Geographic entity generalization matching method based on spatial and semantic features and related device

Abstract

The invention discloses a geographic entity generalization matching method based on spatial and semantic features and a related device, and relates to the technical field of geographic entity matching. According to the method, in the matching process of a planar geographic entity object and a linear geographic entity object, shape feature segmentation is carried out on the complete geographic entity object, and the shape feature of the linear geographic entity object is obtained; according to the method, representative feature line segments after segmentation are selected as matching primitives, shape features of the geographic entity object can be accurately extracted, tiny detail features of the geographic entity object are captured, and basic data are provided for subsequent mode combination and matching; coarse matching and fine matching are carried out based on the spatial features of the matching primitives, so that feature line segments with similarity in spatial positions can be quickly found out, the matching range is narrowed, and the matching efficiency and the matching precision are improved; spatial matching results are checked and optimized on the basis of semantic features, semantic information of geographic entity objects can be fully considered, and matching accuracy and reliability are ensured.

Description

Technical Field

[0001] The present application relates to the field of geographic entity matching technology, and in particular to a geographic entity generalization matching method based on spatial and semantic features and related devices. Background Art

[0002] Geographic entity matching refers to the process of determining the same-named targets of spatial data from different sources according to a certain relationship. Geographic entity matching can not only be used for multi-source data fusion and information association sharing, but also is an important means of data inconsistency change detection, geospatial data production and update. In recent years, with the rapid development of geospatial data acquisition and preprocessing technology, the amount of geospatial data has shown an increasing trend, and the contradictions and problems between multi-phase, multi-scale, heterogeneous and heterogeneous geospatial data have become increasingly prominent. How to identify the same entity in data from different sources and then establish corresponding relationships for related entities has become a research hotspot in the field of geographic information.

[0003] In order to improve the complexity of matching types and flexibility of linear features, some scholars proposed an optimal road matching method based on the minimum cost network flow model and relaxation strategy. However, this research method uses complete entity objects as matching primitives and cannot capture the matching features of the subtle details of the entity, which makes it easy to have wrong matching and missed matching problems, affecting the overall matching accuracy. Summary of the invention

[0004] The purpose of this application is to provide a geographic entity generalization matching method and related devices based on spatial and semantic features, which can capture the tiny detail features of entity objects, reduce mismatches and missed matches, and improve the accuracy of overall matching.

[0005] To achieve the above objectives, this application provides the following solutions:

[0006] In a first aspect, the present application provides a geographic entity generalization matching method based on spatial and semantic features, and the geographic entity generalization matching method based on spatial and semantic features includes:

[0007] Acquire at least two data sets from different sources in the same study area, wherein the first data set includes the at least one first geographic entity object, and the second data set includes at least one second geographic entity object;

[0008] A first matching process is used to match a planar first geographic entity object and a planar second geographic entity object, wherein the planar first geographic entity object is a first geographic entity object of a planar geographic entity type in the first data set, and the planar second geographic entity object is a second geographic entity object of a planar geographic entity type in the second data set;

[0009] Using a second matching process to match a linear first geographic entity object and a linear second geographic entity object, wherein the linear first geographic entity object is a first geographic entity object of a linear geographic entity type in the first data set, and the linear second geographic entity object is a second geographic entity object of a linear geographic entity type in the second data set;

[0010] The first matching process includes:

[0011] According to the spatial features of the planar first geographic entity object and the spatial features of the planar second geographic entity object, the planar first geographic entity object and the planar second geographic entity object are roughly matched to obtain a first matching result, wherein the first matching result includes a plurality of planar entity matching combinations, the planar entity matching combination includes at least one planar first geographic entity object and zero planar second geographic entity objects, or the planar entity matching combination includes zero planar first geographic entity objects and at least one planar second geographic entity object, or the planar entity matching combination includes at least one planar first geographic entity object and at least one planar second geographic entity object;

[0012] Performing shape segmentation on the planar first geographic entity object and the planar second geographic entity object in the planar entity matching combination, and correspondingly obtaining a first characteristic line segment and a second characteristic line segment, wherein a closed shape formed by sequentially connecting a plurality of the first characteristic line segments is the shape of the planar first geographic entity object, and a closed shape formed by sequentially connecting a plurality of the second characteristic line segments is the shape of the planar second geographic entity object;

[0013] According to the spatial features of the first feature line segment and the spatial features of the second feature line segment in each of the entity matching combinations, the first feature line segment and the second feature line segment are precisely matched to obtain a second matching result and a third matching result, respectively, wherein the second matching result is a precise matching result of the first feature line segment, and the third matching result is a precise matching result of the second feature line segment;

[0014] According to the second matching result and the third matching result, the first matching result is corrected to obtain a planar entity space matching result;

[0015] Calculating a first semantic similarity based on the semantic features of the planar first geographic entity object and the semantic features of the planar second geographic entity object in the planar entity space matching result;

[0016] Determine whether the first semantic similarity is greater than a first threshold, and obtain a first determination result;

[0017] When the first judgment result is yes, the planar entity spatial matching result is used as the final planar entity matching result; otherwise, return to the step of "roughly matching the planar first geographic entity object and the planar second geographic entity object according to the spatial features of the planar first geographic entity object and the spatial features of the planar second geographic entity object to obtain a first matching result";

[0018] The second matching process includes:

[0019] Performing shape segmentation on all the linear first geographic entity objects and the linear second geographic entity objects, and correspondingly obtaining a plurality of third characteristic line segments and a plurality of fourth characteristic line segments;

[0020] According to the spatial features of the third feature line segment and the spatial features of the fourth feature line segment, the third feature line segment and the fourth feature line segment are roughly matched to obtain a fourth matching result and a fifth matching result, wherein the fourth matching result is a rough matching result of the third feature line segment, the fifth matching result is a rough matching result of the fourth feature line segment, the fourth matching result includes a plurality of first linear matching combinations, the first linear matching combination includes at least one of the third feature line segment and zero of the fourth feature line segment, or the first linear combination includes at least one of the third feature line segment and at least one of the fourth feature line segment, and the fifth matching result includes a plurality of second linear matching combinations, the second linear matching combination includes at least one of the fourth feature line segment and zero of the third feature line segment, or the second linear combination includes at least one of the fourth feature line segment and at least one of the third feature line segment;

[0021] According to the spatial features of the third characteristic line segment and the spatial features of the fourth characteristic line segment in the first linear matching combination, the third characteristic line segment and the fourth characteristic line segment in the first linear matching combination are precisely matched to obtain a sixth matching result;

[0022] According to the spatial features of the third characteristic line segment and the spatial features of the fourth characteristic line segment in the second linear matching combination, the third characteristic line segment and the fourth characteristic line segment in the second linear matching combination are precisely matched to obtain a seventh matching result;

[0023] Determine a linear entity matching result according to the sixth matching result and the corresponding seventh matching result;

[0024] Calculating a second semantic similarity according to the semantic features of the first linear geographic entity object and the semantic features of the second linear geographic entity object in the linear entity matching result;

[0025] Determine whether the second semantic similarity is greater than a second threshold, and obtain a second determination result;

[0026] When the second judgment result is yes, the linear entity matching result is used as the final linear entity matching result; otherwise, return to the step of "roughly matching the third feature line segment and the fourth feature line segment according to the spatial features of the third feature line segment and the spatial features of the fourth feature line segment to obtain the fourth matching result and the fifth matching result".

[0027] In a second aspect, the present application provides a computer device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the geographic entity generalization matching method based on spatial and semantic features as described in the first aspect above.

[0028] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the geographic entity generalization matching method based on spatial and semantic features described in the first aspect above.

