Land space planning and surveying and mapping engineering collaborative measurement method

Abstract

The invention relates to a land space planning and surveying and mapping engineering collaborative measurement method, which comprises the following steps: collecting multi-source surveying and mapping data of a target area, and constructing a live-action three-dimensional semantic model bearing ground feature semantics through fusion processing; key planning elements are extracted based on the model, the spatial relation of the key planning elements is quantified, and a spatial relation graph with geometric entities as nodes and the spatial relation as edges is generated; the planning management and control terms are formalized into computable logic assertions, and a machine executable planning rule base is constructed; performing collaborative traversal and logical reasoning on the spatial relation graph based on the rule base, automatically comparing and identifying violation situations, and generating a planning conflict report; and carrying out fusion mapping on the report and the three-dimensional model to generate a three-dimensional visual review layer for intuitively indicating the position, type and degree of the violation. According to the method, deep fusion of surveying and mapping data and planning rules is realized, planning conformity review can be automatically and intelligently completed, and the problems that the prior art depends on manpower, and is low in efficiency and different in standard are effectively solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of surveying and mapping engineering and spatial information technology, and in particular to a land space planning and surveying and mapping engineering collaborative measurement method. BACKGROUND

[0002] With the continuous improvement of China's land space governance system, the requirements for the scientificity of planning compilation and the accuracy of implementation supervision are increasing. Under this background, the collaborative measurement method of land space planning and surveying and mapping engineering, as a key technology aimed at breaking through the whole chain from spatial data collection to planning compliance review, has attracted widespread attention. The core goal of this method is to establish an automated technical process to deeply integrate high-precision surveying and mapping geographic information data with legally effective land space planning control rules and intelligently compare them, so as to realize rapid and objective evaluation of planning compliance of construction projects, land use, etc., and provide core technical support for planning "one map" implementation supervision.

[0003] However, the existing technical path still faces significant bottlenecks in achieving this goal. The current conventional approach relies on a semi-automated mode of "surveying and mapping providing drawings, manual rule comparison", which has obvious efficiency and accuracy ceiling. The fundamental problem is that surveying and mapping results (such as traditional two-dimensional line drawing maps or emerging real scene three-dimensional models) are essentially digital records of spatial geometry and surface features, while planning rules are abstract provisions containing complex spatial relationships and professional logic. The existing method lacks an effective mechanism for the computer to "understand" the rules and automatically and intelligently verify them with massive and complex spatial data, resulting in a serious dependence on manual experience in review work, with inherent defects such as low efficiency, inconsistent standards, and difficulty in tracing. SUMMARY

[0004] Based on this, the purpose of the present application is to provide a land space planning and surveying and mapping engineering collaborative measurement method that can deeply integrate surveying and mapping data and planning rules and achieve automated intelligent review.

[0005] The purpose of the present application is achieved by the following scheme:

[0006] In a first aspect, the present application provides a land space planning and surveying and mapping engineering collaborative measurement method, comprising the following steps:

[0007] S1: Collecting multi-source remote sensing and near-earth surveying and mapping data of the target area, fusing optical images and laser point clouds and identifying land feature semantics, performing three-dimensional geometric reconstruction and semantic attribute association on the fused data, and constructing a real scene three-dimensional semantic model;

[0008] S2: Extract the building base contour polygon and the plot red line polygon based on the real scene three-dimensional semantic model, and quantify the spatial position and topological relationship between entities by calculating the minimum distance between polygons and judging their adjacent and containing relationship, to generate a spatial relationship graph with polygons as nodes and spatial relationships as edges;

[0009] S3: Formally convert the land space planning and control clauses associated with the entity types in the real scene three-dimensional semantic model, analyze the control subjects, object objects and spatial constraint conditions in the clauses and convert them into computable logical assertions, and generate a machine executable planning rule library;

[0010] S4: Based on the planning rule library, the spatial relationship graph is iterated and logically reasoned, the entity relationship in the spatial relationship graph is compared with the constraint condition in the planning rule library, the violation situation is automatically identified, and the risk quantitative index of the conflict is calculated, and a planning conflict report is generated;

[0011] S5: The planning conflict report and the real scene three-dimensional semantic model are fused and mapped, the abstract conflict items in the report are positioned to the specific entities in the three-dimensional model and are highlighted and labeled, a visual review layer is generated, and the visual review layer is used to visually indicate the specific spatial position, type and severity of the planning violation in the three-dimensional scene.

[0012] In one embodiment, the present application provides a land space planning and surveying engineering cooperative measurement method, which specifically comprises the following steps:

[0013] S11: The obtained multi-source remote sensing and near-earth surveying data of the target area are subjected to time and space reference unification and format standardization processing, the optical images, laser point clouds and positioning and attitude data obtained by different sensors are aligned, and a time and space consistent multi-source data set is generated;

[0014] S12: The multi-source data set is subjected to joint adjustment and dense matching processing, three-dimensional point cloud reconstruction is performed based on image features and point cloud structures, and a triangular network model with topological connection relationship is constructed, and a real scene three-dimensional grid model is generated;

[0015] S13: The real scene three-dimensional grid model and the laser point cloud of the multi-source data set are subjected to deep learning semantic segmentation processing, the pre-trained instance segmentation network is used to identify and classify the ground features, the classification results are associated with the three-dimensional grid vertices, and a real scene three-dimensional semantic model carrying semantic information is generated.

[0016] In one embodiment, the present application provides a land space planning and surveying engineering cooperative measurement method, which specifically comprises the following steps:

[0017] S21: Real scene three-dimensional semantic model is subjected to ground feature element screening and two-dimensional projection processing, building and land contour are extracted based on semantic label, and three-dimensional surface geometry is projected to horizontal plane to generate initial vector set of building base contour polygon and land red line polygon;

[0018] S22: The polygons in the initial vector set are subjected to spatial relationship quantization processing, the minimum Euclidean distance between the boundaries of any two polygons is calculated, and the topological state of intersection, separation or inclusion is analyzed to generate relationship data with distance measurement and topological type;

[0019] S23: The relationship data is subjected to graph structure modeling processing, each polygon is abstracted as a node, and the measurement and type in the relationship data are connected as edge attributes to corresponding nodes to construct a spatial relationship graph with polygons as nodes and spatial relationships as edges.

[0020] In one of the embodiments, the S3 of the land space planning and surveying and mapping engineering collaborative measurement method provided by the application specifically comprises the following steps:

[0021] S31: The pre-set land space planning control clauses are subjected to structured analysis processing, the applicable objects, spatial relationship descriptions, constraint objects and specific numerical limits in the clause text are split, and a machine-readable rule element set is generated;

[0022] S32: The rule element set is subjected to logical programming conversion processing, the spatial relationship description is mapped to a formal predicate, and the numerical limit and comparison operator are combined to generate a computable logical assertion;

[0023] S33: The computable logical assertion is subjected to library management and index association processing, the assertion is stored according to the type of applicable object, and a mapping relationship with the entity semantic label in the real scene three-dimensional semantic model is established to generate a machine-executable planning rule library.

[0024] In one of the embodiments, the S4 of the land space planning and surveying and mapping engineering collaborative measurement method provided by the application specifically comprises the following steps:

[0025] S41: The spatial relationship graph is subjected to rule-based subgraph retrieval processing, all node and edge sequences that meet the subject and relationship conditions in the graph are matched in the graph based on each rule in the planning rule library as a premise, and a candidate rule application instance set is generated;

[0026] S42: The rule application instance set is subjected to constraint satisfaction verification processing, the spatial measurement attribute value of the edge in each instance is extracted, and the numerical limit condition of the corresponding rule is compared to generate a verification result list containing compliance and violation labels;

[0027] S43: Perform a comprehensive risk assessment on the violation instances in the verification results list, aggregate all rules violated in the same spatial location, calculate risk indicators based on the degree of violation deviation, scope of impact, and rule weight, and generate a planning conflict report.

[0028] In one embodiment, S43 of the collaborative measurement method for land spatial planning and surveying engineering provided by the present invention specifically includes the following steps:

[0029] S431: Extract multi-dimensional impact factors for each violation instance in the verification result list, calculate the deviation ratio between its actual spatial metric value and the specified threshold, the geometric area of ​​the violation spatial element, and the preset priority coefficient of the violated rule, and generate a quantitative impact factor.

[0030] S432: Perform weighted fusion analysis on the quantitative influencing factors. Using a pre-defined weighting model, integrate the distance deviation factor, area influence factor, and rule priority factor to generate a comprehensive risk assessment value. The formula for calculating the comprehensive risk assessment value is as follows:

[0031]

[0032] in, The calculated comprehensive risk assessment value, These are the actual spatial metrics extracted from the spatial relationship map. The standard threshold specified in the planning rules, The normalized relative deviation, The area affected by the violation This represents the total area of ​​the relevant region. The preset priority level for the violated rules, , , , These are preset model adjustment coefficients used to control the contribution ratio and growth curve of the deviation component, area component, and priority component to the total risk;

[0033] S433: Perform structured encapsulation of all violation instances and their corresponding comprehensive risk assessment values ​​in the verification result list, associate the violation element identifier, the specific rule violated, and the calculated comprehensive risk assessment value, and generate a planning conflict report.

