A city multi-source spatial data fusion method and system based on graph matching
By using a graph-based comparison method, point, line, and surface graphs are independently extracted and cleaned. Combined with multiple comparison strategies and grid topology verification, the inefficiency and chaotic topological relationships in the fusion of multi-source spatial data in cities are solved. This achieves efficient and automated data fusion and conflict correction, generating high-quality fusion results that meet the needs of smart cities.
Patent Information
- Application Number
- CN202610834181.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-06-10
- Publication Date
- 2026-07-24
AI Technical Summary
Existing methods for fusing multi-source spatial data in cities are inefficient, rely on manual comparison which is highly subjective, and are difficult to handle the problems of inconsistent geometric accuracy and chaotic topological relationships in multimodal data.
By employing a graph-based comparison method, three types of graphs—points, lines, and surfaces—are extracted and cleaned independently. Combined with multiple comparison strategies and grid topology verification, automated data fusion and conflict determination are achieved, generating high-quality fusion results.
It significantly improves the efficiency and stability of large-scale urban spatial data processing, reduces labor costs, ensures the geometric accuracy and topological compliance of the fusion results, and meets the business application requirements of smart cities.
Smart Images
Figure CN122454416A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of data fusion technology, and more specifically to a method and system for fusion of multi-source urban spatial data based on graphic comparison. Background Technology
[0002] Currently, in fields such as smart city construction and scientific urban governance, data management and sharing service departments often aggregate various spatial data from different departments or time periods (such as administrative division areas, building outlines produced by the housing and construction department, building outlines and road edges produced by the land planning department, road edges produced by the transportation department, POIs, and address census points). These data sources are heterogeneous, and the collection standards are not uniform, often resulting in the following problems: (1) Data redundancy and conflict: The same feature (such as a building or a road) is repeatedly recorded in different data sources, but there are different degrees of differences in geometric shape, geometric location and attribute information; (2) Inconsistent geometric accuracy: Due to coordinate system transformation errors and different data acquisition equipment accuracy, vector graphics of the same ground feature may be offset or deformed; (3) Chaotic topological relationships: The hierarchical relationships between point features (such as manhole covers and surveillance cameras) and area features (such as buildings and plots of land) or line features (such as roads) are unclear and difficult to automatically associate; To address the problems associated with multi-source, multi-modal spatial data, existing data fusion methods largely rely on simple attribute matching, spatial overlay, or manual visual comparison, which are inefficient and have the following drawbacks: (1) Manual comparison: inefficient, highly subjective, and difficult to handle massive amounts of data; (2) Matching based on attribute information: It relies too much on attribute information and is limited by problems such as non-standard attribute text descriptions and inconsistent coding systems, resulting in a low success rate; (3) Simple spatial overlay: It can only handle the inclusion of points and surfaces, and cannot effectively handle complex situations where the graphics of surfaces are inconsistent but actually point to the same feature (such as the difference between the planning map and the current map). At the same time, it only considers the location overlap and ignores the comprehensive similarity calculation of geometric features such as length, curvature, direction, centroid distribution, etc.
[0003] Therefore, how to provide a graph-based method for fusing urban multi-source spatial data that can solve the above problems is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention
[0004] In view of this, the present invention provides a method and system for urban multi-source spatial data fusion based on graphic comparison, which realizes high-quality fusion of multi-source data, provides reliable data support for urban digital construction and intelligent management, and has significant technical advantages and practical application value.
[0005] To achieve the above objectives, the present invention adopts the following technical solution: A method for fusing multi-source spatial data of cities based on graph comparison includes the following steps: S1: Obtain urban multi-source spatial data of the area to be processed, and preprocess the urban multi-source spatial data; S2: Extract the preprocessed urban multi-source spatial data to obtain the corresponding point graph dataset, area graph dataset, and line graph dataset; S3: Performs fusion and conflict determination on point-based, area-based, and line-based graphic datasets, and generates corresponding fusion results.
[0006] Preferably, S2 specifically includes: S21: Extract and classify the preprocessed urban multi-source spatial data to obtain the corresponding initial point graphics, initial area graphics and initial line graphics; S22: Perform data cleaning on the initial point-like pattern, the initial area-like pattern, and the initial line-like pattern respectively; S23: Construct a semantic mapping database for urban multi-source data, and use the urban multi-source data semantic mapping database to label the cleaned initial point graphics, initial area graphics, and initial line graphics to obtain the corresponding point graphics dataset, area graphics dataset, and line graphics dataset.
