Historical and cultural block three-dimensional space data rapid acquisition and display method and system

Abstract

The embodiment of the present application relates to the technical field of space collection, and discloses a kind of historical and cultural block three-dimensional space data fast collection and display method, comprising: obtaining the multi-source three-dimensional space data of historical and cultural block by cooperative collection system;Multi-source three-dimensional space data is input to semantic data processing engine;In semantic data processing engine, multi-source three-dimensional space data is processed to generate digital twin model including semantic information;Through interactive display platform, digital twin model is presented to user and based on semantic information provides interactive function.The method in the embodiment of the present application can integrate ground scanning, unmanned aerial vehicle aerial photography, handheld device detection and other various collection methods, which can not only quickly cover the large area of block roof, high-altitude facilities and other areas by unmanned aerial vehicle, but also rely on handheld device to penetrate into narrow alley and other hidden corners, to make up for the blind area of single collection method.Based on semantic digital twin model, the block can be comprehensively and richly displayed.

Description

Technical Field

[0001] This invention relates to the field of three-dimensional spatial acquisition technology, specifically to a method and system for rapid acquisition and display of three-dimensional spatial data of historical and cultural blocks. Background Technology

[0002] Currently, the protection and renewal of historical and cultural blocks heavily rely on accurate and comprehensive three-dimensional spatial data. Existing technologies, such as total station surveying, ground close-range photogrammetry, and UAV oblique photography, suffer from drawbacks including low efficiency, susceptibility to environmental factors, blind spots in data acquisition, inability to capture indoor and detailed information, and a lack of semantic information in the generated data. This results in subsequent displays and applications remaining at a simple 3D browsing level, failing to achieve a deep, immersive, and interactive experience integrated with the historical context. Therefore, designing a solution capable of efficient data acquisition and display has become a pressing technical problem for those skilled in the art. Summary of the Invention

[0003] To address the aforementioned shortcomings, this invention discloses a rapid acquisition and display system for three-dimensional spatial data of historical and cultural blocks, which enables efficient acquisition and display of three-dimensional spatial data.

[0004] The first aspect of this invention discloses a system for rapid acquisition and display of three-dimensional spatial data of historical and cultural blocks, comprising: Acquisition module: Used to acquire multi-source 3D spatial data of historical and cultural blocks through a collaborative acquisition system; Processing module: used to input the multi-source 3D spatial data into the semantic data processing engine; in the semantic data processing engine, the multi-source 3D spatial data is processed to generate a digital twin model including semantic information; Display module: Used to present the digital twin model to users through an interactive display platform and provide interactive functions based on the semantic information.

[0005] The second aspect of this invention discloses a method for rapid acquisition and display of three-dimensional spatial data of historical and cultural blocks, comprising: Multi-source three-dimensional spatial data of historical and cultural blocks are acquired through a collaborative acquisition system; The multi-source 3D spatial data is input into a semantic data processing engine; the semantic data processing engine processes the multi-source 3D spatial data to generate a digital twin model including semantic information. The digital twin model is presented to users through an interactive display platform, and interactive functions are provided based on the semantic information.

[0006] As an optional implementation, in a second aspect of the present invention, the collaborative acquisition system includes a cluster of unmanned aerial vehicles (UAVs) and a ground-based mobile acquisition unit. The acquisition of multi-source three-dimensional spatial data of historical and cultural blocks through the collaborative acquisition system includes: Receive initial scan data collected by a drone swarm of historical and cultural blocks, and obtain an initial real-world 3D model based on the initial scan data; Based on the initial real-world 3D model, the corresponding areas to be optimized are determined, including blind spots and detail areas. The location information of the area to be optimized is sent to the corresponding ground mobile acquisition unit, and the supplementary spatial data collected by the ground mobile acquisition unit is received. Multi-source 3D spatial data is obtained based on the initial real-world 3D model and supplementary spatial data.

[0007] As an optional implementation, in a second aspect of the present invention, after obtaining the initial real-world 3D model based on the initial scan data, the method further includes: Based on the initial real-world 3D model, determine the data missing areas and model quality score of the historical and cultural district. When the data collection status of the drone cluster is stable, the maximum acceptable area of ​​the missing data region is determined based on historical data collection. Obtain the current area of ​​the data missing region, and determine the data integrity characterization value corresponding to the block based on the current area and the maximum acceptable area; When the data collection state corresponding to the drone cluster is not stable, the data integrity characterization value corresponding to the corresponding block is determined according to the rate of change of the model quality score. Before sending the location information of the area to be optimized to the corresponding ground mobile acquisition unit, the method further includes: Calculate the comprehensive data acquisition strategy score, which is a weighted sum based on intermediate evaluation parameters, including data completeness index, spatial accessibility score, equipment capability matching degree, and historical and cultural value. The acquisition strategy is determined by comparing the comprehensive acquisition strategy score with the set strategy range. The acquisition strategy includes an aggressive strategy, a balanced strategy, or a passive strategy.

[0008] As an optional implementation, in the second aspect of the present invention, before acquiring the multi-source three-dimensional spatial data of the historical and cultural block through the collaborative acquisition system, the method further includes: Based on the received geographic information data, the historical and cultural blocks are divided into multiple collection units with different spatial and collection characteristics. A corresponding acquisition feature vector is generated for each acquisition unit. The acquisition feature vector includes unit type, average width, facade complexity, and recommended acquisition method. The recommended acquisition method includes UAV acquisition or personnel acquisition. Multiple data collection units are identified as the set of data collection points that need to be served. Model drone data collection methods and personnel data collection methods as heterogeneous service points; Construct a data acquisition path model with the goal of minimizing the total data acquisition completion time. The model constraints include the service time window of each data acquisition unit, the endurance or working time constraints of each data acquisition method, and the requirement constraints of each unit for the data acquisition type. The acquisition path model is calculated by optimizing the algorithm, generating an ordered sequence of task units for each UAV acquisition method and each personnel acquisition method, and transmitting the sequence of task units to the corresponding collaborative acquisition system.

[0009] As an optional implementation, in a second aspect of the present invention, the acquisition of multi-source three-dimensional spatial data of historical and cultural blocks through a collaborative acquisition system further includes: Based on the current acquisition progress and environmental perception data, a first challenge distribution model is constructed in the digital twin model to characterize the spatial distribution of the current acquisition challenges; the acquisition challenges include structural occlusion areas, dynamic interference areas, or areas requiring high-precision acquisition. The candidate digital performance models are virtually placed at different candidate deployment points within the digital twin model; The locations where the digital performance model does not interfere with the first challenge distribution model are determined as the first set of feasible deployment points; wherein, the performance interference refers to the inability to effectively overcome the corresponding challenge and complete the data collection task at that location.