[0034] In one embodiment, S5 of the collaborative measurement method for land spatial planning and surveying engineering provided by the present invention specifically includes the following steps:

[0035] S51: Analyze and perform three-dimensional spatial positioning processing on the planning conflict report. Based on the entity identifier associated with each conflict item in the report, find the corresponding three-dimensional geometry in the real-world three-dimensional semantic model and generate a mapping table from conflict description to three-dimensional object.

[0036] S52: Visualize and enhance the 3D scene guided by the mapping relationship table, assign different visual styles to the relevant 3D geometry according to the conflict type and risk level, and generate dynamic annotations between the non-compliant objects with spatial relationships to generate a 3D conflict scene that intuitively shows the violation situation.

[0037] S53: Perform layered output and association processing on the 3D conflict scene, save the enhanced rendering result as an independent overlapping layer, and link it with the text content of the planning conflict report to generate a visual review layer.

[0038] Secondly, this invention provides a collaborative measurement system for land spatial planning and surveying engineering, which is configured with the following modules:

[0039] The real-scene 3D semantic model construction module is used to collect multi-source remote sensing and near-ground mapping data of the target area, fuse optical images and laser point clouds and identify the semantics of ground features, perform 3D geometric reconstruction and semantic attribute association on the fused data, and construct a real-scene 3D semantic model.

[0040] The spatial relationship map generation module is used to extract the building base outline polygons and the land parcel red line polygons based on the real scene 3D semantic model. It quantifies the spatial position and topological relationship between entities by calculating the minimum distance between polygons and judging their adjacency and inclusion relationships, and generates a spatial relationship map with polygons as nodes and spatial relationships as edges.

[0041] The formal generation module for the planning rule base is used to perform formal transformation on the land space planning control clauses associated with entity types in the real-world 3D semantic model. It parses the control subjects, objects and spatial constraints in the clauses and converts them into computable logical assertions, generating a machine-executable planning rule base.

[0042] The planning conflict identification and report generation module is used to perform collaborative traversal and logical reasoning on the spatial relationship map based on the planning rule base. By comparing the entity relationships in the spatial relationship map with the constraints in the planning rule base, it automatically identifies violations and calculates the risk quantification index of the conflict, and generates a planning conflict report.

[0043] The visualization review layer generation module is used to fuse and map planning conflict reports with real-world 3D semantic models. It locates abstract conflict items in the report to specific entities in the 3D model and highlights and annotates them to generate a visualization review layer. The visualization review layer is used to intuitively indicate the specific spatial location, type, and severity of planning violations in the 3D scene.

[0044] Thirdly, this application provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement any of the above-mentioned collaborative measurement methods for land spatial planning and surveying engineering.

[0045] Fourthly, this application provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements any of the aforementioned collaborative measurement methods for land spatial planning and surveying engineering.

[0046] In summary, the collaborative measurement method for land spatial planning and surveying engineering provided in this application offers a unified, refined, and computable spatial information foundation for planning analysis by integrating multi-source surveying data and associating semantic attributes to construct a real-scene 3D semantic model. Based on this model, a spatial relationship map is generated by extracting key elements and quantifying their spatial relationships, explicitly transforming implicit spatial logic into a traversable and inferable knowledge network, thereby enabling computers to achieve a structured understanding of complex spatial relationships. A planning rule base constructed by formalizing natural language clauses into computable logical assertions provides accurate and scalable judgment criteria for automated review. Collaborative traversal and logical reasoning of the map based on the rule base can automatically and efficiently identify violations. Furthermore, a risk quantification model integrating multi-dimensional factors scientifically assesses the severity of conflicts, enabling a leap from qualitative to quantitative review conclusions and significantly improving the objectivity and refinement of the review process. By mapping abstract conflict reports back to the 3D scene to generate a visual review layer, the review results can be presented intuitively, and the problem location can be accurately traced back, greatly assisting planning decision-making and supervision. This method constructs a complete closed loop from data collection, knowledge representation, rule reasoning to result feedback, which can effectively solve the core problems of existing technologies such as reliance on manual labor, low efficiency, inconsistent standards, and difficulty in traceability.

[0047] To better understand and implement this invention, the following detailed description is provided in conjunction with the accompanying drawings. Attached Figure Description

[0048] Figure 1 A flowchart illustrating a collaborative measurement method for land spatial planning and surveying engineering provided in this application embodiment;

[0049] Figure 2A schematic diagram illustrating the process of generating a planning conflict report provided in an embodiment of this application;

[0050] Figure 3 This is a schematic diagram of the structure of a collaborative measurement system for land spatial planning and surveying engineering, provided as another embodiment of this application. Detailed Implementation

[0051] To facilitate understanding of the present invention, a more complete description will be given below with reference to the accompanying drawings. Preferred embodiments of the invention are shown in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a thorough and complete understanding of the disclosure of the invention.

[0052] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0053] In one embodiment, such as Figure 1 As shown, a collaborative measurement method for land spatial planning and surveying engineering is provided. This embodiment illustrates the method applied to a terminal. It is understood that this method can also be applied to a server, and further to a system including both a terminal and a server, and is implemented through interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0054] S1: Collect multi-source remote sensing and near-ground mapping data of the target area, fuse optical images and laser point clouds and identify the semantics of ground features, perform three-dimensional geometric reconstruction and semantic attribute association on the fused data, and construct a real-scene three-dimensional semantic model.

[0055] Specifically, the system collects multi-source remote sensing and near-ground mapping data for the target area. The remote sensing data includes high-resolution satellite imagery and UAV oblique photogrammetry, while the near-ground mapping data includes laser point clouds, ground imagery, and POS data. During the data acquisition phase, the system obtains the coordinates of reference points in the target area through GNSS static measurements, providing baseline control for subsequent data fusion. After data acquisition, the system performs registration between optical imagery and laser point clouds based on the RANSAC algorithm. It uses the 3D geometric information of the laser point clouds to correct image elevation errors and simultaneously leverages the texture information of the imagery to supplement the semantic features of the point clouds, achieving complementary advantages between the two types of data.

[0056] Preferably, the system employs a deep learning model that fuses CNN and PointNet, using registered multi-source data as input. Through feature extraction and attention mechanism optimization, it automatically classifies land cover types, covering common land cover types such as buildings, plots, roads, green spaces, water bodies, and municipal facilities. The output is point cloud and image data with semantic labels, and the fused data is standardized to a specific format, preserving core attributes such as 3D coordinates, reflection intensity, semantic labels, and image texture. After semantic recognition, the system performs 3D geometric reconstruction and semantic attribute association. A Poisson reconstruction algorithm is used to reconstruct the surface of the semantically labeled point cloud, generating a continuous 3D mesh model. Simultaneously, oblique photogrammetry is used for texture mapping to ensure the geometric integrity and texture rendering effect of the model. During semantic attribute association, the system establishes a unique identifier mapping mechanism, binding attributes such as land cover type, size parameters, and ownership information from the semantic recognition results to the geometric units of the 3D mesh model through unique identifiers, constructing a real-world 3D semantic model that supports geometric queries and semantic attribute retrieval.

[0057] S2: Based on the real-scene 3D semantic model, extract the building base outline polygons and the plot red line polygons, and quantify the spatial position and topological relationship between entities by calculating the minimum distance between polygons and judging their adjacency and inclusion relationships, and generate a spatial relationship map with polygons as nodes and spatial relationships as edges.

[0058] Specifically, based on the constructed real-scene 3D semantic model, the system extracts building base outline polygons and land parcel boundary polygons. For building base outline polygons, the system performs bottom-surface slicing on the building entities in the real-scene 3D semantic model, extracts the discrete point set of the slice boundaries, and uses the Douglas-Peucker algorithm to simplify the outline, removing redundant vertices to form closed polygon data. For land parcel boundary polygons, the system retrieves the land parcel planning boundary data of the target area from the national land spatial planning "one map" database, and unifies it to the coordinate system of the real-scene 3D semantic model through a coordinate transformation model, ensuring the spatial reference consistency of the two types of polygon data.