[0029] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0030] The present application provides a generalized matching method and related device for geographic entities based on spatial and semantic features. The method uses different matching processes to match planar geographic entity objects and linear geographic entity objects. Both matching processes involve shape feature segmentation, rough matching and fine matching based on spatial features, and checking and optimizing the spatial matching results based on semantic features. By segmenting the shape features of the complete geographic entity object and selecting the representative feature line segments after segmentation as matching primitives, the shape features of the geographic entity object can be accurately extracted, the minute detail features of the geographic entity object can be captured, and basic data can be provided for subsequent pattern combination and matching; by performing rough matching and fine matching based on the spatial features of the matching primitives, feature line segments with similar spatial positions can be quickly found, the matching range can be narrowed, and the matching efficiency and matching accuracy can be improved; by checking and optimizing the spatial matching results based on semantic features, the semantic information of the geographic entity object can be fully considered to ensure the accuracy and reliability of the matching. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0032] Figure 1 This is a schematic diagram of the results of geographic entity object matching by the existing method in Example 1 of the present application;

[0033] Figure 2 A flowchart of a geographic entity generalization matching method based on spatial and semantic features provided in Example 1 of the present application;

[0034] Figure 3 This is a schematic diagram of the result of rough spatial matching of the planar lake in Example 1 of the present application;

[0035] FIG4( a ) is a schematic diagram showing that the morphological feature of the geographic entity object in Example 1 of the present application is a curvature feature;

[0036] FIG4( b ) is a schematic diagram showing that the morphological feature of the geographic entity object in Example 1 of the present application is a feature without curvature;

[0037] Figure 5 This is a schematic diagram of the feature point extraction result of the natural morphological feature object in Example 1 of the present application;

[0038] Figure 6 This is a schematic diagram of the feature point extraction result of the artificial morphological feature object in Example 1 of the present application;

[0039] Figure 7 This is a schematic diagram of the result of feature line segment extraction in Example 1 of the present application;

[0040] FIG8( a ) is a schematic diagram showing the result of a basic 1:1 mode when the characteristic line segments partially overlap in Example 1 of the present application;

[0041] FIG8( b ) is a schematic diagram of the result of the basic 1:1 mode when there is no intersection between the characteristic line segments in Example 1 of the present application;

[0042] FIG8( c ) is a schematic diagram of the result of the basic 1:1 mode when there is one intersection point between the characteristic line segments in Example 1 of the present application;

[0043] FIG8( d ) is a schematic diagram of the result of the 1:1 mode when there are two intersection points between the characteristic line segments in Example 1 of the present application;

[0044] FIG9( a ) is a schematic diagram of the result of the overlapping area when there is no intersection between the characteristic line segments in Example 1 of the present application;

[0045] FIG9( b ) is a schematic diagram of the result of the overlapping area when there is one intersection point between the characteristic line segments in Example 1 of the present application;

[0046] FIG9( c ) is a schematic diagram of the result of the overlapping area when there are two intersection points between the characteristic line segments in Example 1 of the present application;

[0047] Figure 10(a)-Figure 10(d) This is a schematic diagram of the result of the characteristic line segment belonging to the derived 1:1 mode in case II in Example 1 of the present application;

[0048] Figure 10(e)-Figure 10(f) This is a schematic diagram of the result of the characteristic line segment belonging to the derived 1:1 mode in case III in Example 1 of the present application;

[0049] Figure 10(g)-Figure 10(h) This is a schematic diagram of the result of the characteristic line segment in Example 1 of the present application belonging to the derived 1:1 mode in case IV;

[0050] FIG. 11( a ) is a schematic diagram showing the result of the characteristic line segment belonging to the 1:n mode in case V in Example 1 of the present application;

[0051] FIG11( b ) is a schematic diagram showing the result of the characteristic line segment belonging to the 1:n mode in case VI in Example 1 of the present application;

[0052] FIG. 11( c ) is a schematic diagram showing the result of the characteristic line segment belonging to the 1:n mode in case VII in Example 1 of the present application;

[0053] Figure 11(d)-Figure 11(f) This is a schematic diagram of the result when the characteristic line segment in Example 1 of the present application belongs to the 1:n mode in case VIII;

[0054] FIG. 12( a ) is a schematic diagram showing the result of the characteristic line segment belonging to the m:n mode in case IX in Example 1 of the present application;

[0055] FIG12( b ) is a schematic diagram showing the result of the characteristic line segment belonging to the m:n mode in case X in Example 1 of the present application;

[0056] Figure 12(c)-Figure 12(d) This is a schematic diagram of the result when the characteristic line segment in Example 1 of the present application belongs to the m:n mode in case XI;

[0057] FIG. 12( e ) is a schematic diagram showing the result of the characteristic line segment belonging to the m:n mode in case XII in Example 1 of the present application;

[0058] FIG. 12( f ) is a schematic diagram showing the result of the characteristic line segment belonging to the m:n mode in case XIII in Example 1 of the present application;

[0059] FIG. 12( g ) is a schematic diagram showing the result of the characteristic line segment belonging to the m:n mode in case XIV in Example 1 of the present application;

[0060] Figure 12(h)-Figure 12(i) This is a schematic diagram of the result when the characteristic line segment in Example 1 of the present application belongs to the m:n mode in case XV;

[0061] Fig.13This is a schematic diagram of the result of the feature line segment space similarity measurement in Example 1 of the present application;

[0062] FIG. 14( a ) is a result schematic diagram showing that the entity matching result is 1:0 after precise matching of the planar geographic entity object in Example 1 of the present application;

[0063] FIG. 14( b ) is a schematic diagram showing a result in which the entity matching result is 1:1 after precise matching of the planar geographic entity object in Example 1 of the present application;

[0064] FIG. 14( c ) is a schematic diagram showing the result of the entity matching result being converted from 1:N to 1:1 after precise matching of the planar geographic entity object in Example 1 of the present application;

[0065] Figure 14 (d)-(e) is a schematic diagram of the result of the entity matching result of 1:N after the precise matching of the planar geographic entity object in Example 1 of the present application;

[0066] Figure 14(f)-Figure 14(g) This is a result schematic diagram of an entity matching result of M:N after precise matching of a planar geographic entity object in Example 1 of the present application;

[0067] FIG. 15( a ) is a schematic diagram of a processing result of feature line segments of artificial morphological features having an adjacent relationship in Example 1 of the present application;

[0068] FIG. 15( b ) is a schematic diagram of a processing result of a feature line segment of an artificial morphological feature having a phase separation relationship in Example 1 of the present application;

[0069] FIG. 16( a ) is a schematic diagram showing the result of unprocessing the characteristic line segments of the natural morphological features having a phase separation relationship in Example 1 of the present application;

[0070] FIG. 16( b ) is a schematic diagram showing the result of buffering and expanding the characteristic line segments of the natural morphological features having a separation relationship in Example 1 of the present application;

[0071] FIG. 16( c ) is a schematic diagram showing the result of buffering and retracting the characteristic line segments of the natural morphological features having a separation relationship in Example 1 of the present application;

[0072] FIG. 16( d ) is a schematic diagram showing the result of overlay analysis of the characteristic line segments of the natural morphological features with adjacent relationships that have not been processed in Example 1 of the present application and the reduced overall buffer zone;

[0073] FIG16( e ) is a schematic diagram of the result of adding feature points in Example 1 of the present application;

[0074] FIG16( f ) is a schematic diagram showing the result of adding a feature line segment in Example 1 of the present application;

[0075] Fig.17 This is a schematic diagram of the result of accurate road matching in Example 1 of the present application;

[0076] FIG. 18( a ) is a schematic diagram of the result of road space matching ambiguity in Example 1 of the present application;

[0077] FIG18( b ) is a schematic diagram of the result of ambiguous spatial matching of residential areas in Example 1 of the present application;

[0078] FIG. 19( a ) is a schematic diagram of residential area data at a scale of 1:500 in Example 1 of the present application;

[0079] FIG19( b ) is a schematic diagram of residential area data at a scale of 1:10000 in Example 1 of the present application;

[0080] FIG. 20( a ) is a schematic diagram of river data of data set A in Example 1 of the present application;

[0081] Figure 20(b) is a schematic diagram of the river data of data set B in Example 1 of the present application. DETAILED DESCRIPTION

[0082] The study found that the optimal road matching method based on the minimum cost network flow model and relaxation strategy has several obvious limitations and shortcomings: First, when selecting matching primitives, the method uses complete entity objects without considering the tiny details of the entity objects. This leads to the fact that in actual operation, the method often cannot accurately capture these subtle matching features, resulting in more mismatches and missed matches. Secondly, when designing the research method, it relies too much on the spatial data type of the matching object, which makes the method less robust when dealing with different data sets. In other words, when we apply this method to other data sets, its effect may be greatly reduced, which undoubtedly limits the scope of application and application scenarios of the method. Therefore, future research needs to further improve its ability to capture detailed features while maintaining the effectiveness of the method, and enhance its robustness on different data sets.