[0010] As an optional implementation, in a second aspect of the present invention, the method further includes: After the data acquisition system operates for the preset duration according to the first optimized working mode, a second challenge distribution model is constructed in the digital twin model based on the updated environmental perception data. Calculate the similarity of scene changes between the second challenge distribution model and the first challenge distribution model; If the similarity of the scene changes is lower than the preset change threshold, a global replanning is triggered, and the optimal deployment location and working mode of the acquisition system are re-determined based on the second challenge distribution model. If the similarity of the scene changes is not lower than the preset change threshold, local parameter optimization is triggered, and the optimized working mode of the acquisition system is re-determined based on the current data features at the first optimized deployment location.

[0011] As an optional implementation, in a second aspect of the present invention, the step of processing the multi-source three-dimensional spatial data in the semantic data processing engine to generate a digital twin model including semantic information includes: The point cloud semantic segmentation neural network deployed in the data processing engine is used to process the spatially registered and fused 3D point cloud data to classify the points in the point cloud data into predefined building component categories. For each subset of building component point clouds identified by the point cloud semantic segmentation neural network, a cross-modal feature fusion model deployed in the data processing engine is used for processing. The cross-modal feature fusion model is used to perform the following steps: Receive the point cloud subset of the building components; Receive image data blocks that spatially correspond to the building components; Output architectural style detail tags associated with the building components; The image data blocks are analyzed using a material analysis convolutional neural network deployed in the data processing engine to output the surface material classification results of the building components.

[0012] As an optional implementation, in a second aspect of the present invention, the data processing engine further includes: From an unstructured text database related to historical and cultural blocks, an entity recognition model is used to extract multiple historical entities and multiple historical relationships. The historical entities include building names and historical event names, and the historical relationships include construction time and spatial adjacency. Each extracted historical entity is associated with a subset of building component point clouds identified by the point cloud semantic segmentation neural network. The association is based on the matching degree between the spatial location information of the historical entity described in the text and the three-dimensional coordinates of the building component point cloud subset in the three-dimensional point cloud data. The associated results are stored as nodes and edges in a graph structure database, where each building component and each historical entity is stored as a node, each historical relationship is stored as an edge connecting two nodes, and each building component node includes a set of semantic attributes obtained from the semantic understanding process.

[0013] As an optional implementation, in a second aspect of the present invention, presenting the digital twin model to the user through an interactive display platform and providing interactive functions based on the semantic information includes: Receive natural language query strings input by users through an interactive display platform; The natural language query string is sent to the data processing engine for parsing to convert the query string into a query statement for the graph structure database; Execute the query statement to obtain a list of three-dimensional component identifiers that meet the query conditions; When rendering the digital twin model, the interactive display platform is controlled to highlight the geometric parts corresponding to the list of three-dimensional component identifiers; The interactive display platform is controlled to display detailed semantic attributes and historical information text associated with the list of three-dimensional component identifiers, retrieved from the graph structure database, in the user interface panel.

[0014] Compared with the prior art, the embodiments of the present invention have the following beneficial effects: The method in this invention integrates multiple data collection methods, including ground scanning, drone aerial photography, and handheld device detection. It can quickly cover large areas such as street rooftops and high-altitude facilities using drones, while also utilizing handheld devices to penetrate narrow alleys and other hidden corners, compensating for the blind spots of single data collection methods. Furthermore, based on a semantic digital twin model, it can provide a comprehensive and rich display of the streetscape. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a flowchart illustrating the method for rapid acquisition and display of three-dimensional spatial data of historical and cultural blocks disclosed in an embodiment of the present invention; Figure 2 This is a schematic diagram of the process for acquiring multi-source three-dimensional spatial data disclosed in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of a system for rapid acquisition and display of three-dimensional spatial data of historical and cultural blocks provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] It should be noted that the terms first, second, third, fourth, etc., in the specification and claims of this invention are used to distinguish different objects, not to describe a specific order. The terms used in the embodiments of this invention include and have, and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products, or devices.

[0019] Example 1 Please see Figure 1 , Figure 1 This is a flowchart illustrating the method for rapid acquisition and display of 3D spatial data of historical and cultural blocks disclosed in this invention. The execution entity of the method described in this invention is an execution entity composed of software and / or hardware. This execution entity can receive relevant information via wired and / or wireless means and can send certain instructions. It can also have certain processing and storage functions. This execution entity can control multiple devices, such as remote physical servers or cloud servers and related software, or local hosts or servers and related software that perform related operations on devices located in a certain place. In some scenarios, multiple storage devices can also be controlled; these storage devices can be placed in the same location as the devices or in different locations. Figure 1 As shown, this method for rapid acquisition and display of 3D spatial data of historical and cultural blocks includes the following steps: S101: Acquire multi-source three-dimensional spatial data of historical and cultural blocks through a collaborative acquisition system; S102: Input the multi-source three-dimensional spatial data into the semantic data processing engine; in the semantic data processing engine, process the multi-source three-dimensional spatial data to generate a digital twin model including semantic information; S103: Present the digital twin model to the user through an interactive display platform and provide interactive functions based on the semantic information.

[0020] In practice, the system primarily employs non-contact data acquisition methods, such as 3D laser scanning and drone oblique photography, eliminating the need for physical contact with the historical buildings. This avoids potential damage to ancient buildings, such as scratches and structural disturbances, caused by traditional measuring tools, making it particularly suitable for data collection needs of key protected objects like ancient buildings and structures in historical and cultural districts.

[0021] In this embodiment of the invention, the semantic data processing engine can fuse and optimize multi-source data, such as integrating spatial coordinate data of laser point clouds with texture data of images. The generated model after processing can completely restore the micro-details of building carvings, weathering marks, etc., and can also present building materials through texture mapping, giving the model a photorealistic feel and providing accurate digital archives for preserving the streetscape.

[0022] Unlike simple 3D geometric models, the models generated by this engine include semantic information, such as the building's construction date, structural type, and facility functions. This semantic feature transforms the model from a mere visualization tool into a platform for intelligent analysis. For example, it can automatically identify repair traces on buildings from different eras or distinguish between different spatial types such as streets, squares, and ancient buildings, providing structured data support for subsequent management and research.

[0023] The interactive display platform of this invention can be adapted to different user scenarios. For tourists, the platform enables AR real-view navigation and allows viewing of building history; for managers, it allows access to facility information and spatial layout; and for researchers, it allows measurement of building dimensions and analysis of spatial relationships through interactive functions. Different groups can efficiently obtain the information they need.

[0024] The above methods can be used not only by scenic area staff but also by delegating the details of street-level data collection to everyone for optimization. By designing a method that allows for broader participation in the data collection process and setting up corresponding reward mechanisms, the comprehensiveness of data collection can be optimized, and visitor participation can be increased. The reward mechanism here needs to be officially certified.

[0025] Specifically, in implementation, a dedicated crowdsourcing app is developed, allowing administrators, merchants, and volunteers to upload mobile photos or videos of specific areas. The system then utilizes visual SLAM and photogrammetry technologies to perform near real-time registration and fusion of this non-professional data with a high-precision base model. Furthermore, blockchain technology is introduced to document the source, time, and contributors of the crowdsourced data, ensuring data credibility and enabling rapid detection and updates of model changes.

[0026] Once crowdsourced data is uploaded, the system not only integrates it but also uses generative AI models to automatically generate simulation data for other perspectives and different lighting conditions at that location, based on the photo's angle and content.