[0059] Furthermore, the system performs spatial relationship quantification analysis, employing a rotating caliper algorithm to calculate the shortest distance between the building base outline polygon and the plot boundary polygon, using the sign of the distance value to determine whether the polygons overlap. Simultaneously, based on topological spatial relationship theory, the system uses ray casting to determine the inclusion relationship of polygon vertices and edge intersection detection to determine adjacency relationships, clarifying the spatial relationship types between two types of polygons, including inclusion, adjacency, overlap, and separation. After spatial relationship quantification, the system constructs a spatial relationship graph. The graph adopts a directed graph structure, with extracted polygons as core nodes. Each node contains key information such as polygon identifier, corresponding entity type, geometric parameters, and semantic labels. Quantified spatial relationships are used as edges, containing attributes such as relationship type, minimum distance value, and topological association strength. The system stores the constructed spatial relationship graph in a graph database, which supports node traversal, relationship query, and complex path analysis functions.

[0060] S3: Based on the formal transformation of the land space planning control clauses associated with entity types in the real-world 3D semantic model, the control subjects, objects and spatial constraints in the clauses are analyzed and converted into computable logical assertions, generating a machine-executable planning rule base.

[0061] Specifically, the system filters land use planning control clauses associated with entity types in the real-world 3D semantic model. The filtering scope covers both mandatory and guiding clauses. Mandatory clauses include regulations related to land use control, building setback distances, and maximum floor area ratios, while guiding clauses include requirements for regional green space protection areas and spacing of municipal facilities. The system establishes a mapping relationship between control clauses and the real-world 3D semantic model through entity type tags, enabling specific types of entities to be associated with corresponding exclusive control clauses. Furthermore, the system formalizes the filtered control clauses, using first-order predicate logic for formal description, decomposing the clauses expressed in natural language into triplets consisting of the control subject, the object, and spatial constraints.

[0062] Preferably, the system can employ regular expressions and semantic analysis algorithms to extract key information from the clauses, clarify the specific types of the controlled entities, the scope of the objects, and the specific requirements of spatial constraints. The parsed results are then converted into machine-executable logical expressions, described using the OWL language to ensure logical rigor and machine readability. Regarding the rule base architecture, the system constructs a layered planning rule base, including a basic rule layer, a specialized rule layer, and a parameter configuration layer. The basic rule layer stores general spatial constraint rules, the specialized rule layer stores exclusive rules categorized by entity type, and the parameter configuration layer stores adjustable threshold parameters, supporting dynamic configuration by users according to the planning level. The system stores the planning rule base in an XML database, using a rule engine to load, match, and update rules. It also supports rule version management and traceability functions to ensure the practicality and maintainability of the planning rule base.

[0063] S4: Based on the planning rule base, the spatial relationship map is collaboratively traversed and logically reasoned. By comparing the entity relationships in the spatial relationship map with the constraints in the planning rule base, violations are automatically identified and risk quantification indicators of conflicts are calculated to generate planning conflict reports.

[0064] Specifically, the system performs collaborative traversal and logical reasoning on the spatial relationship graph based on the planning rule base. It employs a strategy combining entity type grouping and bidirectional traversal. First, entities are grouped by plot, then the building entities within each group are traversed. A depth-first search algorithm is used to traverse the nodes and edges of the spatial relationship graph to obtain spatial relationship data between entities. During reasoning, the system can use a Prolog inference engine, treating entity relationships in the spatial relationship graph as facts and logical assertions in the planning rule base as rules. Logical reasoning is completed through pattern matching to determine whether the spatial relationships between entities comply with planning rule requirements. After reasoning, the system identifies violations and quantifies risks. Violations are categorized into mandatory violations and guidance violations based on rule type. Mandatory violations include exceeding the red line and inconsistent land use, while guidance violations include non-compliance with building spacing standards.

[0065] Furthermore, the system calculates the risk value by multiplying the ratio of the violation quantification value to the rule threshold by the rule weight. The violation quantification value is the deviation between the actual value and the threshold, while the rule weight is set based on the legal validity of the clause. The system categorizes violations into different risk levels based on the risk value. Finally, the system generates a planning conflict report, which includes both structured and unstructured data. The structured data includes fields such as conflict identifier, violation entity identifier, entity type, violation type, risk value, risk level, violation quantification data, corresponding rule number, and conflict occurrence time. The unstructured data includes conflict statistics charts, typical violation case descriptions, and risk level distribution heatmaps, clearly defining the core characteristics and scope of impact of the violations.

[0066] S5: The planning conflict report is fused and mapped with the real-world 3D semantic model. The abstract conflict items in the report are located to specific entities in the 3D model and highlighted and annotated. A visual review layer is generated, which is used to intuitively indicate the specific spatial location, type and severity of planning violations in the 3D scene.

[0067] Specifically, the system establishes a fusion mapping mechanism between planning conflict reports and real-world 3D semantic models, using unique entity identifiers to link the two. During the mapping process, the system extracts the unique identifiers of the non-compliant entities from the planning conflict reports and matches them with the unique identifiers of the corresponding geometric units in the real-world 3D semantic models. This precisely binds the abstract conflict entries to concrete entities in 3D space, ensuring that the accuracy of conflict location matches the geometric accuracy of the 3D model. After mapping, the system highlights and annotates the non-compliant entities, employing a tiered rendering strategy. Rendering parameters are configured according to different risk levels, and for entities with localized violations, only the violating areas are rendered, ensuring clear presentation of the violation regions.

[0068] Preferably, the system can adopt a Billboard rendering mode. The annotation content includes information such as violation type, risk level, violation quantification data, and corresponding planning clause number. The annotation position is set above the center of the entity and supports dynamic adjustment with the 3D scene view to avoid obscuring the annotation information. After highlighting and annotation, the system generates a visual review layer. The layer uses a specific format, supports loading on mainstream 3D GIS platforms, and has interactive operation functions such as zooming, panning, and rotation. Users can click on the highlighted entity to view a detailed conflict report. Simultaneously, the system supports grouping and managing the visual review layers by violation type and risk level. Users can show or hide specific groups according to actual needs. The system provides two output formats: offline layer files and an online service interface. Offline layer files support local deployment, while the online service interface supports web access, meeting the planning review needs of different scenarios.

[0069] In summary, the collaborative measurement method for land spatial planning and surveying engineering provided in this application offers a unified, refined, and computable spatial information foundation for planning analysis by integrating multi-source surveying data and associating semantic attributes to construct a real-scene 3D semantic model. Based on this model, a spatial relationship map is generated by extracting key elements and quantifying their spatial relationships, explicitly transforming implicit spatial logic into a traversable and inferable knowledge network, thereby enabling computers to achieve a structured understanding of complex spatial relationships. A planning rule base constructed by formalizing natural language clauses into computable logical assertions provides accurate and scalable judgment criteria for automated review. Collaborative traversal and logical reasoning of the map based on the rule base can automatically and efficiently identify violations. Furthermore, a risk quantification model integrating multi-dimensional factors scientifically assesses the severity of conflicts, enabling a leap from qualitative to quantitative review conclusions and significantly improving the objectivity and refinement of the review process. By mapping abstract conflict reports back to the 3D scene to generate a visual review layer, the review results can be presented intuitively, and the problem location can be accurately traced back, greatly assisting planning decision-making and supervision. This method constructs a complete closed loop from data collection, knowledge representation, rule reasoning to result feedback, which can effectively solve the core problems of existing technologies such as reliance on manual labor, low efficiency, inconsistent standards, and difficulty in traceability.

[0070] In one embodiment, S1 of the collaborative measurement method for land spatial planning and surveying engineering provided by the present invention specifically includes the following steps:

[0071] S11: Perform spatiotemporal benchmark unification and format standardization processing on the acquired multi-source remote sensing and near-ground mapping data of the target area, align the optical images, laser point clouds and positioning and attitude data acquired by different sensors, and generate a spatiotemporally consistent multi-source dataset.

[0072] Specifically, the system preprocesses the acquired multi-source remote sensing and near-ground mapping data of the target area, performs integrity checks on various types of raw data, verifies the completeness of core information in optical imagery, laser point clouds, and positioning and attitude determination data, and removes data entries lacking key parameters to ensure the effectiveness of subsequent processing. In the spatiotemporal reference unification stage, the system determines a unified coordinate reference and time reference. The coordinate reference adopts the national legal spatial reference, and the time reference adopts Coordinated Universal Time (UTC). The system performs coordinate transformation processing on data acquired from different sensors, aligning the coordinate systems of various data types to the preset reference through a spatial reference transformation model. This transformation process requires matrix transformation to complete the coordinate mapping based on the differences between the original coordinate parameters of the data source and the parameters of the preset reference, eliminating spatial offsets caused by coordinate system differences between different sensors.

[0073] In terms of time synchronization, the system extracts the original timestamps of various data types, analyzes the time sampling frequency and timestamp accuracy of different data sources, calibrates the timestamps based on the time synchronization algorithm, and compensates for differences in time sampling intervals through interpolation operations to ensure that data at the same observation time correspond to the same time identifier, achieving consistent data alignment in the time dimension. In the format standardization process, the system parses the storage formats of various raw data types, clarifies the data structure characteristics and field definition rules of different formats, and performs format conversion according to a preset unified data standard, adjusting the storage structure, field definitions, and encoding methods of optical images, laser point clouds, and positioning and attitude determination data to a unified and standardized form.