[0083] Therefore, this embodiment provides a geographic entity generalization matching method and related devices that can capture minute detail features of entity objects, reduce mismatches and missed matches, and improve the accuracy of overall matching.

[0084] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0085] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0086] Example 1

[0087] ① Existing methods use complete entity objects as matching primitives, which cannot capture the matching features of the subtle details of the entity, resulting in problems such as wrong matching and missed matching. Figure 1 As shown in the figure, u1 in the reference dataset should be matched one-to-many with v1 and v3, but the actual match is u1, v1, v2. This is because u1 is not disconnected at the corner, resulting in it not matching v3.

[0088] ② Most existing methods are designed based on the spatial data type of the matching object. The robustness of the same method applied to different data sets is weak, which limits the applicability of the method. Figure 1 As shown, road u3 is incorrectly matched with the neighboring v4. This is because the research algorithm fails to cover this matching situation, resulting in a wrong match.

[0089] In order to overcome the above technical defects, this embodiment provides a geographic entity generalization matching method based on spatial and semantic features. The method is executed by a computer device, and can be executed by a computer device such as a terminal or a server alone, or by a terminal and a server together. In the embodiment of the present application, the method is applied to a server as an example for explanation. Figure 2 As shown, the following steps 201 to 208 are included.

[0090] in:

[0091] Step 201: Acquire at least two data sets from different sources in the same research area, wherein the first data set includes the at least one first geographic entity object, and the second data set includes at least one second geographic entity object.

[0092] Step 202: Use a first matching process to match a first planar geographic entity object and a second planar geographic entity object, wherein the first planar geographic entity object is a first geographic entity object of a planar geographic entity type in the first data set, and the second planar geographic entity object is a second geographic entity object of a planar geographic entity type in the second data set.

[0093] The first matching process includes:

[0094] Step 202-1: According to the spatial features of the first planar geographic entity object and the spatial features of the second planar geographic entity object, the first planar geographic entity object and the second planar geographic entity object are roughly matched to obtain a first matching result, wherein the first matching result includes a plurality of planar entity matching combinations, the planar entity matching combination includes at least one of the first planar geographic entity object and zero of the second planar geographic entity objects, or the planar entity matching combination includes zero of the first planar geographic entity object and at least one of the second planar geographic entity objects, or the planar entity matching combination includes at least one of the first planar geographic entity object and at least one of the second planar geographic entity objects.

[0095] This embodiment performs rough matching of planar geographic entity objects based on the minimum enclosing rectangle method. For planar geographic entity objects in different data sets, if their corresponding minimum enclosing rectangles intersect, the rough matching is successful, and the planar geographic entity objects that have successfully been roughly matched (i.e., planar entity matching combinations) are stored in the same set. The rough matching results of planar lakes in data set A and data set B are shown in Figure 1. Figure 3 And as shown in Table 1.

[0096] Table 1 Rough matching results of surface lakes in dataset A and dataset B

[0097] entity model <![CDATA[a1,b1]]> 1：1 <![CDATA[a2,b2,b3,b4]]> 1：N <![CDATA[a3,a4,b5,b6]]> M:N <![CDATA[a5]]> 1：0

[0098] Step 202-2: Perform shape segmentation on the planar first geographic entity object and the planar second geographic entity object in the planar entity matching combination, and obtain first characteristic line segments and second characteristic line segments correspondingly, wherein a closed shape formed by sequentially connecting a number of the first characteristic line segments is the shape of the planar first geographic entity object, and a closed shape formed by sequentially connecting a number of the second characteristic line segments is the shape of the planar second geographic entity object.

[0099] In the process of visual perception, people often obey certain specific composition rules of graphics and naturally pursue the structural integrity or conformity of things. For example, they distinguish whether the contour line of a geographical entity has curvature, and thus draw the following conclusion: natural features such as rivers and lakes (i.e. geographical entity objects with natural morphological characteristics) are more likely to have curvature characteristics, and artificial features such as settlements and roads (i.e. geographical entity objects with artificial morphological characteristics) often have right-angle characteristics. However, it is not rigorous to make a distinction based on such a conclusion. Most artificial objects have the morphological characteristics of natural objects. Therefore, in order to distinguish the morphological characteristics of geographic entity objects (to determine whether the geographic entity object is a geographic entity object with natural morphological characteristics or a geographic entity object with artificial morphological characteristics), this embodiment sets a distance threshold Δ after constructing the point-line topology. If the distance of the target line segment on the arc segment is less than Δ, the target line segment has a curvature feature. Otherwise, the target line segment has no curvature feature, that is, it has a right-angle feature. The feature line segment is a line segment formed by re-dividing the arc segment. As the matching primitive of this embodiment, the feature line segment has a curvature feature or a right-angle feature, which is the minimum shape factor of the geographic entity object. Multiple feature line segments are connected in sequence to form the complete shape of the geographic entity object.

[0100] The first characteristic line segment and the second characteristic line segment mentioned in step 202-2 are obtained by the same method, and the two are only different in the source of the planar geographic entity objects. Therefore, the following takes the acquisition process of the first characteristic line segment as an example for specific explanation.

[0101] (202-2-1) Constructing a geographic entity model according to the spatial relationship between the first geographic entity objects.

[0102] In this embodiment, in order to distinguish the morphological features of the first geographic entity object, the constructed geographic entity model is a point-line topological structure, and redundant points are removed from the point-line topological structure.

[0103] (202-2-2) Determine in turn whether each target line segment on the arc segment has a curvature feature, and obtain a third judgment result, wherein the arc segment is a line segment in the geographic entity model that represents the contour of the first geographic entity object, and the target line segment is a line between the first point and the last point in a preset point group, and the preset point group includes a preset number of points arranged in sequence on the arc segment, and the points include nodes and geometric points. Of two adjacent target line segments, the last point in the preset point group of the former target line segment serves as the first point in the preset point group of the latter target line segment.

[0104] As shown in Figure 4(a), the points (nodes and geometric points) on the arc segment include A1~A 19, select 5 adjacent geometric points arranged in sequence on the arc segment as the preset point group, set the distance threshold Δ, when the distance between the last geometric point and the first geometric point in the preset point group is less than the threshold Δ, the target line segment has a curvature feature, otherwise, the target line segment has no curvature feature but a right-angle feature.

[0105] FIG4( b ) shows a target line segment without curvature features.

[0106] (202-2-3) When the third judgment result is yes, the first geographic entity object is determined to be a natural morphological feature object, and the first extraction rule is used to determine the feature points; otherwise, the first geographic entity object is determined to be an artificial morphological feature object, and the second extraction rule is used to determine the feature points.

[0107] The first extraction rule includes:

[0108] (1) Selecting nodes whose number of associated arc segments is greater than or equal to a threshold value as feature points. In this embodiment, geometric points whose number of associated arc segments is greater than or equal to 3 are extracted as feature points.

[0109] (2) The vertical distance limit method is used to extract the feature points on the arc segment.

[0110] When the arc segment has curvature characteristics, such as Figure 5 As shown, the set of geometric points that constitute the arc segment is A i (i=1,2,...,7), connect A1 and A7 virtually into a straight line, and calculate other geometric points A on the arc segment i The distance d from (i=2,3...,6) to the straight line A1A7 i At this time, if the maximum distance d2 is greater than the pre-given threshold, A2 is retained. Repeat the above steps to extract the next batch of feature points between A2 and A7, and so on, until the maximum distance from the geometric point on the curve between the two points to the virtual straight line between the two points is less than the threshold.

[0111] The line between two adjacent feature points arranged in sequence is taken as the first feature line segment, and the extraction results are shown in Table 2.