[0027] In this embodiment of the invention, the crowdsourced data acquisition module is used to receive and verify local image data of a street from a user terminal, and to fuse the verified local image data with the preliminary real-world 3D model. The crowdsourced data acquisition module further includes: using blockchain technology to store the source and time of the local image data; and using a generative artificial intelligence model to generate multiple simulated viewpoint images of the location based on the local image data, for training and improving the recognition model in the system.

[0028] More preferably, such as Figure 2 As shown, the collaborative acquisition system includes a drone swarm and a ground-based mobile acquisition unit. The acquisition of multi-source three-dimensional spatial data of historical and cultural blocks through the collaborative acquisition system includes: S1011: Receive initial scan data collected by a drone cluster on a historical and cultural block, and obtain an initial real-world 3D model based on the initial scan data; S1012: Determine the corresponding area to be optimized based on the initial real-world 3D model, the area to be optimized including the blind spot and the detail area; S1013: Send the location information of the area to be optimized to the corresponding ground mobile acquisition unit, and receive the supplementary spatial data collected by the ground mobile acquisition unit; S1014: Obtain multi-source three-dimensional spatial data based on the initial real-world three-dimensional model and supplementary spatial data.

[0029] In practice, drone swarms can conduct parallel aerial photography and laser scanning of the entire street area. Relying on the swarm's collaborative operation capabilities, they can quickly cover a wide range of areas, including building rooftops, the entire streetscape, and public spaces in historical and cultural districts.

[0030] By accurately identifying blind spots and detailed areas in the initial 3D model, such as narrow alleys, building gaps, underground space entrances, carved patterns, inscriptions, and building component joints, supplementary data collection is performed only on the areas requiring optimization using ground-based mobile data collection units. This avoids blindly traversing the entire area by ground-based units, reduces unnecessary data collection, and significantly shortens the overall data collection cycle. Furthermore, combining the data from drone scans allows for the determination of corresponding location information, facilitating user positioning later.

[0031] Due to factors such as flight angle and obstructions, drone swarms have difficulty covering some hidden areas. However, ground-based mobile data acquisition units can flexibly enter narrow spaces, building interiors, and other areas that drones cannot reach. By supplementing data collection, they can fill data gaps in blind spots, ensuring that multi-source 3D spatial data covers all physical spaces in the block and avoiding the data loss problems caused by traditional single acquisition methods.

[0032] The core value of historical and cultural blocks is often reflected in the micro-details of architectural carvings, ancient inscriptions, and traditional craft decorations. Ground-based mobile acquisition units can be equipped with high-precision laser scanners, macro cameras, and other devices to collect detailed areas at close range and in high resolution, obtaining high-precision spatial data and texture information. This ensures that the final multi-source 3D spatial data not only meets the requirements for restoring the overall macro-layout but also accurately presents the micro-detail features, providing high-quality data support for subsequent semantic modeling and cultural value mining.

[0033] The initial real-world 3D model provides a global spatial framework and basic texture information, supplementing the spatial data with high-precision data for focusing blind spots and details. The multi-source 3D spatial data formed by the fusion of the two has both the integrity of global coverage and the high precision of local details. At the same time, through the unified coordinate system of the collaborative acquisition system, such as GNSS positioning calibration, the consistency of spatial coordinates of data from different sources is ensured, avoiding problems such as data misalignment and splicing gaps, and improving the overall reliability of the data.

[0034] Specifically, the steps for supplementary data collection include: acquiring color and depth images using an RGB-D camera integrated on the ground mobile acquisition unit; and scanning surface details of building components using a macro laser scanner integrated on the ground mobile acquisition unit.

[0035] More preferably, after obtaining the initial real-world 3D model based on the initial scan data, the method further includes: Based on the initial real-world 3D model, determine the data missing areas and model quality score of the historical and cultural district. When the data collection status of the drone cluster is stable, the maximum acceptable area of ​​the missing data region is determined based on historical data collection. Obtain the current area of ​​the data missing region, and determine the data integrity characterization value corresponding to the block based on the current area and the maximum acceptable area; When the data collection state corresponding to the drone cluster is not stable, the data integrity characterization value corresponding to the corresponding block is determined according to the rate of change of the model quality score. Before sending the location information of the area to be optimized to the corresponding ground mobile acquisition unit, the method further includes: Calculate the comprehensive data acquisition strategy score, which is a weighted sum based on intermediate evaluation parameters, including data completeness index, spatial accessibility score, equipment capability matching degree, and historical and cultural value. The acquisition strategy is determined by comparing the comprehensive acquisition strategy score with the set strategy range. The acquisition strategy includes an aggressive strategy, a balanced strategy, or a passive strategy.

[0036] Specifically, different data integrity assessment logics are adopted for the two acquisition states of drone swarms: stable and unstable. In the stable state, the maximum acceptable area of ​​the missing region is used as the benchmark, while in the unstable state, the rate of change of the model quality score is used as the core indicator. This avoids misjudgment caused by a single assessment standard, such as overemphasizing minor missing parts in stable acquisition or ignoring the continuous decline in quality in unstable acquisition, and ensures that the data integrity representation value can truly reflect the acquisition effect.

[0037] By combining data integrity characterization values ​​with the location of missing areas and model quality scores, complete information on defect localization and severity quantification is formed. This transforms the re-collection work of ground-based mobile data acquisition units from fuzzy re-collection to precise re-collection, avoiding insufficient or excessive re-collection. The data integrity characterization values ​​are used to determine whether to continue data collection.

[0038] In addition to using data integrity metrics for judgment, a comprehensive evaluation can be conducted. The comprehensive data collection strategy score encompasses four core parameters: data completeness index, spatial accessibility score, equipment capability matching degree, and historical and cultural value. Through weighted calculations, a multi-dimensional quantitative assessment of supplementary data collection needs is achieved, avoiding strategy errors caused by traditional experience-based decision-making, such as aggressive collection in high-value but low-accessibility areas or excessive resource allocation to low-value areas. This significantly improves data collection efficiency. Specifically, the aggressive strategy involves immediately scheduling the optimal resource network for collection; the balanced strategy involves including data in a task queue and performing regular resource scheduling; and the passive strategy involves only recording data and waiting for crowdsourcing or the next full-domain scan.

[0039] Specifically, when the acquisition strategy is aggressive: the region is micro-partitioned based on the geometric texture complexity and historical and cultural value weight of the data-missing region; a first type of fine acquisition path is planned for high-complexity, high-value partitions, a second type of standard acquisition path is planned for medium-complexity partitions, and a third type of fast acquisition path is planned for low-complexity partitions; based on the path optimization algorithm, the first, second, and third types of paths are collaboratively calculated with the starting positions of available acquisition devices to generate a multi-device collaborative acquisition sequence with the shortest total time.