[0074] Furthermore, the system fully preserves the core attribute information of various data types during the format conversion process, including texture information of optical images, 3D geometric information of laser point clouds, and attitude and position parameters of positioning and orientation data. Simultaneously, it generates a data format conversion log, recording information such as the format type and field mapping relationships before and after conversion, facilitating subsequent traceability. Through spatiotemporal benchmark unification and format standardization, the system integrates multi-source data with different formats and benchmarks into a spatiotemporally consistent multi-source dataset. This dataset meets the data compatibility and consistency requirements of subsequent joint adjustment, dense matching, and 3D reconstruction, laying a data foundation for the smooth progress of the entire technical process.

[0075] S12: Perform joint adjustment and dense matching on multi-source datasets, reconstruct 3D point clouds based on image features and point cloud structure, construct a triangular mesh model with topological connectivity, and generate a real-world 3D mesh model.

[0076] Specifically, the system takes a spatiotemporally consistent multi-source dataset as input and performs joint adjustment and dense matching processing. During joint adjustment, the system integrates image point observation data from optical images, 3D coordinate data from laser point clouds, and attitude and position observation data from positioning and orientation data, constructing multi-source observation equations based on the observation characteristics of each data type. The system assigns appropriate weights to different types of observation data, with the weight allocation determined based on sensor observation accuracy and data reliability, ensuring that high-reliability data plays a greater role in the adjustment calculation. The system solves the observation equations using the least squares principle, iteratively optimizing the deviation between observed and theoretical values, eliminating systematic and random errors in various data types, and further optimizing the sensor's exterior orientation elements and the 3D coordinates of the densified points, achieving high-precision fusion of multi-source data. After joint adjustment, the system performs dense matching processing, conducting bidirectional matching based on image features and point cloud structure.

[0077] Furthermore, the system extracts feature points and feature descriptors from optical images, identifies stable feature points in the images using a feature detection algorithm, and generates feature vectors for these feature points using a feature description algorithm. Simultaneously, it extracts structural feature information from the laser point cloud, including the local curvature of the point cloud and the distribution characteristics of neighboring points. The system compares the image feature vectors with the point cloud structural features using a feature matching algorithm to establish correspondences between images and between images and point clouds. Then, a random sampling consensus algorithm is used to filter the matching results, eliminating incorrect matches and improving matching accuracy. Based on the established reliable correspondences, the system uses a dense matching strategy to generate a large number of dense 3D points, supplementing the density of the initial point cloud and enhancing the detail representation capability of the 3D data.

[0078] In the 3D point cloud reconstruction stage, the system performs denoising and redundancy removal on the densely matched 3D point sets. Based on the statistical characteristics of neighboring point sets, it identifies and removes incorrectly matched and redundant points, retaining only valid 3D point data. The system then constructs a triangular mesh model with topological connections based on the valid 3D point sets. Using a triangulation algorithm, the 3D point sets are connected into continuous triangular facets according to spatial distribution rules, forming a complete 3D surface structure. During the triangular mesh model construction process, the system verifies the connection relationships of the triangular facets to ensure topological consistency and avoid problems such as facet overlap and gaps. Regions with topological anomalies are re-triangulated and adjusted. After construction, the system performs surface smoothing on the triangular mesh model, optimizing the surface continuity by adjusting the vertex positions of the triangular facets, ultimately generating a realistic 3D mesh model that accurately reflects the 3D geometry of the target area.

[0079] S13: Deep learning semantic segmentation is performed on the real-scene 3D mesh model and the laser point cloud of the multi-source dataset. The pre-trained instance segmentation network is used to identify and classify ground features. The classification results are associated with the 3D mesh vertices to generate a real-scene 3D semantic model carrying semantic information.

[0080] Specifically, the system uses a real-world 3D mesh model and laser point clouds from multi-source datasets as joint inputs to perform deep learning semantic segmentation. The geometric information of the real-world 3D mesh model is converted into point cloud format and fused with the laser point clouds from the multi-source datasets, integrating the 3D coordinates, reflection intensity, and geometric topology information of both types of data. The system performs scale normalization on the fused data, adjusting the data feature values ​​to a preset range to meet the input requirements of the deep learning model and generating a standardized input data format. The system calls a pre-trained instance segmentation network, which has been trained on a large number of ground feature samples and has the ability to identify and classify different types of ground features. The instance segmentation network extracts features from the fused standardized data, performs feature mapping on the data through convolutional layers to generate shallow feature maps, and then performs dimensionality reduction and feature aggregation on the shallow features through pooling layers. Deep semantic features are gradually extracted through multi-layer convolutional operations, and a skip connection mechanism is used to fuse feature information from different levels to enhance the completeness of feature representation. Based on the extracted feature information, the instance segmentation network realizes instance-level recognition and classification of land features. The system divides land features into preset categories, covering common land feature types such as buildings, plots, roads, green spaces, water bodies, and municipal facilities, and outputs the classification label and spatial range information of each land feature instance.

[0081] After classification, the system associates the classification results with the vertices of the real-world 3D mesh model, establishing a mapping relationship between classification labels and mesh vertices. The system traverses all vertices of the real-world 3D mesh model, matching the spatial coordinates of the vertices with the spatial extent of the feature instances to determine the feature classification label corresponding to each vertex. This label is then stored in the vertex attributes, ensuring that each geometric unit of the 3D mesh model carries corresponding semantic information. The system performs consistency checks on the associated 3D mesh model. By traversing the classification labels of adjacent vertices, it detects any abrupt changes in classification labels, re-matches and verifies areas with inconsistent labels, corrects erroneous associations, and ensures accurate association between classification labels and mesh vertices, avoiding mismatches between semantic information and geometric units. Through the above semantic segmentation, classification result association, and consistency checks, the system generates a real-world 3D semantic model carrying semantic information. This model possesses both accurate 3D geometric form and clear feature semantic attributes.

[0082] In one embodiment, S2 of the collaborative measurement method for land spatial planning and surveying engineering provided by the present invention specifically includes the following steps:

[0083] S21: Perform feature selection and 2D projection processing on the real-scene 3D semantic model, extract building and plot outlines based on semantic tags, and project their 3D surface geometry onto the horizontal plane to generate an initial vector set of building base outline polygons and plot redline polygons.

[0084] Specifically, the system uses the generated real-world 3D semantic model as the processing object, reads the semantic label information of all ground features in the model, establishes the correspondence between semantic labels and target ground feature types, and clarifies that the ground feature categories to be extracted are buildings and land parcels. The system traverses all geometric units of the 3D semantic model, matches and filters them based on the semantic labels associated with each geometric unit, retains geometric units whose labels match the target categories, and removes geometric units that do not correspond to the target ground features, forming a 3D geometric subset containing only buildings and land parcels. After filtering, the system verifies the validity of the filtered 3D geometric subset, checks the topological integrity of the geometric units, and removes geometric units with topological defects to ensure the accuracy of subsequent projection processing. In the 2D projection processing stage, the system determines the projection reference plane as the horizontal plane and establishes the projection mapping relationship between the 3D geometry and the 2D plane.

[0085] Furthermore, the system maps the 3D surface geometry of buildings and plots onto a horizontal plane according to projection rules, maintaining the relative positional relationship of the geometric contours and preserving their topological structure during projection. After projection, the system extracts discrete points from the generated 2D contours to obtain the boundary discrete point set. A contour fitting algorithm then connects these discrete point sets into closed polygons, generating an initial vector set of building base contour polygons and plot boundary polygons. When generating the initial vector set, the system assigns a unique identifier to each polygon while retaining core attributes such as semantic tags and 3D geometric association information of the corresponding features, ensuring the traceability of the initial vector set and the real-world 3D semantic model. After the initial vector set is generated, the system performs a closure check on each polygon in the set, ensuring that the polygon boundaries are free of breaks and overlaps. Polygons that do not meet the requirements are repaired, resulting in a structurally complete and attribute-rich initial vector set.

[0086] S22: Perform spatial relationship quantization on the polygons in the initial vector set, calculate the minimum Euclidean distance between any two polygon boundaries, analyze their intersecting, disjoint, or contained topological states, and generate relational data with accompanying distance metrics and topological types.