[0007] Preferably, S3 includes: S31: Determine multiple comparison strategies, and use multiple comparison strategies to fuse and determine conflicts of point graph datasets, area graph datasets, and line graph datasets. The comparison strategies include: comparison strategy between point graph datasets and line graph datasets, comparison strategy between point graph datasets and area graph datasets, comparison strategy between line graph datasets and line graph datasets, comparison strategy between line graph datasets and area graph datasets, and comparison strategy between area graph datasets and area graph datasets. S32: Obtain and parse the user's actual data output requirements, determine the corresponding comparison strategy based on the actual data output requirements, and generate the corresponding fusion results.
[0008] Preferred options also include: S4: Verify the fusion result and generate the final fusion result.
[0009] Preferably, S4 includes: S41: Construct a city grid topology and attribute verification database, and use the city grid topology and attribute verification database to verify and correct the final fusion result; S42: Calculate the evaluation index results based on the output of S41. When the evaluation index meets the preset threshold requirements, output the fusion result at this time.
[0010] Preferred options also include: S5: Perform format matching on the final fusion result and output it.
[0011] This invention also provides a multi-source spatial data fusion system for cities based on image comparison, comprising: The data acquisition module is used to acquire urban multi-source spatial data of the area to be processed and to preprocess the urban multi-source spatial data. The data extraction module is used to extract the preprocessed urban multi-source spatial data to obtain the corresponding point-based graphic dataset, area-based graphic dataset, and line-based graphic dataset. The data fusion module is used to fuse point-based graphic datasets, area-based graphic datasets, and line-based graphic datasets, and to determine conflicts, generating corresponding fusion results. The data verification module is used to verify the fusion result and generate the final fusion result; The output module is used to perform format matching and output the final fusion result.
[0012] As can be seen from the above technical solution, compared with the prior art, the present invention discloses a method and system for fusing urban multi-source spatial data based on image comparison, which has the following beneficial effects: 1. This invention extracts and cleans three types of graphics—points, lines, and surfaces—independently, avoiding interference caused by mixed calculations of multiple types of elements. This significantly improves the processing efficiency and stability of large-scale urban spatial data. By first determining lines and surfaces separately, and then performing step-by-step fusion of lines-surfaces and points-line-surfaces, invalid comparisons and mismatches are reduced, significantly improving the accuracy of identifying and classifying duplicate features. At the same time, it replaces manual subjective judgment, achieving full automation of deduplication, conflict correction, and attribute association, greatly reducing labor costs and processing time. 2. This invention ensures that the fusion results are geometrically accurate, topologically compliant, and attribute-complete by combining grid topology verification and attribute library verification with closed-loop correction of quantitative evaluation indicators, thus meeting the requirements of smart city business applications. The fusion results support format matching and unified output, improving data availability and reuse value. Attached Figure Description
[0013] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.
[0014] Figure 1 This invention provides an overall flowchart of a method for fusing multi-source urban spatial data based on image comparison. Figure 2 The present invention provides a structural principle block diagram of an urban multi-source spatial data fusion system. Detailed Implementation
[0015] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0016] See Figure 1 As shown in the figure, this invention discloses a method for fusing urban multi-source spatial data based on graph comparison, including the following steps: S1: Obtain urban multi-source spatial data of the area to be processed, and preprocess the urban multi-source spatial data. The urban multi-source spatial data may include spatial data from various departments and Internet maps from multiple departments and platforms. Data types include: building outlines, land parcels, road edges / centerlines, POIs, urban components, pipelines, administrative divisions and other types of data. The preprocessing process may specifically include: (1) Format conversion; (2) Spatial coordinate system one, specifically, can be the conversion of data coordinate system one to CGCS2000 coordinate system or local coordinate system; (3) Data cleaning: remove empty data, data with incorrect topological relationships, and self-repeating data; S2: Extract the preprocessed urban multi-source spatial data to obtain the corresponding point graph dataset, area graph dataset, and line graph dataset; S3: Performs fusion and conflict determination on point-based, area-based, and line-based graphic datasets, and generates corresponding fusion results.
[0017] In one specific embodiment, S2 specifically includes: S21: Extract and classify the preprocessed urban multi-source spatial data to obtain the corresponding initial point graphics, initial area graphics and initial line graphics; S22: Perform data cleaning on the initial point-like pattern, the initial area-like pattern, and the initial line-like pattern respectively; S23: Construct a semantic mapping database for urban multi-source data, and use the urban multi-source data semantic mapping database to label the cleaned initial point graphics, initial area graphics and initial line graphics to obtain the corresponding point graphics dataset (such as street lights, manhole covers), area graphics dataset (such as land parcels, building outlines, lakes) and line graphics dataset (such as road centerlines, pipelines).