[0040] When the data collection strategy is passive, the data-missing area is divided into multiple crowdsourced sub-grids. The real-time location and mobile terminal type of the user are sensed through a sensor network deployed in the street. When a user enters the preset range of a crowdsourced sub-grid and the mobile terminal type meets the data collection requirements, an AR navigation interface is automatically pushed to the user's terminal. The AR navigation interface is used to guide the user to the best data collection point and assist them in completing the data collection operation.

[0041] More preferably, before acquiring the multi-source three-dimensional spatial data of the historical and cultural block through the collaborative acquisition system, the method further includes: Based on the received geographic information data, the historical and cultural blocks are divided into multiple collection units with different spatial and collection characteristics. A corresponding acquisition feature vector is generated for each acquisition unit. The acquisition feature vector includes unit type, average width, facade complexity, and recommended acquisition method. The recommended acquisition method includes UAV acquisition or personnel acquisition. Multiple data collection units are identified as the set of data collection points that need to be served. Model drone data collection methods and personnel data collection methods as heterogeneous service points; Construct a data acquisition path model with the goal of minimizing the total data acquisition completion time. The model constraints include the service time window of each data acquisition unit, the endurance or working time constraints of each data acquisition method, and the requirement constraints of each unit for the data acquisition type. The acquisition path model is calculated by optimizing the algorithm, generating an ordered sequence of task units for each UAV acquisition method and each personnel acquisition method, and transmitting the sequence of task units to the corresponding collaborative acquisition system.

[0042] Traditional data collection often employs partitioned traversal or random allocation methods, which can easily lead to inefficiencies such as overlapping tasks between drones and personnel, path reversals, and waiting for refueling. This solution constructs a path model with the goal of minimizing the total data collection completion time, integrates the requirements of all data collection units and the constraints of heterogeneous service points, and generates a globally optimal sequence of task units, thus avoiding invalid path consumption and task waiting time.

[0043] Specifically, the model incorporates service time window constraints for each acquisition unit, and optimizes the algorithm to achieve precise task timing. This avoids quality degradation due to inappropriate timing (e.g., backlighting affecting texture accuracy) and resource idleness caused by waiting time windows, improving the continuity and efficiency of the acquisition process. The optimization algorithm employs a genetic algorithm. Each chromosome encodes an acquisition scheme containing multiple sub-paths. A fitness function is designed to evaluate each chromosome, and its value is negatively correlated with the total estimated time of the scheme. During the crossover or mutation operations of the genetic algorithm, the task of an acquisition unit can be simultaneously assigned to both drones and personnel, indicating collaborative data acquisition for that unit.

[0044] To address constraints such as drone endurance limitations and personnel work time limits, the path model can intelligently break down task units and plan charging / rest nodes. For example, it can plan a return charging route before the drone's endurance reaches its limit, or insert reasonable rest intervals into the personnel task sequence to avoid task stagnation caused by resource interruptions and ensure the continuous and efficient progress of the data collection process.

[0045] In practical implementation, a feature vector is generated for each collection unit, including unit type, average width, facade complexity, and recommended collection method, to achieve quantitative representation of collection requirements. For example, for units with wide streets and low facade complexity, a drone swarm is recommended for efficient collection, while for units with narrow alleys and high facade complexity, personnel carrying portable equipment are recommended for accurate collection.

[0046] By modeling drone data collection and personnel data collection as heterogeneous service points, the limitations of traditional single resource scheduling are overcome. Through model algorithms, the two types of resources complement and coordinate. Drones are responsible for collecting the overall framework of the street, while personnel simultaneously follow up with detailed data collection, forming a data collection mode with global coverage and local precision, which ensures both collection efficiency and data quality.

[0047] Specifically, for UAV-based data acquisition, the following approach can be used when acquiring data from a facade: For the target facade, calculate the required set of sampling points covering its surface; within the safe airspace where the UAV can fly, generate a set of candidate viewpoints, where each viewpoint corresponds to a UAV pose that can cover part of the sampling point set; select a minimum subset of viewpoints from the candidate viewpoint set, such that the subset of viewpoints can jointly cover the entire sampling point set and satisfy a preset image overlap rate constraint; based on the selected subset of viewpoints, use a path smoothing algorithm to generate a flight trajectory for the UAV to perform data acquisition within the unit.

[0048] More preferably, the acquisition of multi-source three-dimensional spatial data of historical and cultural blocks through a collaborative acquisition system further includes: Based on the current acquisition progress and environmental perception data, a first challenge distribution model is constructed in the digital twin model to characterize the spatial distribution of the current acquisition challenges; the acquisition challenges include structural occlusion areas, dynamic interference areas, or areas requiring high-precision acquisition. The candidate digital performance models are virtually placed at different candidate deployment points within the digital twin model; The locations where the digital performance model does not interfere with the first challenge distribution model are determined as the first set of feasible deployment points; wherein, the performance interference refers to the inability to effectively overcome the corresponding challenge and complete the data collection task at that location.

[0049] The solution of this invention is based on the current collection progress and environmental perception data. It constructs a first challenge distribution model in the digital twin model, transforming abstract challenges such as building gaps, tree obstructions, densely populated areas, temporary construction areas, carved patterns, and inscribed text areas into spatial distribution data. This ensures that all collection difficulties are accurately identified without omission.

[0050] The digital performance models of candidate data acquisition devices, such as the scanning range, accuracy threshold, and anti-interference capabilities of drones, and the effective detection distance and operating space requirements of personnel data acquisition devices, are virtually placed in a digital twin model. This allows for the prediction of the suitability of device deployment locations for data acquisition challenges in advance. For example, it avoids deploying drones that rely on direct viewing angles in structurally obstructed areas, and avoids deploying data acquisition devices that require long-term static operation in densely populated and dynamically disturbed areas. This approach avoids performance interference problems that cannot be overcome after device deployment from the source.

[0051] For the high-precision data collection areas identified in the first challenge distribution model, the optimal deployment points that meet the accuracy thresholds are selected through virtual verification of the digital performance model, providing high-quality data support for subsequent semantic modeling and cultural value mining.

[0052] The digital performance model mentioned in this embodiment of the invention includes the core parameters of the device (such as battery life, accuracy, and operational constraints). Combined with the spatial characteristics of the first challenge distribution model, the first set of feasible deployment points naturally adapts to the device performance and scenario requirements. Long-endurance, large-area scanning drones are deployed in open areas without dynamic interference, while high-precision, small-area operation personnel data collection devices are deployed in narrow alleys or detailed areas of ancient buildings.

[0053] Specifically, through virtual simulation using a digital twin model, all feasible deployment points are determined in advance. During the data collection process, there is no need to frequently adjust equipment positions or change collection methods, avoiding unnecessary waste such as equipment backtracking and waiting due to improper deployment, and reducing energy consumption and unnecessary labor for operators. Traditional data collection deployment requires on-site testing to verify the feasibility of equipment locations, which is time-consuming and may interfere with historical and cultural blocks. This solution, through virtual placement and performance verification using a digital twin model, completes the selection of deployment points in virtual space, determining the optimal solution without on-site trial and error.