[0087] Specifically, the system takes an initial vector set as input and performs polygon spatial relationship quantification. It preprocesses the polygons in the initial vector set, optimizing the boundary point set of the polygons using a contour simplification algorithm to remove redundant boundary points. This reduces data complexity and improves subsequent computational efficiency without altering the core contour shape. After preprocessing, the system calculates the minimum Euclidean distance between any two polygon boundaries. It iterates through all polygon combinations in the initial vector set. For each pair of polygons, it first extracts their boundary point sets. Then, using a distance calculation algorithm, it iterates through all point pairs in the two boundary point sets, calculating the straight-line distance between the point pairs. The minimum value among all point pair distances is selected as the minimum Euclidean distance for that polygon boundary set. During the distance calculation process, the system records the polygon identifier corresponding to each distance value to ensure a clear correspondence between distance data and polygons. In the topology analysis phase, the system determines the topological relationship between any two polygons based on the polygon boundary point sets and the minimum Euclidean distance result.

[0088] Preferably, the system first checks whether the boundary point sets of two sets of polygons have overlapping or intersecting coordinates. If so, it is determined to be an intersecting topological state. If no intersection is found, the system then uses the minimum Euclidean distance result for further judgment. If the distance is greater than zero, it is determined to be a disjoint topological state. If all boundary points of one polygon are located inside the other polygon and there is no intersection, it is determined to be an inclusive topological state. The system performs multiple checks on the topological state of each set of polygons to avoid errors caused by a single judgment logic. After quantization, the system integrates the identification information, minimum Euclidean distance value, and topological state type of each set of polygons to generate relational data with distance measurement and topological type. The relational data is stored in a structured format, and each record contains core fields such as the unique identifiers of the two polygons, the minimum Euclidean distance data, and the topological state identifier.

[0089] S23: Perform graph structure modeling on the relational data, abstract each polygon into a node, and connect the corresponding nodes with the metrics and types in the relational data as edge attributes to construct a spatial relational graph with polygons as nodes and spatial relations as edges.

[0090] Specifically, the system uses the generated relational data as a foundation to perform graph structure modeling and construct a spatial relational graph. Before modeling, the system preprocesses the relational data, first cleaning the data to remove duplicate records and invalid data, ensuring the uniqueness and validity of the relational data; then classifying and organizing the relational data, grouping it according to topological state type or polygon type, providing ordered data for the modeling process. In the graph structure modeling stage, the system first defines nodes, abstracting each polygon in the initial vector set into a node in the graph structure. Each node is associated with corresponding attribute information, including the polygon's unique identifier, the corresponding semantic label of the ground feature, the polygon's geometric parameter information, and the association identifier in the 3D semantic model, giving the node a complete attribute description.

[0091] Furthermore, the system defines edge attributes, using the minimum Euclidean distance metric and topology type from the relational data as the core attributes of the edges, establishing a mapping relationship between edge attributes and nodes. The system traverses all relational data records, creating an edge connecting two polygonal nodes in each record, and assigning the corresponding minimum Euclidean distance value and topology type to the edge attributes, achieving precise association between edges and nodes. During edge construction, the system ensures that each relational data corresponds to only one edge, avoiding duplicate edge creation. After the graph structure is constructed, the system stores the constructed graph structure using graph database storage specifications, forming a spatial relational graph. The system performs integrity verification on the graph, checking whether each polygonal node has been included in the graph and whether each relational data has been converted into a corresponding edge, ensuring full coverage of nodes and edges. Simultaneously, the system verifies the consistency of the association between node attributes and edge attributes, avoiding attribute mismatches. The spatial relational graph uses polygons as nodes and spatial relationships as edges, fully preserving the distance metrics and topological association information between polygons, supporting subsequent collaborative traversal and logical reasoning based on a planning rule base.

[0092] In one embodiment, S3 of the collaborative measurement method for land spatial planning and surveying engineering provided by the present invention specifically includes the following steps:

[0093] S31: Perform structured parsing on the pre-defined land spatial planning control clauses, break down the applicable objects, spatial relationship descriptions, constrained objects and specific numerical restrictions in the clause text, and generate a set of machine-readable rule elements.

[0094] Specifically, the system uses pre-defined land spatial planning control clauses as its processing objects. It preprocesses these clauses, filtering those related to entity types in the real-world 3D semantic model, eliminating irrelevant clauses, verifying the completeness of the filtered clauses, marking and temporarily storing clauses with incomplete expressions or logical breaks, and only parsing complete clauses. In the structured parsing stage, the system can use text segmentation algorithms to break down the clause text into independent lexical units, and use keyword extraction algorithms to identify core words corresponding to applicable objects, spatial relationship descriptions, constrained objects, and specific numerical restrictions. The system establishes element mapping rules, matching core words with four major element categories. For ambiguous expressions or polysemous words, it uses semantic similarity calculations based on the clause's context to assist in analysis. Preferably, the system uses a formula to verify element relevance:

[0095]

[0096] in, The logical consistency coefficient. Here, is the matching value for the corresponding relationship of the i-th element group, and n is the total number of element groups under the same clause. This formula is used to quantify the logical matching degree between the applicable object and the constrained object, and between the spatial relationship description and the numerical restriction. After the verification is passed, the system integrates the four major categories of elements corresponding to each clause, assigns a unique identifier, records the original information of the clause, and stores the generated rule element set in a structured format, including core fields such as clause identifier and applicable object.

[0097] S32: Perform logical programming transformation on the set of rule elements, map spatial relationship descriptions into formal predicates, combine numerical constraints with comparison operators, and generate computable logical assertions.

[0098] Specifically, the system takes a set of rule elements as input and performs logical programming transformation. Before transformation, the system validates the set of rule elements, checking for missing element fields, incorrect associations, and other issues, and removes invalid records. In the logical programming transformation stage, the system first processes the spatial relationship description elements, establishes mapping rules, calls a preset predicate dictionary, and calculates the matching degree using a formula. The calculation expression is:

[0099]

[0100] in, The degree of matching between spatial relations and formal predicates, This refers to the set of features describing spatial relationships within the clauses. The system first sets features for formal predicates to ensure they accurately reflect the spatial constraint intent. Then, it extracts specific numerical constraint information and combines it with constraint logic matching and comparison operators to form quantified constraint expressions. The system integrates formal predicates, quantified constraint expressions, applicable objects, and constraint object information, constructing computable logical assertions following logical programming syntax specifications. The system performs syntax validation on the generated logical assertions, checking for and correcting syntax errors and logical contradictions. It establishes a mapping between assertions and rule element sets, records original element information for easy traceability and modification, and transforms unstructured element information into structured, computable logical assertions.

[0101] S33: Perform library-based management and index association processing on computable logical assertions, classify and store assertions according to applicable object types, establish mapping relationships with entity semantic labels in the real-world 3D semantic model, and generate a machine-executable planning rule library.

[0102] Specifically, the system uses computable logical assertions as its core, conducting library management and index association processing. In the library management stage, the system categorizes logical assertions according to their applicable object types, forming subsets of categorized assertions. A hierarchical storage architecture is constructed, allocating independent storage space to each subset, establishing a category directory to record management information, and assigning a unique identifier to each assertion to ensure retrieval. In the index association stage, the system extracts the applicable object types of the categorized assertion subsets and matches them with entity semantic tags in the real-world 3D semantic model. The matching accuracy is calculated using a formula:

[0103]

[0104] in, For semantic matching coefficients, For the feature vector of the applicable object type, A one-to-one correspondence is established for the feature vectors of entity semantic labels. The system constructs a semantic label index, associating and storing assertion unique identifiers with semantic labels to form an index table, supporting fast querying. The system performs consistency verification on the mapping relationship, integrates the categorized stored assertions, category directories, index tables, and mapping relationships, and generates a machine-executable planning rule base.

[0105] In one embodiment, such as Figure 2 As shown, S4 of the collaborative measurement method for land spatial planning and surveying engineering provided by the present invention specifically includes the following steps:

[0106] S41: Perform rule-based subgraph retrieval processing on the spatial relationship graph. Based on each rule in the planning rule base, match all node and edge sequences in the graph that meet its subject and relationship conditions to generate a set of candidate rule application instances.

[0107] Specifically, the system uses a spatial relationship graph and a planning rule base as processing objects. It decomposes the premises of each rule in the planning rule base, extracts core search conditions such as the subject type and spatial relationship type corresponding to the rule, and establishes a mapping relationship between the search conditions and the graph nodes and edge attributes. Preferably, the system can use a subgraph isomorphic matching algorithm to traverse the spatial relationship graph, using the premises of each rule as a search template to match node and edge sequences in the graph that match the subject type and spatial relationship. To quantify the matching accuracy, the system calculates the subgraph matching degree using a formula:

[0108]

[0109] in, For subgraph matching degree, To the number of successfully matched nodes, The target number of nodes required by the rule. The number of edges that were successfully matched. This represents the target number of edges required by the rule. The system sets a matching degree threshold, filters out node and edge sequences that meet the threshold, and integrates them into a candidate set of rule application instances. Each instance contains core content such as rule identifier, matching node information, and edge attribute data.

[0110] S42: Perform constraint satisfaction verification on the set of rule application instances, extract the spatial metric attribute value of the edge in each instance, compare it with the numerical constraint conditions of the corresponding rule, and generate a verification result list containing compliance and violation markers.