[0018] In one specific embodiment, S3 includes: S31: Determine multiple comparison strategies, and use multiple comparison strategies to fuse and determine conflicts of point graph datasets, area graph datasets, and line graph datasets. The comparison strategies include: comparison strategy between point graph datasets and line graph datasets, comparison strategy between point graph datasets and area graph datasets, comparison strategy between line graph datasets and line graph datasets, comparison strategy between line graph datasets and area graph datasets, and comparison strategy between area graph datasets and area graph datasets. S32: Obtain and parse the user's actual data output requirements, determine the corresponding comparison strategy based on the actual data output requirements, and generate the corresponding fusion results.
[0019] Specifically, the working process of several comparison strategies in S31 includes: A. The specific processing steps for fusing and conflict determination of linear graph datasets include: 1. Application scenario: Comparing the same linear features from different sources, such as road centerlines, river lines, and pipelines, to identify and merge duplicate line features; 2. Judgment method: 2.1 Integration 2.1.1 Length Similarity S len Taking the lengths of two linear data points L1 and L2 as an example, the specific expression is as follows: S len =min(L1,L2) / max(L1,L2) 2.1.2 Orientation and Morphological Similarity S shape : The modified Fréchet distance or Hausdorff distance is used to measure the maximum deviation between two linear data points; 2.1.3 The specific expression for the comprehensive similarity model is: S scoreline =w1 S len +w2 S shape +w3 (1-Norm) In the formula, w1, w2, and w3 are all proportionality coefficients, and Norm is the normalized average vertical distance between L1 and L2.
[0020] 2.2 Conflict Resolution: If S scoreline If the similarity exceeds the threshold, it is considered to be the same road.
[0021] Otherwise, a geometric offset is determined. If two lines are the same feature but are offset, a new line is generated by weighted average, with the weights set according to the accuracy of the data source. For linear data that should intersect but has dangling points, extend or trim to the intersection point; For partially overlapping linear data, extract the common and differing parts of the two lines, and decide which part to keep based on the confidence score.
[0022] B. The specific process for fusing and conflict determination of planar graphic datasets includes: 1. Application scenario: Compare areal features from different sources, such as building outlines, land parcels, and land use boundaries, and carry out data deduplication and fusion to eliminate gaps and overlaps.
[0023] 2. Similarity Calculation
[0024] 2.1 Overlap Rate Calculation
[0025] Calculate the intersection area S of two planar data points A and B. inter =Area(A∩B) and the area of the union S union =Area(A∩B); 2.2 Define the overlap index IOU=S inter / S union ; 2.3 Shape Similarity Calculation Calculate the Hausdorff distance between the boundary lines of two planar features, take the median, and determine whether it is less than the shape similarity threshold.
[0026] 2.3 Conflict Determination
[0027] High overlap: If IOU > θ1 and Hausdorff distance < θ2, θ1 is the judgment threshold, which can be set between 0.7 and 1; θ2 is the distance judgment threshold, which can be set to 2. Then, it is initially determined that they are the same ground feature. Partial overlap: If 0 < IOU < θ1, it is determined that there are local differences or misalignments between the two feature contours.
[0028] 2.4 Fusion processing
[0029] High overlap: According to the data source confidence, select to retain the surface with high confidence, or take the outer envelope as the new geometry, and merge the attribute information according to the principle of "latest time first" or "high-confidence source first"; Partial overlap: Identify the overlapping area, decide whether to keep or discard according to the confidence, or retain the union (if both actually exist but represent different features).
[0030] C. The specific processing process for fusing and conflict judgment between linear graphic datasets and areal graphic datasets includes: (1) Application scenario: Check the spatial relationship between linear features and areal features, such as whether pipelines cross building surfaces (prohibited), whether road lines are completely within the road surface, whether administrative boundaries follow river lines, etc.
[0031] (2) Judgment method: Calculate the length ratio of the intersection part between the line and the surface, and the length ratio of the line outside the surface.
[0032] If the line is entirely within the surface (or the surface completely contains the line), the relationship is inclusion.
[0033] If the line is partially within the surface and partially outside the surface, it is determined as crossing or partial overlap.
[0034] (3) Conflict judgment
[0035] Crossing conflict: If the line should not cross the surface (such as pipelines crossing buildings), the line needs to be broken, the internal line segments of the surface are deleted, and the endpoints are retained at the surface boundary.
[0036] Boundary alignment: If the line should be the boundary of the surface (such as a river line should be the administrative boundary), but the line does not completely coincide with the surface boundary, the line is snapped to the surface boundary, or the surface boundary is adjusted to be consistent with the line (according to the confidence level).