[0054] The challenge distribution model is built based on the current collection progress and real-time environmental perception data. It has the ability to be dynamically updated. For example, when there is a sudden peak in pedestrian traffic in the block (the dynamic interference area expands) or the weather changes (such as the decrease in light intensity affecting texture collection), the first challenge distribution model can be updated in real time, and feasible deployment points can be re-selected to ensure that the collection deployment decision always adapts to the dynamically changing scenario and avoids collection failure caused by sudden environmental changes.

[0055] The spatial distribution of data collection challenges, equipment performance parameters, and deployment site feasibility are all quantified and characterized through data. The entire decision-making process is traceable and verifiable, allowing project managers to clearly understand the logic and basis of deployment decisions, avoiding biases caused by experience-based decisions, and improving the overall controllability of the data collection process. By selecting feasible deployment points, sensitive spaces such as the vicinity of fragile ancient buildings and densely populated areas can be avoided, thus preventing physical disturbances caused by close contact between equipment and ancient buildings, or interference with the visitor experience and daily operation of the street. At the same time, virtual verification reduces the additional personnel and equipment movement caused by on-site trial and error, further reducing the impact on the street's appearance and safety.

[0056] More preferably, the method further includes: After the data acquisition system operates for the preset duration according to the first optimized working mode, a second challenge distribution model is constructed in the digital twin model based on the updated environmental perception data. Calculate the similarity of scene changes between the second challenge distribution model and the first challenge distribution model; If the similarity of the scene changes is lower than the preset change threshold, a global replanning is triggered, and the optimal deployment location and working mode of the acquisition system are re-determined based on the second challenge distribution model. If the similarity of the scene changes is not lower than the preset change threshold, local parameter optimization is triggered, and the optimized working mode of the acquisition system is re-determined based on the current data features at the first optimized deployment location.

[0057] During the data collection process, the environmental conditions of historical and cultural blocks may change dynamically over time. By constructing a second challenge distribution model based on updated environmental perception data after a preset time period, these dynamic changes can be captured in real time. By calculating the similarity of scene changes between the second and first challenge distribution models, scene changes are quantified into a quantifiable indicator. When the similarity is lower than a preset threshold (drastic scene changes, such as temporary closure of the core collection area or sudden severe weather), a global replanning is triggered, completely adjusting the deployment location and working mode. When the similarity is not lower than the threshold (minor scene changes, such as a local increase in pedestrian traffic or slight fluctuations in light intensity), only local parameter optimizations are performed (such as adjusting the drone's flight altitude and the photosensitivity of personnel's data collection equipment).

[0058] The scene change similarity in this embodiment of the invention is a quantitative and fusion-based metric used to assess the overall pattern of changes in the challenges faced by the historical and cultural district's data collection environment within two perception intervals. Scene change similarity is calculated based on the similarity between two maps. High similarity indicates that the spatial distribution, type, and intensity of the challenges remain relatively stable; low similarity indicates that the challenge pattern has changed significantly.

[0059] Global replanning requires recalculating deployment locations and operating modes, consuming significant computing resources and time; while local parameter optimization only adjusts core operating parameters, resulting in lower costs. By determining the similarity of scene changes, global replanning is triggered only when the scene changes drastically, while local optimization is used for minor changes. This ensures optimization effectiveness while avoiding data collection interruptions and resource waste caused by frequent global replanning.

[0060] The pre-set periodic evaluation and differentiated optimization mechanism enables the data acquisition system to respond quickly in the early stages of scenario changes. The differentiated optimization strategy avoids data gaps or quality fluctuations caused by scenario changes. During local optimization, the working mode is adjusted based on the current data characteristics of the original deployment location to ensure compatibility between new and existing data. During global replanning, the deployment location is redefined based on the new challenge model to ensure the integrity of data coverage.

[0061] When the scene changes drastically, global replanning can reallocate drone and personnel data collection resources, deploying high-precision equipment to newly added high-precision demand areas and allocating redundant resources to unfinished areas. When the scene changes slightly, local parameter optimization can adjust only the equipment's operating parameters without mobilizing additional resources, achieving dynamic and accurate matching of data collection resources. Through dynamic iterative optimization, data collection quality issues caused by scene changes can be avoided in a timely manner, ensuring that the task of each data collection unit can be completed on the first attempt, reducing secondary data collection and modeling rework.

[0062] Dynamic optimization mechanisms prevent the data collection system from lingering in sensitive areas. For example, if a sudden increase in pedestrian traffic occurs around an ancient building after a dramatic change in the scene, global replanning can reschedule data collection tasks to off-peak hours or change the collection route. Conversely, if local pedestrian traffic increases during minor scene changes, local parameter optimization can improve equipment collection speed, shorten operation time, and reduce interference with the daily operation of the street, visitor experience, and the safety of cultural relics. Differentiated optimization strategies prevent project delays caused by scene changes. Rapid local optimization during minor scene changes ensures that data collection progress is not affected; and rapid redeployment through global replanning during dramatic scene changes avoids prolonged stagnation. Simultaneously, precise resource allocation and optimized cost control prevent cost overruns due to ineffective investment, keeping project progress and costs within a controllable range.

[0063] More preferably, the step of processing the multi-source three-dimensional spatial data in the semantic data processing engine to generate a digital twin model including semantic information includes: The point cloud semantic segmentation neural network deployed in the data processing engine is used to process the spatially registered and fused 3D point cloud data to classify the points in the point cloud data into predefined building component categories. For each subset of building component point clouds identified by the point cloud semantic segmentation neural network, a cross-modal feature fusion model deployed in the data processing engine is used for processing. The cross-modal feature fusion model is used to perform the following steps: Receive the point cloud subset of the building components; Receive image data blocks that spatially correspond to the building components; Output architectural style detail tags associated with the building components; The image data blocks are analyzed using a material analysis convolutional neural network deployed in the data processing engine to output the surface material classification results of the building components.

[0064] Specifically, by deploying a point cloud semantic segmentation neural network, the spatially registered and fused 3D point cloud data is classified point by point, accurately dividing the point cloud into predefined categories such as roofs, doors and windows, columns, carvings, steps, and walls. Compared to traditional manual annotation or rule-based classification methods, neural networks have stronger feature extraction capabilities, effectively distinguishing components with similar shapes but different functions, such as the brackets of ancient buildings versus ordinary load-bearing columns. This achieves automated and refined semantic segmentation of building components, laying a precise structural semantic foundation for digital twin models.

[0065] The cross-modal feature fusion model links a subset of point clouds of architectural components with corresponding image data blocks. Through multimodal data complementarity, it accurately outputs architectural style detail labels, such as Ming and Qing style carvings, European style reliefs, Lingnan arcade window frames, and the white walls and black tiles of Jiangnan water towns. This approach avoids style misjudgments caused by relying solely on point clouds or images, extending semantic information from component type to style features. This allows the digital twin model to not only reproduce form but also characterize style, enhancing the depth and richness of semantic modeling.