[0112] Table 2. Results of feature point extraction of natural morphological features

[0113]

[0114] The second extraction rule includes:

[0115] (a) Selecting nodes whose number of associated arc segments is greater than or equal to a threshold value as feature points. In this embodiment, geometric points whose number of associated arc segments is greater than or equal to 3 are extracted as feature points.

[0116] (b) Taking the vertex of the common angle as a feature point, wherein the common angle refers to the angle formed by taking the common geometric points of two adjacent target line segments as vertices, and the shared angle is an angle less than or equal to a preset value.

[0117] (c) If there is a concave-convex structure within the outline of the first geographic entity object, then removing feature points in an independent concave-convex structure whose area is smaller than an area threshold, and removing feature points on a common edge of two adjacent concave-convex structures, wherein the independent concave-convex structure refers to a concave-convex structure that exists alone.

[0118] like Figure 6 As shown, when the arc segment has no curvature feature, geometric points with obvious angle changes are extracted as feature points.

[0119] For linear features, when the common angle is less than or equal to θ (θ can be ), the vertices of the common angles of the linear features are extracted as feature points.

[0120] For planar features, when the common angle is less than or equal to θ (θ can be ), the vertices of the common corners of the surface features are extracted as feature points. Since surface features (such as residential areas) usually contain concave-convex structures, when the area of ​​the concave-convex structure is less than a certain threshold, the geometric points in the concave-convex structure are not extracted, such as Figure 6 A9, A 10 ; When several concave-convex structures have shared edges, the geometric points within the concave-convex structures are not extracted, such as Figure 5 A2-A7 in.

[0121] The line between two adjacent feature points arranged in sequence is taken as the first feature line segment. The results are shown in Table 3.

[0122] Table 3. Feature point extraction results of artificial morphological features

[0123]

[0124] It should be noted that the following process of extracting characteristic line segments of linear geographic entity objects also adopts this method, the only difference is that the process of extracting characteristic line segments of linear geographic entity objects does not need to consider the concave-convex structure.

[0125] Therefore, this embodiment extracts characteristic line segments of planar geographic entities with curvature characteristics in the to-be-matched data sets A and B, and sets the starting characteristic points to be non-unique. The results are shown in Tables 4 and 5.

[0126] Table 4. Feature segment extraction results of dataset A

[0127] Dataset A

[0128] Geometry Point Feature Segment entity A1,A2 La1,La2 Pa1 A3,A4 La3,La4 Pa2 A5, A6, A7 La5, La6, La7 Pa3 A5, A7, A8 La7, La8, La9 Pa4

[0129] Table 5 Feature segment extraction results of dataset A

[0130] Dataset B

[0131] Geometry Point Feature Segment entity B1, B2 Lb1,Lb2 Pb1 B3,B4 Lb3,Lb4 Pb2 B5,B6 Lb5,Lb6 Pb3 B7,B8 Lb7,Lb8 Pb4 B9,B10 Lb9,Lb10 Pb5 B11,B12 Lb11,Lb12 Pb6

[0132] Step 202-3: Based on the spatial features of the first feature line segment and the spatial features of the second feature line segment in each of the entity matching combinations, the first feature line segment and the second feature line segment are precisely matched to obtain a second matching result and a third matching result, respectively, wherein the second matching result is a precise matching result of the first feature line segment, and the third matching result is a precise matching result of the second feature line segment.

[0133] In this embodiment, based on the recognition of feature line segment patterns, the precise matching of planar geographic entity objects is further achieved based on the spatial similarity measurement of the projection distance and relative offset rate between feature line segments. Therefore, on this basis, this embodiment needs to first define different matching modes for feature line segments.

[0134] (1) 1:1 Mode (basic one-to-one mode)

[0135] When the spatial relationship between feature segments from different datasets meets any of the following conditions, it is called a basic 1:1 mode.

[0136] ① Two characteristic line segments partially overlap. As shown in Figure 8(a), characteristic line segment A1A2 partially overlaps with characteristic line segment B1B2.

[0137] ② There is an overlapping area between the two feature line segments after mutual projection. As shown in Figure 8(b), there is an overlapping area when there is no intersection between A1A2 and B1B2. As shown in Figure 8(c), there is an overlapping area when there is one intersection between A1A2 and B1B2. As shown in Figure 8(d), there is an overlapping area when there are two intersections between A1A2 and B1B2.

[0138] In actual matching scenarios, the same feature often changes in shape and position in different data sets, and the edges that constitute the feature are often inconsistent. Therefore, this embodiment focuses on situation ②, and situation ② is expressed by the formula:

[0139]

[0140] Where S is the overlapping area between A1A2 and B1B2.

[0141] As matching primitives, feature line segments have a spatial relationship of separation and intersection. Under this spatial relationship, feature line segments may have an overlapping area S after being projected in relative directions. As shown in Figure 9, if the projection of a feature point to a feature line segment is on the feature line segment, the horizontal distance between the feature points is defined as the length of the overlapping area, represented by l. As shown in Figure 9(a), the average width of the overlapping area is represented by d, which is specifically the ratio of the area to the length of the overlapping area. If the projection of any feature point of a feature line segment is on another feature line segment, feature points with a shorter projection distance are selected respectively, and the horizontal distance between the feature points is the length l of the overlapping area, such as feature points A2 and B1 in Figure 9(b) and Figure 9(c).

[0142] (2) Derived 1:1 model (derived one-to-one model)

[0143] When the number of spatially similar feature line segments is greater than 2, according to the actual situation, only two of the feature line segments may be combined to form a derived 1:1 pattern. Under the premise that there is an overlapping area after mutual projection, if A i A i+1 With B j B j+1 If any of the conditions in formula (2) is satisfied, it is a derived 1:1 mode. This patent uses the values ​​of i and j to indicate whether the characteristic line segments in the same data set are continuous. When i (or j) is a continuous value, the characteristic line segments in data set A (or B) are continuous. If i (or j) is a discontinuous value, the characteristic line segments in data set A (or B) are discontinuous. Under this rule, when multiple characteristic line segments in the same data set are continuous, if there is an angle requirement, this situation is only applicable to characteristic line segments with geometric morphological characteristics, such as situation II.

[0144]

[0145] Among them, angle[B j B j+1 ] is the angle between the continuous matching feature line segments with right-angle features; according to the actual matching scene, the angle threshold α is set to 150°; l ij The overlapping area S ij The length, d ij is the average projection distance of the overlapping area; Indicates the target feature line segment A1A2 and the matching feature line segment B j B j+1The overlapping areas between them intersect, or the shared edge of the overlapping area is the target feature line segment or the matching feature line segment; Case II means that when there are multiple continuous matching feature line segments and the matching feature line segments have right-angle features, condition 1 is used for identification, and the matching feature line segment corresponding to the maximum overlapping area length is selected to match the target feature line segment to form a one-to-one matching combination; Case III means that when two matching feature line segments in the same data set are discontinuous and the longest overlapping area length is much larger than any other overlapping area length, condition 2 is used for identification, and the matching feature line segment corresponding to the maximum overlapping area length is selected to match the target feature line segment to form a one-to-one matching combination; Case IV means that when two matching feature line segments in the same data set are discontinuous and the overlapping area length is close to the longest overlapping area length, condition 3 is used for identification, and the matching feature line segment corresponding to the minimum average projection distance is selected to match the target feature line segment to form a one-to-one matching combination.

[0146] As shown in Figure 10, after the edge relationship of the actual object is abstracted, it is expressed in the form of a combination of feature line segments. According to formula (2), the matching results are A1A2 and B1B2. When the angle of the continuous feature line segments approaches 90°, situation II usually occurs in the matching of planar geographic entity objects. Otherwise, it occurs in the one-to-one matching of linear geographic entity objects. Figure 10(a)-Figure 10(d) is the combination of feature line segments matched in case II; when the feature line segments in the same data set are discontinuous, such as Figure 10(e)-Figure 10(h) , Situations III and IV may occur in densely distributed surface entity matching and one-to-one matching of linear entities. Figure 10(e)-Figure 10(f) is the combination of characteristic line segments matched in case III, Figure 10(g)-Figure 10(h) is the combination of feature line segments matched in case IV.