[0111] Specifically, the system takes a set of rule application instances as input and performs constraint satisfaction verification. Before verification, the system checks the attribute integrity of each instance in the instance set, verifying whether the spatial metric attribute values ​​of the edges are complete, and removing instances with missing attributes. In the verification phase, the system extracts the spatial metric attribute values ​​of the edges in each instance, associates them with the corresponding planning rules, and obtains the numerical constraints in the rules. The system then uses a formula to determine constraint satisfaction:

[0112]

[0113] in, To constrain compliance (1 indicates compliance, 0 indicates violation). The actual spatial metric attribute value of the edges in the instance. The system sets numerical limit ranges for the rules. It iterates through all rule application instances, performs constraint satisfaction checks on each, and adds compliance or violation tags to each instance. After verification, the system integrates the verification results of all instances, sorts them by rule type and instance identifier, and generates a verification result list. The list includes fields such as instance identifier, rule identifier, spatial metric attribute value, numerical limit conditions, and compliance / violation tags.

[0114] S43: Perform a comprehensive risk assessment on the violation instances in the verification results list, aggregate all rules violated in the same spatial location, calculate risk indicators based on the degree of violation deviation, scope of impact, and rule weight, and generate a planning conflict report.

[0115] Specifically, the steps for generating the planning conflict report in this application are as follows:

[0116] S431: Extract multi-dimensional impact factors for each violation instance in the verification result list, calculate the deviation ratio between its actual spatial metric value and the specified threshold, the geometric area of ​​the violation spatial element, and the preset priority coefficient of the violated rule, and generate a quantitative impact factor.

[0117] Specifically, the system performs attribute correlation verification on violation instances. By traversing the core data fields of the instance, it checks the completeness of the actual spatial measurement value, the corresponding planning rule standard threshold, and the geometric information of the violation spatial elements, eliminating violation instances with missing data or incorrect correlations. In the extraction phase, the system locates the node and edge attributes corresponding to the violation instance from the spatial relationship graph, extracts the actual spatial measurement value, and retrieves the prescribed threshold of the rule violated by the instance from the planning rule base. The deviation ratio is then calculated using a formula:

[0118]

[0119] in, This represents the deviation ratio. This is the actual spatial measurement value. This is the standard threshold specified in the planning rules. The system calculates the geometric area based on the contour vertex coordinates of the non-compliant spatial elements in the real-world 3D semantic model using the following formula:

[0120]

[0121] in, The planar coordinates of the vertices of the non-compliant element's outline. The system calculates the total number of contour vertices and determines the total area of ​​the relevant region where the violation element is located through spatial retrieval. The system retrieves the preset priority coefficient of the corresponding rule from the planning rule base; this coefficient is directly related to the rule's legal validity level. The system integrates the deviation ratio, the geometric area of ​​the violation spatial element, and the preset priority coefficient of the rule in a unified data format, and clarifies the quantitative expression of each factor through field mapping, generating standardized quantitative impact factors.

[0122] S432: Perform weighted fusion analysis on the quantitative influencing factors. Through a preset weight model, integrate the distance deviation factor, area influence factor and rule priority factor to generate a comprehensive risk assessment value.

[0123] Specifically, the system verifies the validity of the quantified influencing factors and normalizes each factor using a formula:

[0124]

[0125] in, These are the normalized factor values. These are the original factor values. This represents the minimum value boundary for this type of factor. This sets the maximum value boundary for this type of factor, ensuring that factors of different dimensions are on the same quantitative scale. The system calls a preset weighting model, incorporating the normalized distance deviation factor, area influence factor, and rule priority factor into the fusion calculation. The formula for calculating the comprehensive risk assessment value is as follows:

[0126]

[0127] in, The calculated comprehensive risk assessment value, These are the actual spatial metrics extracted from the spatial relationship map. The standard threshold specified in the planning rules, The normalized relative deviation, The area affected by the violation This represents the total area of ​​the relevant region. The preset priority level for the violated rules, , , , These are preset model adjustment coefficients used to control the contribution ratio and growth curve of the deviation component, area component, and priority component to the total risk; Define the percentage contribution of relative deviation to total risk. Adjusting the growth rate of the deviation component, Set the weighting coefficients for the area influence factor. The system determines the impact weight of rule priority levels. It substitutes the normalized, quantified impact factors into the formula to complete the calculation, verifies the calculation results to ensure the values ​​are within a preset reasonable range, and finally obtains the comprehensive risk assessment value for each violation instance.

[0128] S433: Perform structured encapsulation of all violation instances and their corresponding comprehensive risk assessment values ​​in the verification result list, associate the violation element identifier, the specific rule violated, and the calculated comprehensive risk assessment value, and generate a planning conflict report.

[0129] Specifically, the system uses all violation instances and their corresponding comprehensive risk assessment values ​​as the core to perform structured encapsulation processing. Before encapsulation, the system performs integrity checks on the associated data of each violation instance and calculates the data integrity coefficient using a formula:

[0130]

[0131] in, The integrity coefficient, This represents the actual number of complete core fields. To determine the preset total number of core fields, instance data with completeness coefficients that do not meet the preset requirements are removed. In the structured encapsulation stage, the system establishes a multi-dimensional data association mapping mechanism, binding the violation element identifier of the violation instance with the geometric unit identifier of the corresponding entity in the real-world 3D semantic model. This associates the spatial coordinate range of the violation element, the number of the violation rule, the specific content of the rule, the category to which the rule belongs, and the level of legal effect. At the same time, it associates the calculated comprehensive risk assessment value with the original data of each quantitative impact factor.

[0132] Furthermore, the system integrates the aforementioned related information into standardized data entries according to a preset data structure specification. Each entry includes fields such as violation element identifier, spatial coordinate range, rule details, quantitative impact factor, comprehensive risk assessment value, and calculation process log. The system sorts and organizes all standardized data entries according to their comprehensive risk assessment values, and then performs a secondary verification to check the logical consistency between entries, generating a structured planning conflict report. The report is stored in a common data format, adapting to the interface requirements for subsequent fusion and mapping with the real-world 3D semantic model.

[0133] In one embodiment, S5 of the collaborative measurement method for land spatial planning and surveying engineering provided by the present invention specifically includes the following steps:

[0134] S51: Analyze and perform 3D spatial positioning processing on the planning conflict report. Based on the entity identifier associated with each conflict item in the report, find the corresponding 3D geometry in the real-world 3D semantic model and generate a mapping table from conflict description to 3D object.

[0135] Specifically, the system processes planning conflict reports, performing analysis and 3D spatial positioning. In the analysis phase, the system traverses all conflict entries in the report according to a preset data structure, extracting core information such as entity identifiers, conflict types, and risk levels associated with each entry, and ensuring the standardization of the extracted information through data format validation. In the 3D spatial positioning phase, the system matches the extracted entity identifiers with entity identifiers in the real-world 3D semantic model, calculating the positioning matching degree using a formula:

[0136]

[0137] in, To determine the matching degree, The number of entity identifier features that are successfully matched. This represents the total number of features for entity identification. Based on the matching results, the system filters out the corresponding 3D geometries, verifies the spatial integrity and topological validity of the geometries, and ensures the accuracy of the positioning results. After positioning, the system establishes an association between conflict descriptions and 3D objects, clarifying the unique mapping path between each conflict entry and the 3D geometry. It integrates conflict entry information, 3D geometry identifiers, spatial coordinate ranges, and other data according to a unified field specification to generate a mapping table. This table supports bidirectional queries, enabling both the positioning of 3D objects through conflict descriptions and the retrieval of associated conflict information through 3D objects, providing accurate associated data support for subsequent visualization enhancement processing.

[0138] S52: Visualize and enhance the 3D scene guided by the mapping relationship table. Assign different visual styles to the relevant 3D geometries according to the conflict type and risk level, and generate dynamic annotations between the non-compliant objects with spatial relationships to generate a 3D conflict scene that intuitively shows the non-compliance situation.

[0139] Specifically, the system uses a mapping table as a guide to perform visualization enhancement processing on the 3D scene. Before processing, the system reads the conflict type and risk level information from the mapping table and establishes a correspondence rule between conflict type, risk level, and visual style. The system determines the visual style parameters of the 3D geometry using a formula, which is:

[0140]

[0141] in, For comprehensive parameters of visual style, The basic visual parameters corresponding to the conflict type, Enhanced visual parameters corresponding to risk levels, and The visual parameter adjustment coefficients control the contribution of conflict type and risk level to the visual style. Based on the visual style parameters, the system assigns corresponding display attributes to relevant 3D geometries, achieving visual differentiation of different conflict scenarios. For violations involving spatial relationships, the system extracts their spatial correlation data and generates dynamic annotations containing core conflict information. The annotation content is linked to the spatial position of the 3D geometry, automatically optimizing the display angle as the scene's perspective adjusts. The system performs consistency verification on the visualized 3D scene, checking the accuracy of the visual style and the completeness of the dynamic annotations to ensure the intuitive presentation of violations and generate a 3D conflict scene.