[0066] Material analysis convolutional neural networks perform specialized analysis on image data blocks, outputting surface material classification results for building components, such as blue bricks, wood, stone, glazed tiles, rammed earth, and paint, and can further identify material states. Compared to traditional visual judgment or single spectral analysis, this network can accurately distinguish similar materials (such as blue bricks and gray bricks, old wood and new wood) through texture features, color distribution, and local details, supplementing digital twin models with key physical attribute semantics.

[0067] Through a three-tiered processing flow, a three-dimensional semantic association is formed between component type, style details, and surface material, transforming previously fragmented multi-source three-dimensional spatial data into structured and interpretable semantic assets. This association system supports intelligent querying and analysis; for example, users can filter all Ming and Qing dynasty carved components within a block to statistically analyze the distribution range of mahogany buildings, providing data support for the research and management of historical and cultural blocks.

[0068] Architectural style detail labels and material classification results are directly linked to the core value of historical and cultural blocks. For example, by statistically analyzing the quantity and distribution of components of a specific style, the historical features of the block can be quantified; by recording the preservation status of traditional materials (such as rammed earth and old wood), the current state of cultural heritage protection can be assessed. This quantitative extraction capability makes digital twin models a tool for cultural value mining, rather than simply a visual medium.

[0069] Structured semantic information enables digital twin models to have intelligent interaction and analysis capabilities. For example, users can use natural language queries to display all stone inscription components to find Lingnan-style building facades, and the model can quickly locate the corresponding components and display relevant information. Managers can use semantic association analysis (such as the matching degree of style and material and the material differences of components from different eras) to provide data basis for the formulation of restoration plans.

[0070] Precise component classification, style tags, and material analysis can provide digital basis for the restoration of ancient buildings. For example, semantic information can be used to quickly locate components that need to be repaired, such as weathered wooden pillars. Style detail tags can be used to ensure that the restoration process is consistent with the original style. Matching restoration materials can be selected based on material classification to avoid restoring old buildings to a new state or causing style distortion, thereby improving the scientific nature and accuracy of restoration work.

[0071] More preferably, the data processing engine further includes: From an unstructured text database related to historical and cultural blocks, an entity recognition model is used to extract multiple historical entities and multiple historical relationships. The historical entities include building names and historical event names, and the historical relationships include construction time and spatial adjacency. Each extracted historical entity is associated with a subset of building component point clouds identified by the point cloud semantic segmentation neural network. The association is based on the matching degree between the spatial location information of the historical entity described in the text and the three-dimensional coordinates of the building component point cloud subset in the three-dimensional point cloud data. The associated results are stored as nodes and edges in a graph structure database, where each building component and each historical entity is stored as a node, each historical relationship is stored as an edge connecting two nodes, and each building component node includes a set of semantic attributes obtained from the semantic understanding process.

[0072] The solution of this invention uses an entity recognition model to accurately extract historical entities such as building names and historical event names from unstructured text databases such as historical documents, local chronicles, archaeological reports, and folk records, as well as historical relationships such as construction time and spatial adjacency. This breaks through the limitations of traditional data processing that only focuses on spatial form and physical attributes, and transforms scattered and difficult-to-use textual historical information into structured and associative semantic elements, so that the digital twin model not only resembles the real appearance of the historical and cultural block, but also captures its essence.

[0073] By calculating the matching degree between the spatial location descriptions of historical entities in the text and the 3D coordinates of a subset of architectural component point clouds, a one-to-one correspondence between historical entities and specific architectural components is established. This transforms architectural components from isolated spatial units into cultural carriers of historical information. A structured historical context knowledge graph is then formed: architectural components and historical entities are stored as nodes, and historical relationships are stored as edges, constructing a visualized knowledge network. This graph-based structure supports multi-dimensional relational queries and tracing.

[0074] The above methods break down the barriers between 3D spatial data (point clouds, images) and unstructured text data. Through coordinate matching and semantic association, spatial data gains historical context support, and text data gains spatial positioning anchors. For example, by combining the 3D coordinates of architectural components with the description of the south corridor of the east wing in the text, the location of historical events can be accurately pinpointed. By combining the material and style data of the components with the record of the reconstruction during the Guangxu period in the text, the authenticity of the historical renovation events can be verified, achieving mutual corroboration between spatial scenes and historical narratives.

[0075] The associative nature of graph-structured databases supports intelligent inference based on existing historical relationships, supplementing undocumented spatial relationships and enriching the completeness of historical context knowledge networks. It constructs permanently traceable digital context archives: binding historical information scattered in texts with specific architectural components and permanently storing it in a graph structure forms a three-in-one digital context archive of components, entities, and relationships, effectively preventing the loss of historical information due to document damage and memory loss. Even if architectural entities are damaged by natural disasters or human-caused destruction, their historical origins and cultural value can still be fully traced through digital twin models, providing core evidence for the regenerative protection of cultural heritage.

[0076] By using entity recognition models and coordinate matching algorithms, the automated and standardized extraction and association of historical semantic information is achieved, avoiding the subjectivity and fragmentation of manual processing (such as differences in the records of the same event in different documents). The standardized storage format of the graph-structured database facilitates the sharing of historical and cultural data across regions and projects. For example, it can link the contextual knowledge networks of different historical and cultural blocks to form a larger-scale regional historical and cultural map, promoting the standardization and large-scale digital transmission of cultural heritage.

[0077] When new historical documents are discovered or archaeological findings are published, the newly added historical entities and relationships can be quickly extracted using an entity recognition model. These are then linked to the corresponding components of the digital twin model using a coordinate matching algorithm, and the graph database is updated, enabling the dynamic improvement of historical context archives. This scalability ensures that the historical semantics of the digital twin model are always synchronized with the latest research results, maintaining the timeliness and integrity of cultural transmission.

[0078] More preferably, the step of presenting the digital twin model to the user through an interactive display platform and providing interactive functions based on the semantic information includes: Receive natural language query strings input by users through an interactive display platform; The natural language query string is sent to the data processing engine for parsing to convert the query string into a query statement for the graph structure database; Execute the query statement to obtain a list of three-dimensional component identifiers that meet the query conditions; When rendering the digital twin model, the interactive display platform is controlled to highlight the geometric parts corresponding to the list of three-dimensional component identifiers; The interactive display platform is controlled to display detailed semantic attributes and historical information text associated with the list of three-dimensional component identifiers, retrieved from the graph structure database, in the user interface panel.

[0079] Users do not need to master professional 3D model operation skills or database query syntax. They can initiate queries simply by using everyday natural language. This breaks the limitation that traditional 3D modeling is exclusive to professionals and allows users with different knowledge backgrounds, such as tourists, students, managers, and researchers, to operate the platform conveniently, greatly improving its versatility.

[0080] The data processing engine parses natural language query strings into precise query statements for graph-structured databases, accurately identifying users' core needs. For example, if a user enters "Where are the carved windows in the Ming and Qing styles?", the system can automatically parse it into three-dimensional components with the architectural style tag "Ming and Qing styles" and the component type "carved window". This avoids query deviations caused by the ambiguity of natural language and ensures consistency between the interaction intent and the search results.