[0147] (3) 1:n Mode (one-to-many mode)

[0148] In the overlapping area S ij Under the premise of satisfying any one of the conditions in formula (3), A i A i+1 With B j B j+1 It is defined as a 1:n mode. If there is an angle requirement in the formula, this situation is only applicable to the matching of continuous feature line segments with right-angle features, such as cases V and VI; otherwise, it is applicable to the matching of continuous feature line segments with curvature features or the matching of non-continuous feature line segments, such as cases VII and VIII.

[0149]

[0150] In the formula, It means that in the first data set (i.e., data set A) and / or in the second data set (i.e., data set B), the distance between any geometric points of the feature line segment is less than the distance threshold Δ. When the distance between any geometric points is less than the threshold, the feature line segment has a curvature feature; case V means that among multiple continuous matching feature line segments, when the matching feature line segment has a right-angle feature and the corner angle is less than the set threshold, condition a is used for identification; case VI means that among multiple continuous matching feature line segments, when the matching feature line segment has a right-angle feature and the corner angle is greater than or equal to the set threshold, condition b is selected for identification; case VII means that among multiple continuous matching feature line segments, when the matching feature line segment has a curvature feature, condition c is selected for identification; case VIII means that among multiple discontinuous matching feature line segments, when the matching feature line segment has a curvature feature, condition d is selected for identification.

[0151] According to the actual scenario, the situations in formula (3) are listed, such as Figure 11(a)-Figure 11(c) , A1A2 is matched with B1B2, B2B3, and B3B4 in a 1:n ratio. Case V often appears in one-to-many matching of linear entities such as roads, as shown in Figure 11(a); and in order to distinguish between surface entity matching, one-to-one matching of linear entities, and one-to-many matching of linear entities, case VI often appears in stepped residential area matching and one-to-many matching of linear entities, as shown in Figure 11(b); case VII appears in one-to-many matching of entities with natural morphological characteristics such as rivers and lakes, as shown in Figure 11(c). Figure 11(d)-Figure 11(f) In the example, A1A2 is matched with B1B2 and B3B4 in a 1:n manner. Case VIII is abstracted from the one-to-many residential matching. Figure 11(d)-Figure 11(e) As shown; Case VIII is shown in Figure 11(f).

[0152] When judging the state of the 1:n and m:n modes, no matter whether the same data set contains a single or multiple feature segments, and whether the feature segments are continuous, we judge based on whether the projections of the left and right feature points in the direction of the feature segment of the same data set are both on the feature segment of another data set. For example, Figure 11(a)-Figure 11(c) , the projections of the feature points B1 and B4 at both ends of the feature line segment of data set B are both on A1A2. At this time, the 1:n mode is in the "positive" state; Figure 11(d)-Figure 11(f) , the projection of feature point A2 of A1A2 is on the feature line segment of data set B, and the projection of the leftmost feature point B1 of the feature line segment of data set B is on A1A2. At this time, the 1:n mode is in the "oblique" state.

[0153] (4) m:n Mode (many-to-many mode)

[0154] The m:n mode is essentially a combination of the 1:1 and 1:n modes. Under the premise that there is an overlapping area, when any of the conditions in formula (4) is met, A i A i+1 With B j B j+1 Perform m:n matching between them.

[0155]

[0156]

[0157] in, Refers to the angles between the feature line segments in dataset A and dataset B respectively; case IX means that when multiple continuous feature line segments in datasets A and B have right-angle features and the angle between any feature line segments in the same dataset is greater than or equal to the set threshold α, condition aa is used for identification; case X means that when multiple continuous feature line segments in both datasets have curved features, condition bb is used for identification; case XI means that when multiple continuous feature line segments in datasets A and B have right-angle features and the angle between feature line segments in the same dataset is less than the set threshold α, condition cc is used for identification; case XII indicates that the feature line segments in data set A are discontinuous, and multiple continuous feature line segments in data set B have right-angle features and the angle between any feature line segments in the same data set is greater than or equal to the angle threshold α, and condition dd is used for identification; case XIII indicates that when the feature line segments in both data sets A and B are discontinuous, condition ee is used for identification; case XIV indicates that the feature line segments in data set A are discontinuous, and multiple continuous feature line segments in data set B have curved features and the angle between any feature line segments in the same data set is greater than or equal to the set threshold α, condition ff is used for identification.

[0158] A series of situations of m:n mode are listed according to actual scenarios. As shown in Figure 12(a) and Figure 12(b), they correspond to typical m:n modes under situations IX and X, respectively. A1A2 and A2A3 are both matched with B1B2 and B2B3 in m:n mode, but the morphological features of the feature segments are different. These two situations often appear in many-to-many matching scenarios of entities; situation XI is usually a combination of 1:1 mode of feature segments, for example, A1A2 and B1B2, A2A3 and B2B3 in Figure 12(c), A1A2 and B1B2, A2A3 and B2B3, A3A4 and B3B4 in Figure 12(d). Situation XI is suitable for one-to-one matching scenarios of linear entities such as roads and residential entities with stepped structures. The feature segments in the above situations are all continuous. When the feature segments in the data set are discontinuous, situations XII (Figure 12(e)-(f)), XIII ( Figure 12(h)-Figure 12(i)) is applicable to one-to-many and many-to-many matching scenarios of surface settlements; when a river with a curved feature in one dataset is dry and the same river in another dataset is not dry, situation XIV will occur, as shown in Figure 12(g), where A1A2, A3A4 are matched with B1B2, B2B3 in an m:n manner.

[0159] On the basis of defining the matching mode of the characteristic line segment, taking the first characteristic line segment and the second characteristic line segment to obtain the second matching result as an example, the precise matching process includes:

[0160] Step A: According to different pattern recognition conditions, the matching pattern between the target feature line segment and the corresponding matching feature line segment is identified, and the feature matching combinations corresponding to the different matching patterns are determined, wherein, for each entity matching combination, the planar first geographic entity object in the entity matching combination is used as the target geographic entity object, and the corresponding planar second geographic entity object is used as the matching geographic entity object.

[0161] Step B: Calculate the maximum projection distance and offset rate between the target feature line segment and the matching feature line segment in the feature matching combination.

[0162] like Fig.13 As shown in the figure, in order to reflect the difference between the two-dimensional feature segments in the X and Y directions, the similarity between the feature segments A (target feature segment) and B (matching feature segment) is calculated based on the spatial similarity metrics in the two directions. The spatial similarity metric in the X direction is the relative offset rate between the feature segments, and the spatial similarity metric in the Y direction is the maximum projection distance between the feature segments.

[0163] ①Relative offset rate: The relative offset rate Δ is between 0 and 1. The larger the relative offset rate, the smaller the possibility of matching between feature segments. The relative offset rate calculation formula between feature segments A and B is as follows:

[0164]

[0165] Where: l is the length of the overlapping area between A and B, l A , l B are the lengths of the lines connecting the two end points of the characteristic line segments A and B. According to the formula, the relative offset rate of A and B facing each other is 0.

[0166] ② Maximum projection distance: The geometric point set on the feature line segment A is The set of geometric points on the characteristic line segment B is Let the geometric points on A draw perpendicular lines to B respectively, and select the longest perpendicular line with its foot on B instead of on the extension line of B. Its length is the maximum projection distance from A to B. Similarly, the length of the longest perpendicular line from the geometric point on B to A is the maximum projection distance from B to A. Therefore, the smaller value of the maximum projection distance from A to B and from B to A is selected as the maximum projection distance between A and B. The improved Hausdorff distance (Enhanced Hausdorff Distance, E_HD) is used for calculation. The larger the projection distance, the smaller the possibility of matching between feature segments. The calculation formula of the improved Hausdorff distance is shown in formula (6):

[0167] E_HD(A,B)=min{H(A,B),H(B,A)} (6);

[0168] Where H(A, B) and H(B, A) represent the maximum projection distances from A to B and from B to A, respectively, as shown in equations (7) and (8).

[0169]

[0170] In the formula, is a geometric point on A and B; They are At the foot of A and B; They are arrive arrive The vertical distance.