[0142] S53: Perform layered output and association processing on the 3D conflict scene, save the enhanced rendering result as an independent overlapping layer, and link it with the text content of the planning conflict report to generate a visual review layer.

[0143] Specifically, the system performs layered output and correlation processing on the 3D conflict scene. In the layered output stage, the system processes the enhanced rendered 3D conflict scene into layers according to data type and display priority, encapsulating the rendering results of the non-compliant geometry and dynamic annotation information into independent data units, integrating them into independent overlapping layers. Preferably, the system uses a standardized format to store the overlapping layers, ensuring layer compatibility and loadability. In the correlation processing stage, the system calculates the correlation strength between the layer and the planning conflict report using a formula:

[0144]

[0145] in, This represents the correlation strength value. For the feature matching value of the conflict report text field, For spatial feature matching values ​​of layer data, and The correlation coefficient is used to balance the matching weights of textual and spatial features. Based on the correlation strength value, the system establishes bidirectional links between overlapping layers and the text content of the planning conflict report, enabling quick navigation between violation objects in the layer and conflict entries in the report. The system performs integrity checks on layer data and related links to ensure no data loss or broken links, generating a visual review layer that accurately displays violation information in a 3D scene and supports linked queries with the conflict report.

[0146] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0147] Based on the same inventive concept, this application also provides a land spatial planning and surveying engineering collaborative measurement system for implementing the aforementioned collaborative measurement method for land spatial planning and surveying engineering. The solution provided by this system is similar to the implementation scheme described in the above method. Therefore, the specific limitations of one or more embodiments of the land spatial planning and surveying engineering collaborative measurement system provided below can be found in the limitations of the land spatial planning and surveying engineering collaborative measurement method described above, and will not be repeated here.

[0148] Preferably, such as Figure 3 As shown, the present invention provides a collaborative measurement system 600 for land spatial planning and surveying engineering, which is configured with the following modules:

[0149] The real-scene 3D semantic model construction module 610 is used to collect multi-source remote sensing and near-ground mapping data of the target area, fuse optical images and laser point clouds and identify the semantics of ground features, perform 3D geometric reconstruction and semantic attribute association on the fused data, and construct a real-scene 3D semantic model.

[0150] The spatial relationship map generation module 620 is used to extract the building base outline polygons and the land parcel red line polygons based on the real scene 3D semantic model, and to quantify the spatial position and topological relationship between entities by calculating the minimum distance between polygons and judging their adjacency and inclusion relationships, and to generate a spatial relationship map with polygons as nodes and spatial relationships as edges.

[0151] The planning rule base formal generation module 630 is used to perform formal transformation on the land space planning control clauses associated with entity types in the real-world 3D semantic model, parse the control subjects, objects and spatial constraints in the clauses and convert them into computable logical assertions, generating a machine-executable planning rule base.

[0152] The planning conflict identification and report generation module 640 is used to perform collaborative traversal and logical reasoning on the spatial relationship map based on the planning rule base. By comparing the entity relationships in the spatial relationship map with the constraints in the planning rule base, it automatically identifies violations and calculates the risk quantification index of the conflict, and generates a planning conflict report.

[0153] The visualization review layer generation module 650 is used to fuse and map the planning conflict report with the real-world 3D semantic model. It locates the abstract conflict items in the report to specific entities in the 3D model and highlights and annotates them to generate a visualization review layer. The visualization review layer is used to intuitively indicate the specific spatial location, type and severity of planning violations in the 3D scene.

[0154] Preferably, the real-scene 3D semantic model construction module 610 provided in this application is configured with the following units:

[0155] The multi-source data spatiotemporal alignment and standardization unit is used to perform spatiotemporal benchmark unification and format standardization processing on the multi-source remote sensing and near-ground mapping data of the acquired target area, and to align optical images, laser point clouds and positioning and attitude data acquired by different sensors to generate spatiotemporally consistent multi-source datasets.

[0156] The 3D mesh model reconstruction unit is used to perform joint adjustment and dense matching processing on multi-source datasets. Based on image features and point cloud structure, it performs 3D point cloud reconstruction and constructs a triangular mesh model with topological connectivity to generate a real-world 3D mesh model.

[0157] The Real-Scene 3D Semantic Association Unit is used to perform deep learning semantic segmentation on the real-scene 3D mesh model and the laser point cloud of the multi-source dataset. It identifies and classifies ground features through a pre-trained instance segmentation network, associates the classification results with the 3D mesh vertices, and generates a real-scene 3D semantic model carrying semantic information.

[0158] Preferably, the spatial relationship map generation module 620 provided in this application is configured with the following units:

[0159] The feature extraction and 2D projection unit is used to filter feature elements and perform 2D projection processing on the real scene 3D semantic model. Based on semantic tags, it extracts the outlines of buildings and plots and projects their 3D surface geometry onto the horizontal plane to generate an initial vector set of building base outline polygons and plot red line polygons.

[0160] The polygon spatial relationship quantification analysis unit is used to quantify the spatial relationships of polygons in the initial vector set, calculate the minimum Euclidean distance between any two polygon boundaries, analyze their intersecting, disjoint, or contained topological states, and generate relationship data with accompanying distance metrics and topological types.

[0161] The spatial relationship graph modeling unit is used to perform graph structure modeling on relationship data. It abstracts each polygon into a node and uses the measure and type in the relationship data as edge attributes to connect the corresponding nodes, thus constructing a spatial relationship graph with polygons as nodes and spatial relationships as edges.

[0162] Preferably, the formal generation module 630 for the planning rule base provided in this application is configured with the following units:

[0163] The planning clause structured parsing unit is used to perform structured parsing processing on the preset land space planning control clauses, breaking down the applicable objects, spatial relationship descriptions, constraint objects and specific numerical restrictions in the clause text, and generating a set of machine-readable rule elements.

[0164] The rule element logic programming transformation unit is used to perform logic programming transformation on the rule element set, mapping spatial relationship descriptions into formal predicates, and combining numerical constraints with comparison operators to generate computable logical assertions.

[0165] The planning rule base construction and indexing unit is used to manage and index computable logical assertions in a library manner. Assertions are classified and stored according to the applicable object type, and a mapping relationship is established with the entity semantic tags in the real scene 3D semantic model to generate a machine-executable planning rule base.

[0166] Preferably, the planning conflict identification and report generation module 640 provided in this application is configured with the following units:

[0167] The rule-matching subgraph retrieval unit is used to perform rule-based subgraph retrieval processing on the spatial relationship graph. Based on each rule in the planning rule base, it matches all node and edge sequences in the graph that meet its subject and relationship conditions, and generates a set of candidate rule application instances.

[0168] The instance constraint satisfaction verification unit is used to perform constraint satisfaction verification on the set of rule application instances. It extracts the spatial metric attribute value of the edge in each instance, compares it with the numerical constraint conditions of the corresponding rule, and generates a verification result list containing compliance and violation markers.

[0169] The violation instance risk assessment and report generation unit is used to conduct a comprehensive risk assessment of violation instances in the verification result list, aggregate all rules violated in the same spatial location, calculate risk indicators based on the violation deviation degree, impact scope and rule weight, and generate a planning conflict report.

[0170] Preferably, the violation instance risk assessment and report generation unit is configured with the following sub-units:

[0171] The sub-unit for quantitative extraction of impact factors of violation instances is used to extract multi-dimensional impact factors for each violation instance in the verification result list, calculate the deviation ratio between its actual spatial metric value and the specified threshold, the geometric area of ​​the violation spatial element, and the preset priority coefficient of the violated rule, and generate quantitative impact factors.

[0172] The multi-factor weighted fusion risk assessment subunit is used to perform weighted fusion analysis on quantitative influencing factors. It integrates and calculates distance deviation factor, area influence factor and rule priority factor through a preset weight model to generate a comprehensive risk assessment value.

[0173] The sub-unit for structured encapsulation and report generation of violation information is used to perform structured encapsulation of all violation instances and their corresponding comprehensive risk assessment values ​​in the verification result list, associate the violation element identifier, the specific rules violated, and the calculated comprehensive risk assessment value, and generate a planning conflict report.

[0174] Preferably, the visualization review layer generation module 650 provided in this application is configured with the following units:

[0175] The conflict resolution and 3D positioning unit is used to resolve and locate the planning conflict report in 3D space. Based on the entity identifier associated with each conflict item in the report, it searches for the corresponding 3D geometry in the real-world 3D semantic model and generates a mapping table from conflict description to 3D object.