[0037] D. The specific processing process for judgment between point graphic datasets and areal graphic datasets includes: 1. Point-line judgment 1.1 Application scenario: Check the topological relationship between point features (such as fire hydrants, pipeline valves, POI points) and linear features (such as water supply pipelines, road centerlines) to determine whether a point should be located on a line or which line a point should belong to.
[0038] 1.2 Judgment method: When the distance between a point and multiple lines is less than θ, the line with the smallest distance is selected as the assigned line; if the distances are close (the difference is less than 10% of the second smallest distance), attribute similarity (such as name and type) is used to assist in the determination.
[0039] 1.3 Conflict Resolution: If a point must logically be located on a line (such as a pipeline valve), then the point will be precisely captured at the perpendicular position of the nearest line.
[0040] If a point should actually exist independently (such as in an independent tree), but is mistakenly covered by a line, then the point's position remains unchanged, and the shortest distance d between the two and the line is retained in the fused data. If d ≤ θ (θ is a judgment threshold, set according to the actual situation), then the point is considered to be on the line; otherwise, the point is considered to be off-line.
[0041] E. The specific process for distinguishing between point-based and area-based graphical datasets includes: The ray casting algorithm is used to determine whether a point is inside a polygon. For points contained within multiple faces, the following multi-factor attribution decision algorithm is employed: Centroid distance: Calculates the Euclidean distance from a point to the centroid of each face. The smaller the distance, the higher the priority.
[0042] Attribute matching score: Calculates the string similarity (e.g., Levenshtein distance) between point attributes (such as name) and polygon attributes (such as name, address). The larger the value, the higher the priority.
[0043] Specifically, conflict detection combines the quantitative metrics output by the comparison engine with preset business rules to determine whether conflicts exist and their severity levels. Inputs include the comparison result set and rule base configuration; output is a set of conflict records. Each record contains the conflict type, involved entities, severity level, and recommended handling action, specifically including: 1. Rule Management Submodule: Store various conflict rules in a structured manner (such as JSON Schema or database tables), and support dynamic addition, deletion, modification and query of rules.
[0044] Rule expressions: Allow users to define composite conditions based on comparison metrics. For example: Face-to-face repetitive conflict: IoU > 0.8.
[0045] Point-to-surface ambiguity conflict: Number of points included > 1.
[0046] Lines suspected of being the same conflict: overall similarity > 0.85 and length ratio > 0.7.
[0047] 2. Conflict Type Library: Predefined conflict types: Covering common conflicts in five major comparison scenarios, such as: Linear-surface conflicts: illegal crossings, misaligned boundaries, and insufficient coverage.
[0048] Surface-to-surface conflict: high overlap, partial overlap, gaps, interlocking.
[0049] Conflict Level Labeling: Based on the degree of deviation or business impact, conflicts are labeled with levels such as "hint", "warning", and "critical".
[0050] 3. Rule Engine: Receive the output of the comparison engine and match each condition in the rule base. If a condition is met, generate a conflict record, which includes the conflict type, participating element ID, conflict index value, and suggested handling method (such as "absorb", "merge", "manual review").
[0051] It supports rule priority; when multiple rules are satisfied simultaneously, the rule with the highest priority is used for judgment.
[0052] Specifically, the implementation process of S32 also includes: Based on the conflict records output by the conflict detection library, the user's data requirements, and combined with the data source confidence level and business logic, the final fusion operation plan is automatically generated, and geometric and attribute fusion is performed. Inputs include the conflict record set, data source confidence level, and business rule configuration; outputs a fusion operation command sequence, corresponding operation parameters, and detailed fusion logs.
[0053] Furthermore, when the user's data requirement analysis results indicate that the processing results of multiple comparison strategies need to be fused, a fusion decision model is constructed, and the output results of the selected comparison strategies are input into the fusion decision model for processing to obtain the corresponding fusion result.
[0054] In one specific embodiment, it also includes: S4: Verify the fusion result and generate the final fusion result.
[0055] In one specific embodiment, S4 includes: S41: Construct a city grid topology and attribute verification database, and use the city grid topology and attribute verification database to verify and correct the final fusion result; S42: Calculate the evaluation index results based on the output of S41. When the evaluation index meets the preset threshold requirements, output the fusion result at this time.
[0056] In one specific embodiment, it also includes: S5: Perform format matching and output the final fusion result, specifically including: 1. Supports multiple output formats, including common GIS formats (Shapefile, GeoPackage), databases (PostGIS), and web services (GeoJSON, vector tiles).
[0057] Supports on-demand output: It can output fully integrated "post-governance" data, output conflict reports only, or output labeled "semi-integrated" data for manual review.