[0171] Step C: When the maximum projection distance is greater than the distance threshold, and the corresponding offset rate is greater than the offset rate threshold, it is determined that the corresponding target feature line segment and the matching feature line segment in the feature matching combination are matched successfully.

[0172] Step D: taking the successfully matched feature matching combination as the second matching result.

[0173] It should be noted that the process of obtaining the third matching result also adopts the above method.

[0174] Step 202-4: According to the second matching result and the third matching result, the first matching result is corrected to obtain a planar entity space matching result.

[0175] When the number of successfully matched feature matching combinations in the second matching result (that is, the number of successfully matched target feature line segments, or the number of target feature line segments corresponding to the matching pattern) is greater than the threshold, the corresponding planar first geographic entity object and the planar second geographic entity object are matched successfully, otherwise the matching fails.

[0176] A coarse matching result may include several different types of matching sets. For example, an M:N match includes several M:N, 1:N, 1:1, and 1:0 matches, or a 1:N match is converted into a 1:N, 1:1, and 1:0 match with a smaller number of entities. Exact matching is the process of differentiating and refining the coarse matching results.

[0177] Due to differences in geographic entity data collection methods, collection time, processing methods, etc., the same planar entity in different data sets may differ in position, angle, shape, spatial distribution, etc. This embodiment matches the same geographic entity with different types of differences through pattern recognition of feature segments. In order to improve the efficiency of accurate matching of planar geographic entity objects, on the basis of pattern definition, all feature line segment associated feature points are made to draw perpendicular lines to the feature line segments of another data set. It is required that the perpendicular line does not pass through any feature line segment, and the foot of the perpendicular falls on the feature line segment. If the feature line segments have overlapping areas, the maximum projection distance and relative offset rate between the feature line segments are calculated.

[0178] As shown in Figure 14(a), residential areas A and B are suspected to be 1:1 matches in the rough matching results. However, based on the precise matching feature line segment matching pattern recognition, B2B3 cannot be combined with A2A3 and A3A4 because the maximum projection distance exceeds the distance threshold. Therefore, the pattern matching of B2B3 with A2A3 and A3A4 is unsuccessful, and the recognition results of the remaining feature line segments A1A2, B1B2, A3A4, B3B4, A4A1, and B4B1 cannot summarize the corresponding entity contours (that is, the number of feature line segments successfully recognized by the planar entity A pattern does not meet the threshold, and the number of feature line segments successfully recognized by the planar entity B pattern does not meet the threshold). Therefore, the planar entities A and B are both 1:0 or 0:1 matching; as shown in Figure 14 (b), the maximum projection distance and offset rate between the feature line segment combinations both meet the set threshold, and the recognition results of the feature line segment pattern type of 1:1 are A1A2, B1B2, A2A3, B2B3, A3A4, B3B4, A4A1, and B4B1, so the planar entities A and B can be matched 1:1 geographical entities.

[0179] As shown in Figure 14(c), the surface entity A and B1 and B2 are suspected to be 1:N matches in the rough matching results. Since the relative offset rate or maximum projection distance between B5B6 and A1A2, B7B8 and A3A4, and B6B7 and A2A3 exceeds the threshold, B5B6, B6B7, and B7B8 are not recognized, and the unrecognized feature line segments all belong to the surface entity B1. Therefore, all feature line segments associated with the surface entity polygon B1 are removed, and the removed feature lines are restored. Segment (i.e., multiple area-shaped geographic entity objects of the same dataset in the rough matching entity combination The feature line segments are removed when processing the feature line segments. This processing process will be mentioned below ), re-screen and match the feature line segments, and the identified feature line segment combination can summarize the outline of entities A and B1, and perform 1:1 geographic entity matching of surface entities A and B1. Of course, "whether the identified feature line segments can summarize the corresponding shape of the object" is the standard for regularized rectangular object matching. When performing more complex surface entity matching, the matching degree is mainly based on the matching degree. In the matching process, due to the large number of concave-convex structures of large-scale data objects and the confusion of the boundaries to be matched, the external feature line segments of the same object in different data sets cannot be fully identified. Therefore, the matching degree (MD) is proposed to analyze the percentage of feature line segments actually identified. The higher the matching degree, the greater the probability of accurate matching. The matching degree calculation formula is as follows:

[0180] MD=l A / l S ×100% (9);

[0181] In the formula, l A is the cumulative length of the external feature line segments actually matched in the pattern, l s is the total length of the external feature segments.

[0182] As shown in Figure 14(d) and Figure 14(e), all external feature line segments belonging to the planar entities A, B1, and B2 are identified according to the pattern definition and threshold limits (threshold corresponding to the maximum projection distance and threshold corresponding to the relative offset rate), and the planar entity A is matched with B1 and B2 in a 1:N manner. Figure 14(f) is a typical spatial heterogeneity distribution of planar entities. According to the matching method of this embodiment, the planar entities A1 and A2 are matched with B1 and B2 in an M:N manner. As shown in Figure 14(g), it is a special spatial heterogeneity distribution caused by different map synthesis methods. The feature line segment identification method is different from that of Figure 14(f). After filtering the internal feature line segments, A2A3 in the data set A is not identified because there is no external feature line segment combined with it. At this time, the filtered B2B3 is restored, re-identified as an external feature line segment, and combined with A2A3. Similarly, B5B6 is combined with the restored A5A6. The recognition feature line segment combination can summarize the outlines of the corresponding surface entities A1, A2 and B1, B2, and the surface entities A1, A2 and B1, B2 are matched in M:N geographic entities.

[0183] Before executing step 202-3, the first matching further includes:

[0184] Processing (feature screening) the characteristic line segments of the target geographic entity object, wherein the target geographic entity object refers to the planar first geographic entity object or the planar second geographic entity object in the multi-planar entity matching combination, and the multi-planar entity matching combination refers to the planar entity matching combination including multiple planar first geographic entity objects or multiple planar second geographic entity objects.

[0185] When the same dataset has multiple planar geographic entity objects in the rough matching results, the feature line segments that can summarize the overall contours of the same dataset with multiple planar geographic entity objects are called external feature line segments, and the feature line segments inside the overall contours of the same dataset with multiple planar geographic entity objects are called internal feature line segments, such as B1B2, B3B4, B4B1, B5B6, B6B7, and B7B8 in Figure 15(a) and Figure 15(b). Before the precise matching of the planar geographic entity objects, it is necessary to screen the internal feature line segments and retain the external feature line segments.

[0186] When the target geographic entity object is an artificial morphological feature, the process of processing the characteristic line segments of the target geographic entity object by using the buffer zone expansion-contraction specifically includes:

[0187] (1) Determine whether there is an adjacency relationship between the plurality of target geographic entity objects, and obtain a fourth determination result.

[0188] (2) When the fourth judgment result is yes, the common feature line segments between the plurality of target geographic entity objects are removed.

[0189] As shown in Figure 15(a), the two polygonal geographic entity objects in dataset B are adjacent, and the shared feature line segments (common feature line segments), namely B8B5, are directly filtered out. At the same time, B2B5 and B8B3 whose angles are close to θ (θ is close to 180°) are queried and filtered.

[0190] (3) When the fourth judgment result is no, the contours of the plurality of target geographic entity objects are expanded according to the preset outward expansion distances to obtain an outward expansion buffer zone.

[0191] (4) All of the externally expanded buffers are merged into an overall buffer.

[0192] (5) The overall buffer zone is reduced according to a preset inward contraction distance to obtain a reduced overall buffer zone, wherein the preset inward contraction distance is greater than the preset outward expansion distance.

[0193] (6) removing characteristic line segments intersecting with the reduced overall buffer zone from the plurality of target geographic entity objects.

[0194] If the two entity polygons in data set B are separated, then an outward expansion buffer distance a is set to merge the housing buffer clusters in data set B, and then an inward contraction buffer distance b (b>a) is set to contract the buffer zone and remove B2B3 and B8B5 that intersect with the buffer zone, as shown in Figure 15(b).

[0195] (7) The feature line segments retained after the removal operation are used as the final feature line segments of the plurality of target geographic entity objects.