[0081] Users only need to complete two steps—entering a query and viewing the results—to obtain target information, eliminating the need for complex model scaling, rotation, and filtering operations. After executing the query, the system accurately locates the target component using a list of 3D component identifiers and highlights it in the digital twin model, intuitively presenting the spatial location and shape of the target component. Users no longer need to search through complex 3D models one by one; they can instantly locate the architectural components of interest, significantly reducing spatial positioning costs. For example, after querying all stone inscriptions, all inscription components in the model are immediately highlighted, and their distribution and spatial relationships are immediately apparent.

[0082] The user interface panel synchronously presents the complete semantic attributes (component type, material, style details) and historical information (construction time, associated historical events, and documentary records) associated with the target component. This aggregates and presents the spatial, physical, and historical semantic information scattered across the graph structure database, avoiding the hassle of users having to search for information across modules and interfaces. Semantic association retrieval based on the graph structure database ensures that query results only include the target component and related information that meet the criteria, avoiding redundant display of irrelevant information.

[0083] The method in this invention integrates multiple data collection methods, including ground scanning, drone aerial photography, and handheld device detection. It can quickly cover large areas such as street rooftops and high-altitude facilities using drones, while also utilizing handheld devices to penetrate narrow alleys and other hidden corners, compensating for the blind spots of single data collection methods. Furthermore, based on a semantic digital twin model, it can provide a comprehensive and rich display of the streetscape.

[0084] Example 2 Please see Figure 3 , Figure 3 This is a schematic diagram of the structure of the system for rapid acquisition and display of three-dimensional spatial data of historical and cultural blocks disclosed in an embodiment of the present invention. Figure 3 As shown, the rapid acquisition and display system for three-dimensional spatial data of this historical and cultural district may include: Acquisition Module 21: Used to acquire multi-source three-dimensional spatial data of historical and cultural blocks through a collaborative acquisition system; Processing module 22: used to input the multi-source three-dimensional spatial data into the semantic data processing engine; in the semantic data processing engine, the multi-source three-dimensional spatial data is processed to generate a digital twin model including semantic information; Display module 23: used to present the digital twin model to users through an interactive display platform and provide interactive functions based on the semantic information.

[0085] The method in this invention integrates multiple data collection methods, including ground scanning, drone aerial photography, and handheld device detection. It can quickly cover large areas such as street rooftops and high-altitude facilities using drones, while also utilizing handheld devices to penetrate narrow alleys and other hidden corners, compensating for the blind spots of single data collection methods. Furthermore, based on a semantic digital twin model, it can provide a comprehensive and rich display of the streetscape.

[0086] Example 3 Please see Figure 4 , Figure 4 This is a schematic diagram of the structure of an electronic device disclosed in an embodiment of the present invention. The electronic device can be a computer, a server, etc. Of course, in certain cases, it can also be a mobile phone, tablet computer, monitoring terminal, or other smart device, as well as an image acquisition device with processing capabilities. Figure 4As shown, the electronic device may include: Memory 510 storing executable program code; Processor 520 coupled to memory 510; The processor 520 calls the executable program code stored in the memory 510 to execute some or all of the steps in the method for rapid acquisition and display of three-dimensional spatial data of historical and cultural blocks in Embodiment 1.

[0087] This invention discloses a computer-readable storage medium storing a computer program that enables a computer to perform some or all of the steps in the method for rapid acquisition and display of three-dimensional spatial data of historical and cultural blocks in Embodiment 1.

[0088] This invention also discloses a computer program product, wherein when the computer program product is run on a computer, the computer performs some or all of the steps in the method for rapid acquisition and display of three-dimensional spatial data of historical and cultural blocks in Embodiment 1.

[0089] This invention also discloses an application publishing platform, which is used to publish computer program products. When the computer program products are run on a computer, the computer executes some or all of the steps in the method for rapid acquisition and display of three-dimensional spatial data of historical and cultural blocks in Embodiment 1.

[0090] In various embodiments of the present invention, it should be understood that the sequence number of each process does not necessarily imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0091] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they can be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0092] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0093] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-accessible memory. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several requests to cause a computer device (which can be a personal computer, server, or network device, specifically a processor in the computer device) to execute some or all of the steps of the methods described in the various embodiments of the present invention.

[0094] In the embodiments provided by this invention, it should be understood that B corresponding to A means that B is associated with A, and B can be determined based on A. However, it should also be understood that determining B based on A does not mean determining B solely based on A; B can also be determined based on A and / or other information.

[0095] Those skilled in the art will understand that some or all of the steps in the various methods of the embodiments described can be implemented by a program instructing related hardware. This program can be stored in a computer-readable storage medium, including read-only memory (ROM), random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, disk storage, magnetic tape storage, or any other computer-readable medium capable of carrying or storing data.

[0096] The foregoing has provided a detailed description of the method, system, electronic device, and storage medium for rapid acquisition and display of three-dimensional spatial data of historical and cultural blocks disclosed in the embodiments of the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A system for rapid acquisition and display of three-dimensional spatial data of historical and cultural blocks, characterized in that, include: Acquisition module: Used to acquire multi-source 3D spatial data of historical and cultural blocks through a collaborative acquisition system; The collaborative data acquisition system includes a cluster of unmanned aerial vehicles (UAVs) and a ground-based mobile data acquisition unit. The acquisition of multi-source three-dimensional spatial data of historical and cultural blocks through a collaborative acquisition system includes: The system receives initial scan data collected from a historical and cultural district via a drone swarm, and obtains an initial real-world 3D model based on the initial scan data. After obtaining the initial real-world 3D model, the system further includes: determining the data-missing region and model quality score of the historical and cultural district based on the initial real-world 3D model; when the drone swarm's acquisition state is stable, determining the maximum acceptable area of ​​the data-missing region based on historical acquisition data; obtaining the current area of ​​the data-missing region, and determining the data integrity characterization value corresponding to the district based on the current area and the maximum acceptable area; when the drone swarm's acquisition state is not stable, determining the data integrity characterization value corresponding to the district based on the rate of change of the model quality score. Based on the initial real-world 3D model, the corresponding areas to be optimized are determined, including blind spots and detail areas. A comprehensive data acquisition strategy score is calculated, which is a weighted sum based on intermediate evaluation parameters, including data completeness index, spatial accessibility score, equipment capability matching degree, and historical and cultural value. The data acquisition strategy is determined by comparing the comprehensive data acquisition strategy score with a set strategy range, which includes an aggressive strategy, a balanced strategy, or a passive strategy. The location information of the area to be optimized is sent to the corresponding ground mobile acquisition unit, and the supplementary spatial data collected by the ground mobile acquisition unit is received. Multi-source 3D spatial data is obtained based on the initial real-scene 3D model and supplementary spatial data; Processing module: used to input the multi-source 3D spatial data into the semantic data processing engine; in the semantic data processing engine, the multi-source 3D spatial data is processed to generate a digital twin model including semantic information; Display module: Used to present the digital twin model to users through an interactive display platform and provide interactive functions based on the semantic information.

