location code to form a complete space-time identifier, achieving an accurate description of the space-time location of geographic entities.

3. The method for spatiotemporal coding of geographic entities in a real-scene three-dimensional system according to claim 2, characterized in that: The time coding rules are as follows: STC is converted and expressed in 43 fixed-length time segment integer codes MTC. The maximum time granularity of MTC is 32768 years, and the minimum time granularity is 1s. It includes 42 different time granularities and is recorded in binary coding. Take the middle value of STC to get STC0, and calculate MTC0 corresponding to time 0 according to the time segmentation level N; According to MTC=(STC0>>(43-N))<<(43-N)+MTC0, the time code is calculated; wherein the time code rule divides the time into a plurality of standard time periods and assigns a unique code to each time period.

4. The method for spatiotemporal coding of geographic entities in a real-scene three-dimensional system according to claim 3, characterized in that: The spatiotemporal embedding model includes an input layer, a hidden layer, and an output layer; The input layer receives the GeoSOT grid position vector and the BTC time vector, learns the intrinsic correlation between time and space through a multi-layer neural network structure, and maps high-dimensional spatiotemporal data into a low-dimensional embedding space. The hidden layer uses attention and HiRA-Pro fusion methods to achieve high-precision alignment of multimodal data in spatiotemporal resolution through process signatures, and effectively fuses temporal and spatial features to generate semantically rich spatiotemporal embedding vectors; The output layer converts the spatiotemporal embedding vector into a fixed-length spatiotemporal code, which contains precise time information and reflects the spatial location of geographic entities.

5. The method for spatiotemporal coding of geographic entities in a real-scene three-dimensional system according to claim 4, characterized in that: The optimization of spatial data encoding by combining Morton code and binary tree coding is as follows: Data preprocessing: Apply Morton code encoding to each point in the multidimensional dataset; and sort the points according to the value of the Morton code. Initialize the root node: The root node represents the area of ​​the entire data set; Recursive partitioning: Select the Morton code value of the first point from the sorted list of data points. The Morton code value is used to determine how to partition the area represented by the current node. Specifically: For a quadtree, the decision on whether to partition along the x-axis or y-axis is based on the bit at a specific position in the Morton code; For the octree, consider multiple bits to decide along which axis of x, y, z and the direction of division; Recursively repeat the partitioning process in each sub-region until the stopping condition is met; Query: For a given query point, first calculate the Morton code of the query point, and then search in the quadtree according to the Morton code; Retrieval: Once the relevant sub-region is located, search for data points that meet the query conditions in the corresponding sub-region.

6. The method for spatiotemporal coding of geographic entities in a real-scene three-dimensional system according to claim 5, characterized in that: The heap sort algorithm is used to sort the space-time codes as follows: Reading data: When the data set exceeds the preset size, divide the data set into multiple data blocks to ensure that the size of each data block can be placed in the memory for processing, implement data segmentation, and load each block into the memory in turn; In-memory heap sorting: For each data block in memory, create a maximum heap or minimum heap, and use the heap sorting method to sort the data blocks in memory, that is, repeatedly remove the maximum or minimum element from the heap, put the maximum or minimum element at the end of the result array or list, and then rearrange the remaining elements to restore the heap properties; Merge sort results: For the sorted data blocks in the memory, use the external merge sort method to merge the sorted data blocks, that is, manage the merging process of data blocks on the external storage, reduce the number of read and write operations to improve efficiency; And create a minimum heap to maintain the minimum element in all the currently unmerged data blocks, ensuring that the currently available minimum element is always taken out and merged, thereby ensuring the efficiency of merging; Detail optimization: Use multi-way merging to improve merging efficiency, optimize reading and writing, reduce the number of read and write times to external storage, and reduce disk I / O overhead through caching mechanisms; when available resources allow, parallelize data reading, in-memory heap sorting or merge sorting results, use parallel processing of multiple data blocks, or perform merging operations in parallel on multi-core processors.

7. A real-scene three-dimensional system geographic entity spatiotemporal coding system, characterized in that: The system is based on the back-end spring framework and the front-end VUE framework, uses the pytorch deep learning library to build a spatiotemporal feature fusion model, cleans the collected data set to remove noise, performs model training, and performs spatiotemporal coding fusion, and finally integrates the real-scene three-dimensional system geographic entity spatiotemporal coding method described in any one of claims 1-6; The system also uses VUE to develop front-end pages, maximizing the removal of redundant functions and retaining core pages and functional points, so that users can easily operate the software platform to query, analyze and display geographic information.

8. The real-scene three-dimensional system geographic entity spatiotemporal coding system according to claim 7, characterized in that: The system includes: The multi-level network partitioning module is used to divide the earth's surface into multi-level and multi-scale grids based on the GeoSOT model. It follows the GB / T 40087-2021 Earth Space Grid Coding Rules, determines the grid level by calculating the bounding box size of the geographic entity, and performs a binary left shift operation on the coordinates of the entity center point. Combined with Morton code conversion and quaternary conversion technology, it generates a unique grid code containing northern and southern hemispheres and eastern and western hemispheres, namely the GeoSOT grid position vector; BTC Beidou time code generation module is used to generate time code using BTC of Beidou satellite navigation system time service, divide time into 42 levels, convert STC time single granularity code to form MTC fixed-length time period integer code, and calculate STC0 and MTC0 to obtain the final time code, namely BTC time vector, to achieve accurate expression and unique coding of time information; The spatiotemporal embedding model building module is used to use the GeoSOT grid position vector and BTC time vector as the input of the deep learning spatiotemporal embedding model, learn the spatiotemporal correlation through a multi-layer neural network structure, and use the attention mechanism and HiRA-Pro fusion method in the hidden layer to achieve feature fusion, generate semantically rich spatiotemporal embedding vectors, and finally output fixed-length spatiotemporal codes; Spatial data coding optimization module is used to introduce Morton code and binary tree coding algorithm based on GeoSOT grid coding. Quadtree is used for two-dimensional spatial data, and octree is used for three-dimensional data, dividing the space into four plane quadrants and eight spatial quadrants. Morton code is used to maintain spatial proximity, and binary tree coding is used to improve data retrieval and indexing efficiency, forming an efficient spatial data coding and indexing mechanism. The spatiotemporal coding sorting module is used to achieve fast sorting and retrieval of data by building a heap structure, ensuring the orderliness and fast accessibility of data and improving the overall efficiency of large-scale data set processing.

9. An electronic device, characterized in that: include: memory and at least one processor; Wherein, the memory stores a computer program; The at least one processor executes the computer program stored in the memory, so that the at least one processor executes the spatiotemporal encoding method of geographic entities in a real-life three-dimensional system as described in any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which can be executed by a processor to implement the spatiotemporal encoding method of geographic entities in a real-life three-dimensional system according to any one of claims 1 to 6.

Images