[0196] When filtering feature line segments for geographic entity objects with natural morphological features, new feature points must be added. This is because the method of extracting feature points based on the vertical distance limit method causes the position and number of feature points to be random. Even a solid polygon contains only one feature point, causing the feature line segment to be a closed curve. When filtering feature line segments subsequently, the entire closed curve will be regarded as an internal feature line segment and will be filtered out, which goes against the original intention of filtering internal feature line segments.

[0197] Therefore, when the target geographic entity object is an artificial morphological feature, the characteristic line segment of the target geographic entity object is processed by using the buffer expansion-contraction and adding feature points. The processing process is the same as steps (1)-(5) in the process of processing the characteristic line segment of the target geographic entity object by using the buffer expansion-contraction. The only difference is that after obtaining the reduced overall buffer zone, the step of processing the characteristic line segment of the target geographic entity object by using the buffer expansion-contraction and adding feature points includes:

[0198] The intersection points of the reduced overall buffer zone and the plurality of target geographic entity objects are taken as newly added feature points; and the newly added feature line segments are removed, wherein the newly added feature line segments are the connecting lines between two adjacent newly added feature points on the same target geographic entity object.

[0199] For the geographical entities Pb5 and Pb6 with natural morphological features in Figure 16(a), the buffer is expanded outward (as shown in Figure 16(b)) and contracted inward at the same distance to obtain the reduced overall buffer in Figure 16(c). Then, Pb5 and Pb6 before buffering are overlapped with the reduced overall buffer for analysis (as shown in Figure 16(d)). The four intersection points in Figure 16(d) are used as new feature point segmentation segments. The new feature points are shown in Figure 16(e). The internal feature line segments Lb14, Lb15, and Lb16 associated with the new feature points are selected, as shown in Figure 16(f).

[0200] Step 202-5: Calculate a first semantic similarity based on the semantic features of the planar first geographic entity object and the semantic features of the planar second geographic entity object in the planar entity space matching result.

[0201] Step 202-6: Determine whether the first semantic similarity is greater than a first threshold, and obtain a first determination result.

[0202] Step 202-7: When the first judgment result is yes, the planar entity space matching result is used as the final planar entity matching result; otherwise, return to step 202-1.

[0203] Step 203: Use a second matching process to match the linear first geographic entity object and the linear second geographic entity object, wherein the linear first geographic entity object is the first geographic entity object of the linear geographic entity type in the first data set, and the linear second geographic entity object is the second geographic entity object of the linear geographic entity type in the second data set.

[0204] The second matching process includes:

[0205] Step 203 - 1 : performing shape segmentation on all the linear first geographic entity objects and the linear second geographic entity objects, and correspondingly obtaining a plurality of third characteristic line segments and a plurality of fourth characteristic line segments.

[0206] Step 203-2: According to the spatial features of the third feature line segment and the spatial features of the fourth feature line segment, the third feature line segment and the fourth feature line segment are roughly matched to obtain a fourth matching result and a fifth matching result, wherein the fourth matching result is a rough matching result of the third feature line segment, and the fifth matching result is a rough matching result of the fourth feature line segment. The fourth matching result includes a plurality of first linear matching combinations, the first linear matching combination includes at least one of the third feature line segments and zero of the fourth feature line segments, or the first linear combination includes at least one of the third feature line segments and at least one of the fourth feature line segments, and the fifth matching result includes a plurality of second linear matching combinations, the second linear matching combination includes at least one of the fourth feature line segments and zero of the third feature line segments, or the second linear combination includes at least one of the fourth feature line segments and at least one of the third feature line segments.

[0207] Step 203 - 3 : Based on the spatial features of the third characteristic line segment and the spatial features of the fourth characteristic line segment in the first linear matching combination, perform precise matching on the third characteristic line segment and the fourth characteristic line segment in the first linear matching combination to obtain a sixth matching result.

[0208] Step 203-4: performing precise matching on the third characteristic line segment and the fourth characteristic line segment in the second linear matching combination according to the spatial characteristics of the third characteristic line segment and the spatial characteristics of the fourth characteristic line segment in the second linear matching combination, to obtain a seventh matching result;

[0209] Step 203 - 5 : Determine a linear entity matching result according to the sixth matching result and the corresponding seventh matching result.

[0210] Step 203 - 6 : Calculate a second semantic similarity based on the semantic features of the first linear geographic entity object and the semantic features of the second linear geographic entity object in the linear entity matching result.

[0211] Step 203 - 7 : Determine whether the second semantic similarity is greater than a second threshold, and obtain a second determination result.

[0212] Step 203 - 8 : When the second judgment result is yes, the linear entity matching result is used as the final linear entity matching result; otherwise, return to step 203 - 2 .

[0213] It should be noted that the rough matching and fine matching methods of the linear geographic entity objects in this embodiment are the same as those of the planar geographic entity objects, the only difference being that the rough matching processing objects of the linear geographic entity objects are feature line segments.

[0214] like Fig.17 As shown in , linear entity matching is mainly divided into two steps: coarse matching and precise matching, and both are based on feature line segments as basic units. Fig.17 For the same area road data of different data sets, the extracted feature segments are respectively expressed as La i (i=1,2,...,18), Lb i (i=1,2,...,21) means, referred to as a i , b i .

[0215] Step 1: In the coarse matching process, a buffer zone is established outward with the minimum circumscribed rectangle of the feature line segment as the center, and the feature line segments of other data sets falling within the buffer zone are searched in sequence. The search result is a one-way coarse matching result, and the one-way matching symbol is represented by →.

[0216] Step 2: In the precise matching process, the one-way precise matching of each feature line segment in the table is determined by combining the mode definition and the projection distance and relative offset rate threshold limits. The combination of feature line segments that satisfy both A→B and B→A is searched to obtain the two-way matching of the feature line segments in the table. The two-way matching symbol is represented by “:”. The one-way rough matching results, one-way precise matching results, and two-way precise matching results are shown in Table 6.

[0217] Table 6 Linear entity matching results One-way rough matching

[0218]

[0219] One-way exact match

[0220]

[0221] Bidirectional exact match

[0222]

[0223] The matching process of the planar geographic entity objects and the linear geographic entity objects in the present implementation involves the steps of shape feature segmentation, coarse matching and fine matching based on spatial features, and verification and optimization of the spatial matching results based on semantic features, thereby improving the accuracy of geographic entity object matching.

[0224] Since spatial matching has limitations, specifically, when the spatial positions of the two data sets are very different, mismatches and missed matches may occur in the matching results. For example, as shown in the schematic diagram of the ambiguous results of road spatial matching in Figure 18(a), the linear entities a2 and b4 and b5 are mismatched in a one-to-many manner. The reason for the error is that the correct matches a2 and b2, a4 and b4, a5 and b5 are not the closest in space, and the lengths of the feature line segments of the mismatches are roughly the same, resulting in failure to recognize the pattern. As shown in the schematic diagram of the ambiguous results of residential spatial matching in Figure 18(b), the locations of residential areas A2 and B2 are very different, and the external feature line segments that constitute the entity polygon are not fully recognized, resulting in missed matches between residential areas A2 and B2.

[0225] Therefore, after matching based on spatial features, this embodiment verifies and optimizes the spatial matching results. Since the process used for linear geographic entity objects and planar entity objects is the same (the only difference is that different objects are processed), this embodiment takes the spatial matching results of planar geographic entity objects as an example to specifically explain the steps of verifying and optimizing the spatial matching results based on semantic features.

[0226] Calculating a first semantic similarity according to the semantic features of the planar first geographic entity object and the semantic features of the planar second geographic entity object in the planar entity space matching result specifically includes:

[0227] Step 202-5-1: Calculate the similarity between the corrected target attribute features and the corrected matching attribute features of the same type, wherein the corrected target attribute features refer to the attribute features of the planar first geographic entity object in the planar entity space matching result, and the corrected matching attribute features are the attribute features of the corresponding planar second geographic entity object in the planar entity space matching result.

[0228] The corresponding attribute items in the geographic entity dataset are selected, the attribute items are divided into different types, and the similarities between the corresponding attribute items of geographic entities are calculated according to different methods, as shown in Table 7.