[0058] 2. Automatically generate data quality reports, including: Integrated statistics: How many elements were processed, how many conflicts were resolved, and the distribution of various types of conflicts.
[0059] Summary of changes: How much were points moved, how much were lines adjusted, and how much area of surfaces were modified.
[0060] Unresolved Conflicts List: Lists complex conflicts that require human intervention for easy review.
[0061] 3. Generate comparison maps before and after fusion (such as overlay display) to intuitively show the fusion effect.
[0062] The corrected elements are highlighted to facilitate quick user review.
[0063] 4. Supports multiple version outputs, storing the fused results as a new version while retaining the original data version, and supports version rollback and comparison.
[0064] See Figure 2 As shown, this embodiment of the invention also provides a system for fusing urban multi-source spatial data based on graph comparison as described in any of the above embodiments, comprising: The data acquisition module is used to acquire urban multi-source spatial data of the area to be processed and to preprocess the urban multi-source spatial data. The data extraction module is used to extract the preprocessed urban multi-source spatial data to obtain the corresponding point-based graphic dataset, area-based graphic dataset, and line-based graphic dataset. The data fusion module is used to fuse point-based graphic datasets, area-based graphic datasets, and line-based graphic datasets, and to determine conflicts, generating corresponding fusion results. The data verification module is used to verify the fusion result and generate the final fusion result; The output module is used to perform format matching and output the final fusion result.
[0065] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.
[0066] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A method for fusing multi-source urban spatial data based on image comparison, characterized in that, Includes the following steps: S1: Obtain urban multi-source spatial data of the area to be processed, and preprocess the urban multi-source spatial data; S2: Extract the preprocessed urban multi-source spatial data to obtain the corresponding point graph dataset, area graph dataset, and line graph dataset; S3: Performs fusion and conflict determination on point-based, area-based, and line-based graphic datasets, and generates corresponding fusion results.
2. The method for fusing multi-source urban spatial data based on image comparison according to claim 1, characterized in that, S2 specifically includes: S21: Extract and classify the preprocessed urban multi-source spatial data to obtain the corresponding initial point graphics, initial area graphics and initial line graphics; S22: Perform data cleaning on the initial point-like pattern, the initial area-like pattern, and the initial line-like pattern respectively; S23: Construct a semantic mapping database for urban multi-source data, and use the urban multi-source data semantic mapping database to label the cleaned initial point graphics, initial area graphics, and initial line graphics to obtain the corresponding point graphics dataset, area graphics dataset, and line graphics dataset.
3. The method for fusing urban multi-source spatial data based on image comparison according to claim 1, characterized in that, S3 includes: S31: Determine multiple comparison strategies, and use multiple comparison strategies to fuse and determine conflicts of point graph datasets, area graph datasets, and line graph datasets. The comparison strategies include: comparison strategy between point graph datasets and line graph datasets, comparison strategy between point graph datasets and area graph datasets, comparison strategy between line graph datasets and line graph datasets, comparison strategy between line graph datasets and area graph datasets, and comparison strategy between area graph datasets and area graph datasets. S32: Obtain and parse the user's actual data output requirements, determine the corresponding comparison strategy based on the actual data output requirements, and generate the corresponding fusion results.
4. The method for fusing urban multi-source spatial data based on image comparison according to claim 1, characterized in that, Also includes: S4: Verify the fusion result and generate the final fusion result.
5. The method for fusing urban multi-source spatial data based on image comparison according to claim 4, characterized in that, S4 includes: S41: Construct a city grid topology and attribute verification database, and use the city grid topology and attribute verification database to verify and correct the final fusion result; S42: Calculate the evaluation index results based on the output of S41. When the evaluation index meets the preset threshold requirements, output the fusion result at this time.
6. The method for fusing urban multi-source spatial data based on image comparison according to claim 4, characterized in that, Also includes: S5: Perform format matching on the final fusion result and output it.
7. A system utilizing the urban multi-source spatial data fusion method based on image comparison as described in any one of claims 1-6, characterized in that, include: The data acquisition module is used to acquire urban multi-source spatial data of the area to be processed and to preprocess the urban multi-source spatial data. The data extraction module is used to extract the preprocessed urban multi-source spatial data to obtain the corresponding point-based graphic dataset, area-based graphic dataset, and line-based graphic dataset. The data fusion module is used to fuse point-based graphic datasets, area-based graphic datasets, and line-based graphic datasets, and to determine conflicts, generating corresponding fusion results. The data verification module is used to verify the fusion result and generate the final fusion result; The output module is used to perform format matching and output the final fusion result.