[0176] The 3D scene visualization enhancement unit is used to enhance the visualization of 3D scenes guided by the mapping relationship table. It assigns different visual styles to relevant 3D geometries according to the conflict type and risk level, and generates dynamic annotations between non-compliant objects with spatial relationships, generating a 3D conflict scene that intuitively displays the violation situation.

[0177] The review layer generation and association unit is used to output and associate 3D conflict scenes in layers, save the enhanced rendering results as independent overlapping layers, and link them with the text content of the planning conflict report to generate a visual review layer.

[0178] In one embodiment, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the above-described collaborative measurement method for land spatial planning and surveying engineering.

[0179] In one embodiment, this application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the above-described collaborative measurement method for land spatial planning and surveying engineering.

[0180] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of those different embodiments or examples.

[0181] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The components described as separate parts may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this disclosure according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0182] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various variations or substitutions within the technical scope disclosed in this application, and these should all be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A collaborative measurement method for land spatial planning and surveying engineering, characterized in that, Includes the following steps: S1: Collect multi-source remote sensing and near-ground mapping data of the target area, fuse optical images and laser point clouds and identify the semantics of ground features, perform three-dimensional geometric reconstruction and semantic attribute association on the fused data, and construct a real-scene three-dimensional semantic model. S2: Based on the real-scene 3D semantic model, extract the building base outline polygons and the land parcel red line polygons, and quantify the spatial position and topological relationship between entities by calculating the minimum distance between polygons and judging their adjacency and inclusion relationships, and generate a spatial relationship map with polygons as nodes and spatial relationships as edges. S3: Based on the formal transformation of the land space planning control clauses associated with the entity types in the real-world 3D semantic model, the control subjects, objects and spatial constraints in the clauses are parsed and converted into computable logical assertions, generating a machine-executable planning rule base. S4: Based on the planning rule base, the spatial relationship graph is collaboratively traversed and logically reasoned. By comparing the entity relationships in the spatial relationship graph with the constraints in the planning rule base, violations are automatically identified and risk quantification indicators of conflicts are calculated to generate a planning conflict report. S5: The planning conflict report and the real-world 3D semantic model are fused and mapped. The abstract conflict items in the report are located to specific entities in the 3D model and highlighted and annotated to generate a visual review layer. The visual review layer is used to intuitively indicate the specific spatial location, type and severity of planning violations in the 3D scene.

2. The method according to claim 1, characterized in that, S1 includes: S11: Perform spatiotemporal benchmark unification and format standardization processing on the multi-source remote sensing and near-ground mapping data of the acquired target area, align the optical images, laser point clouds and positioning and attitude data acquired by different sensors, and generate a spatiotemporally consistent multi-source dataset. S12: Perform joint adjustment and dense matching processing on the multi-source dataset, reconstruct three-dimensional point cloud based on image features and point cloud structure, construct a triangular mesh model with topological connection relationship, and generate a real-scene three-dimensional mesh model. S13: Perform deep learning semantic segmentation processing on the real-scene 3D mesh model and the laser point cloud of the multi-source dataset. Identify and classify ground features through a pre-trained instance segmentation network, associate the classification results with the 3D mesh vertices, and generate a real-scene 3D semantic model carrying semantic information.

3. The method according to claim 1, characterized in that, S2 includes: S21: Perform feature selection and two-dimensional projection processing on the real-scene three-dimensional semantic model, extract building and plot outlines based on semantic tags, and project their three-dimensional surface geometry onto the horizontal plane to generate an initial vector set of building base outline polygons and plot red line polygons. S22: Perform spatial relationship quantization processing on the polygons in the initial vector set, calculate the minimum Euclidean distance between any two polygon boundaries, and analyze their intersecting, disjoint, or contained topological states to generate relationship data with accompanying distance metrics and topological types. S23: Perform graph structure modeling on the relational data, abstract each polygon as a node, and connect the corresponding nodes with the measure and type in the relational data as edge attributes to construct the spatial relational graph with polygons as nodes and spatial relations as edges.

4. The method according to claim 1, characterized in that, S3 includes: S31: Perform structured analysis on the pre-defined land space planning control clauses, break down the applicable objects, spatial relationship descriptions, constrained objects and specific numerical restrictions in the clause text, and generate a set of machine-readable rule elements; S32: Perform logical programming transformation on the set of rule elements, map the spatial relationship description into formal predicates, combine numerical constraints with comparison operators, and generate computable logical assertions; S33: Perform library-based management and index association processing on the computable logical assertions, classify and store the assertions according to the applicable object type, establish a mapping relationship with the entity semantic tags in the real-world 3D semantic model, and generate a machine-executable planning rule library.

5. The method according to claim 1, characterized in that, S4 includes: S41: Perform rule-based subgraph retrieval processing on the spatial relationship graph. Based on each rule in the planning rule base, match all nodes and edge sequences in the graph that meet its subject and relationship conditions to generate a candidate set of rule application instances. S42: Perform constraint satisfaction verification on the set of rule application instances, extract the spatial metric attribute value of the edge in each instance, compare it with the numerical constraint condition of the corresponding rule, and generate a verification result list containing compliance and violation markers; S43: Perform a comprehensive risk assessment on the violation instances in the verification result list, aggregate all rules violated in the same spatial location, calculate risk indicators based on the violation deviation degree, impact range and rule weight, and generate a planning conflict report.

6. The method according to claim 5, characterized in that, S43 includes: S431: Extract multi-dimensional impact factors for each violation instance in the verification result list, calculate the deviation ratio between its actual spatial metric value and the specified threshold, the geometric area of ​​the violation spatial element, and the preset priority coefficient of the violated rule, and generate a quantitative impact factor. S432: Perform weighted fusion analysis on the quantitative influencing factors, and integrate the distance deviation factor, area influence factor, and rule priority factor through a preset weight model to generate a comprehensive risk assessment value. The formula for calculating the comprehensive risk assessment value is as follows: in, The calculated comprehensive risk assessment value, These are the actual spatial metrics extracted from the spatial relationship map. The standard threshold specified in the planning rules, The normalized relative deviation, The area affected by the violation This represents the total area of ​​the relevant region. The preset priority level for the violated rules, , , , These are preset model adjustment coefficients used to control the contribution ratio and growth curve of the deviation component, area component, and priority component to the total risk; S433: Perform structured encapsulation of all violation instances and their corresponding comprehensive risk assessment values ​​in the verification result list, associate the violation element identifier, the specific rule violated, and the calculated comprehensive risk assessment value, and generate a planning conflict report.

7. The method according to any one of claims 1-6, characterized in that, S5 includes: S51: The planning conflict report is parsed and three-dimensional spatially located. Based on the entity identifier associated with each conflict item in the report, the corresponding three-dimensional geometry is found in the real-world three-dimensional semantic model to generate a mapping table from conflict description to three-dimensional object. S52: Perform visualization enhancement processing on the three-dimensional scene guided by the mapping relationship table, assign different visual styles to the relevant three-dimensional geometries according to the conflict type and risk level, and generate dynamic annotations between the non-compliant objects with spatial relationships to generate a three-dimensional conflict scene that intuitively displays the violation situation. S53: Perform layered output and association processing on the three-dimensional conflict scene, save the enhanced rendering result as an independent overlapping layer, and link it with the text content of the planning conflict report to generate a visual review layer.

8. A collaborative measurement system for land spatial planning and surveying engineering, characterized in that, The system includes: The real-scene 3D semantic model construction module is used to collect multi-source remote sensing and near-ground mapping data of the target area, fuse optical images and laser point clouds and identify the semantics of ground features, perform 3D geometric reconstruction and semantic attribute association on the fused data, and construct a real-scene 3D semantic model. The spatial relationship map generation module is used to extract the building base outline polygons and the land parcel red line polygons based on the real scene 3D semantic model, and to quantify the spatial position and topological relationship between entities by calculating the minimum distance between polygons and judging their adjacency and inclusion relationships, and to generate a spatial relationship map with polygons as nodes and spatial relationships as edges. The formal generation module for the planning rule base is used to perform formal transformation on the land space planning control clauses associated with the entity types in the real-world 3D semantic model, parse the control subjects, objects and spatial constraints in the clauses and convert them into computable logical assertions, generating a machine-executable planning rule base. The planning conflict identification and report generation module is used to perform collaborative traversal and logical reasoning on the spatial relationship map based on the planning rule base. By comparing the entity relationships in the spatial relationship map with the constraints in the planning rule base, it automatically identifies violations and calculates the risk quantification index of the conflict, and generates a planning conflict report. The visualization review layer generation module is used to fuse and map the planning conflict report with the real-world 3D semantic model, locate the abstract conflict items in the report to specific entities in the 3D model and highlight and annotate them to generate a visualization review layer. The visualization review layer is used to intuitively indicate the specific spatial location, type and severity of planning violations in the 3D scene.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 7.

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

Images