2. A method for rapid acquisition and display of three-dimensional spatial data of historical and cultural blocks, characterized in that, include: Multi-source three-dimensional spatial data of historical and cultural blocks are acquired through a collaborative acquisition system; The collaborative data acquisition system includes a cluster of unmanned aerial vehicles (UAVs) and a ground-based mobile data acquisition unit. The acquisition of multi-source three-dimensional spatial data of historical and cultural blocks through a collaborative acquisition system includes: The system receives initial scan data collected from a historical and cultural district via a drone swarm, and obtains an initial real-world 3D model based on the initial scan data. After obtaining the initial real-world 3D model, the system further includes: determining the data-missing region and model quality score of the historical and cultural district based on the initial real-world 3D model; when the drone swarm's acquisition state is stable, determining the maximum acceptable area of ​​the data-missing region based on historical acquisition data; obtaining the current area of ​​the data-missing region, and determining the data integrity characterization value corresponding to the district based on the current area and the maximum acceptable area; when the drone swarm's acquisition state is not stable, determining the data integrity characterization value corresponding to the district based on the rate of change of the model quality score. Based on the initial real-world 3D model, the corresponding areas to be optimized are determined, including blind spots and detail areas. A comprehensive data acquisition strategy score is calculated, which is a weighted sum based on intermediate evaluation parameters, including data completeness index, spatial accessibility score, equipment capability matching degree, and historical and cultural value. The data acquisition strategy is determined by comparing the comprehensive data acquisition strategy score with a set strategy range, which includes an aggressive strategy, a balanced strategy, or a passive strategy. The location information of the area to be optimized is sent to the corresponding ground mobile acquisition unit, and the supplementary spatial data collected by the ground mobile acquisition unit is received. Multi-source 3D spatial data is obtained based on the initial real-scene 3D model and supplementary spatial data; The multi-source 3D spatial data is input into a semantic data processing engine; the semantic data processing engine processes the multi-source 3D spatial data to generate a digital twin model including semantic information. The digital twin model is presented to users through an interactive display platform, and interactive functions are provided based on the semantic information.

3. The method for rapid acquisition and display of three-dimensional spatial data of historical and cultural blocks as described in claim 2, characterized in that, Before acquiring multi-source three-dimensional spatial data of historical and cultural blocks through the collaborative acquisition system, the following is also included: Based on the received geographic information data, the historical and cultural blocks are divided into multiple collection units with different spatial and collection characteristics. A corresponding acquisition feature vector is generated for each acquisition unit. The acquisition feature vector includes unit type, average width, facade complexity, and recommended acquisition method. The recommended acquisition method includes UAV acquisition or personnel acquisition. Multiple data collection units are identified as the set of data collection points that need to be served. Model drone data collection methods and personnel data collection methods as heterogeneous service points; Construct a data acquisition path model with the goal of minimizing the total data acquisition completion time. The model constraints include the service time window of each data acquisition unit, the endurance or working time constraints of each data acquisition method, and the requirement constraints of each unit for the data acquisition type. The acquisition path model is calculated by optimizing the algorithm, generating an ordered sequence of task units for each UAV acquisition method and each personnel acquisition method, and transmitting the sequence of task units to the corresponding collaborative acquisition system.

4. The method for rapid acquisition and display of three-dimensional spatial data of historical and cultural blocks as described in claim 3, characterized in that, The acquisition of multi-source three-dimensional spatial data of historical and cultural blocks through a collaborative acquisition system also includes: Based on the current acquisition progress and environmental perception data, a first challenge distribution model is constructed in the digital twin model to characterize the spatial distribution of the current acquisition challenges; the acquisition challenges include structural occlusion areas, dynamic interference areas, or areas requiring high-precision acquisition. The candidate digital performance models are virtually placed at different candidate deployment points within the digital twin model; The locations where the digital performance model does not interfere with the first challenge distribution model are determined as the first set of feasible deployment points; wherein, the performance interference refers to the inability to effectively overcome the corresponding challenge and complete the data collection task at that location.

5. The method for rapid acquisition and display of three-dimensional spatial data of historical and cultural blocks as described in claim 4, characterized in that, The method further includes: After the data acquisition system operates for the preset duration according to the first optimized working mode, a second challenge distribution model is constructed in the digital twin model based on the updated environmental perception data. Calculate the similarity of scene changes between the second challenge distribution model and the first challenge distribution model; If the similarity of the scene changes is lower than the preset change threshold, a global replanning is triggered, and the optimal deployment location and working mode of the acquisition system are re-determined based on the second challenge distribution model. If the similarity of the scene changes is not lower than the preset change threshold, local parameter optimization is triggered, and the optimized working mode of the acquisition system is re-determined based on the current data features at the first optimized deployment location.

6. The method for rapid acquisition and display of three-dimensional spatial data of historical and cultural blocks as described in claim 2, characterized in that, The process of processing the multi-source three-dimensional spatial data in the semantic data processing engine to generate a digital twin model including semantic information includes: The point cloud semantic segmentation neural network deployed in the data processing engine is used to process the spatially registered and fused 3D point cloud data to classify the points in the point cloud data into predefined building component categories. For each subset of building component point clouds identified by the point cloud semantic segmentation neural network, a cross-modal feature fusion model deployed in the data processing engine is used for processing. The cross-modal feature fusion model is used to perform the following steps: Receive the point cloud subset of the building components; Receive image data blocks that spatially correspond to the building components; Output architectural style detail tags associated with the building components; The image data blocks are analyzed using a material analysis convolutional neural network deployed in the data processing engine to output the surface material classification results of the building components.

7. The method for rapid acquisition and display of three-dimensional spatial data of historical and cultural blocks as described in claim 6, characterized in that, The data processing engine also includes: From an unstructured text database related to historical and cultural blocks, an entity recognition model is used to extract multiple historical entities and multiple historical relationships. The historical entities include building names and historical event names, and the historical relationships include construction time and spatial adjacency. Each extracted historical entity is associated with a subset of building component point clouds identified by the point cloud semantic segmentation neural network. The association is based on the matching degree between the spatial location information of the historical entity described in the text and the three-dimensional coordinates of the building component point cloud subset in the three-dimensional point cloud data. The associated results are stored as nodes and edges in a graph structure database, where each building component and each historical entity is stored as a node, each historical relationship is stored as an edge connecting two nodes, and each building component node includes a set of semantic attributes obtained from the semantic understanding process.

8. The method for rapid acquisition and display of three-dimensional spatial data of historical and cultural blocks as described in claim 7, characterized in that, The process of presenting the digital twin model to users through an interactive display platform and providing interactive functions based on the semantic information includes: Receive natural language query strings input by users through an interactive display platform; The natural language query string is sent to the data processing engine for parsing to convert the query string into a query statement for the graph structure database; Execute the query statement to obtain a list of three-dimensional component identifiers that meet the query conditions; When rendering the digital twin model, the interactive display platform is controlled to highlight the geometric parts corresponding to the list of three-dimensional component identifiers; The interactive display platform is controlled to display detailed semantic attributes and historical information text associated with the list of three-dimensional component identifiers, retrieved from the graph structure database, in the user interface panel.

Images