[0229] Step 202-5-2: Establishing a semantic relationship between the modified target attribute feature and the modified matching attribute feature, and determining the weight of each of the modified target attribute feature and each of the modified matching attribute features according to the entropy method.

[0230] According to the semantic connection between the concepts of attribute items, the relationship between attribute items is established: name-NAME, type-TYPE, code-CODE, number of floors-FLOORS. The entropy value of the attribute item is calculated according to the distribution of the attribute value and the corresponding weight is calculated according to the entropy value, as shown in Table 8 and Table 9.

[0231] Step 202-5-3: Calculate the first semantic similarity based on all the similarities and corresponding weights.

[0232] The similarity of the spatial matching results is calculated by combining the similarity between the attribute items and the weights of the attribute items, as shown in Table 7.

[0233] Table 7 Results of semantic similarity measurement

[0234]

[0235]

[0236] Table 8 Entropy values ​​of attribute items

[0237]

[0238] Table 9 Attribute feature combination weight values

[0239]

[0240] Experimental verification:

[0241] Figure 19(a) is a distribution map of residential entities in the same area at a scale of 1:500, and Figure 19(b) is a distribution map of residential entities in the same area at a scale of 1:10000. Table 10 is a statistical display of the matching results obtained using this method. The precision and recall rates are both higher than 90%, and the processing time is significantly shortened compared to manual matching. The results show that this method has the following advantages: (1) It does not distinguish between reference datasets and target datasets, and has obvious advantages in matching residential areas with fuzzy scales; (2) According to the actual boundary relationship of residential areas, a spatial matching model based on feature line segments is proposed, and the matching results have model support and are closer to manual matching results; (3) Based on the chaotic boundary of surface objects, a definition of matching degree is proposed, and the matching results are output in batches, which greatly reduces the number of missed matches; (3) The spatial matching results are optimized based on semantic similarity measurement, and the number of wrong matches is greatly reduced.

[0242] Figure 20(a) and Figure 20(b) are schematic diagrams of river entity dataset A and river entity dataset B in the same space and different time periods, respectively. The statistical information of the matching results is shown in Table 11. The recall and precision of river matching are both above 90%, and the time is shortened by more than 7 times compared with the manual matching results. The results show that this method has the following advantages: (1) The matching result is independent of the matching direction. Matching dataset A to B or matching dataset B to A will obtain the same result. This advantage is particularly prominent in matching datasets with different LODs; (2) According to the actual situation, a series of spatial patterns between line segments that will appear when matching objects are proposed. With the pattern as support, the matching results are closer to human eye recognition and can handle many-to-many complex situations well; (3) Based on the spatial pattern, the initial matching is performed, and then combined with the semantic similarity estimation for detection and correction, the accuracy of the matching results is greatly improved; (4) It can match linear entities with artificial and natural morphological characteristics such as roads and rivers, and has good robustness.

[0243] Table 10 Statistics of residential area matching results

[0244]

[0245] Table 11 Statistics of river matching results

[0246]

[0247] Example 2

[0248] This embodiment provides a computer device, including a memory and a processor. The memory stores a computer program. When the processor executes the computer program, the geographic entity generalization matching method based on spatial and semantic features in the above-mentioned embodiment 1 is implemented.

[0249] Example 3

[0250] This embodiment provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the geographic entity generalization matching method based on spatial and semantic features in the above-mentioned embodiment 1 is implemented.

[0251] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0252] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A generalized matching method for geographic entities based on spatial and semantic features, characterized in that: The geographic entity generalization matching method based on spatial and semantic features includes: Acquire at least two data sets from different sources in the same study area, wherein the first data set includes at least one first geographic entity object, and the second data set includes at least one second geographic entity object; A first matching process is used to match a planar first geographic entity object and a planar second geographic entity object, wherein the planar first geographic entity object is a first geographic entity object of a planar geographic entity type in the first data set, and the planar second geographic entity object is a second geographic entity object of a planar geographic entity type in the second data set; Using a second matching process to match a linear first geographic entity object and a linear second geographic entity object, wherein the linear first geographic entity object is a first geographic entity object of a linear geographic entity type in the first data set, and the linear second geographic entity object is a second geographic entity object of a linear geographic entity type in the second data set; The first matching process includes: According to the spatial features of the planar first geographic entity object and the spatial features of the planar second geographic entity object, the planar first geographic entity object and the planar second geographic entity object are roughly matched to obtain a first matching result, wherein the first matching result includes a plurality of planar entity matching combinations, the planar entity matching combination includes at least one planar first geographic entity object and zero planar second geographic entity objects, or the planar entity matching combination includes zero planar first geographic entity objects and at least one planar second geographic entity object, or the planar entity matching combination includes at least one planar first geographic entity object and at least one planar second geographic entity object; Performing shape segmentation on the planar first geographic entity object and the planar second geographic entity object in the planar entity matching combination, and correspondingly obtaining a first characteristic line segment and a second characteristic line segment, wherein a closed shape formed by sequentially connecting a plurality of the first characteristic line segments is the shape of the planar first geographic entity object, and a closed shape formed by sequentially connecting a plurality of the second characteristic line segments is the shape of the planar second geographic entity object; According to the spatial features of the first feature line segment and the spatial features of the second feature line segment in each of the entity matching combinations, the first feature line segment and the second feature line segment are precisely matched to obtain a second matching result and a third matching result, respectively, wherein the second matching result is a precise matching result of the first feature line segment, and the third matching result is a precise matching result of the second feature line segment; According to the second matching result and the third matching result, the first matching result is corrected to obtain a planar entity space matching result; Calculating a first semantic similarity based on the semantic features of the planar first geographic entity object and the semantic features of the planar second geographic entity object in the planar entity space matching result; Determine whether the first semantic similarity is greater than a first threshold, and obtain a first determination result; When the first judgment result is yes, the planar entity spatial matching result is used as the final planar entity matching result; otherwise, return to the step of "roughly matching the planar first geographic entity object and the planar second geographic entity object according to the spatial features of the planar first geographic entity object and the spatial features of the planar second geographic entity object to obtain a first matching result"; The second matching process includes: Performing shape segmentation on all the linear first geographic entity objects and the linear second geographic entity objects, and correspondingly obtaining a plurality of third characteristic line segments and a plurality of fourth characteristic line segments; According to the spatial features of the third feature line segment and the spatial features of the fourth feature line segment, the third feature line segment and the fourth feature line segment are roughly matched to obtain a fourth matching result and a fifth matching result, wherein the fourth matching result is a rough matching result of the third feature line segment, the fifth matching result is a rough matching result of the fourth feature line segment, the fourth matching result includes a plurality of first linear matching combinations, the first linear matching combination includes at least one of the third feature line segment and zero of the fourth feature line segment, or the first linear combination includes at least one of the third feature line segment and at least one of the fourth feature line segment, and the fifth matching result includes a plurality of second linear matching combinations, the second linear matching combination includes at least one of the fourth feature line segment and zero of the third feature line segment, or the second linear combination includes at least one of the fourth feature line segment and at least one of the third feature line segment; According to the spatial features of the third characteristic line segment and the spatial features of the fourth characteristic line segment in the first linear matching combination, the third characteristic line segment and the fourth characteristic line segment in the first linear matching combination are precisely matched to obtain a sixth matching result; According to the spatial features of the third characteristic line segment and the spatial features of the fourth characteristic line segment in the second linear matching combination, the third characteristic line segment and the fourth characteristic line segment in the second linear matching combination are precisely matched to obtain a seventh matching result; Determine a linear entity matching result according to the sixth matching result and the corresponding seventh matching result; Calculating a second semantic similarity according to the semantic features of the first linear geographic entity object and the semantic features of the second linear geographic entity object in the linear entity matching result; Determine whether the second semantic similarity is greater than a second threshold, and obtain a second determination result; When the second judgment result is yes, the linear entity matching result is used as the final linear entity matching result; otherwise, return to the step of "roughly matching the third feature line segment and the fourth feature line segment according to the spatial features of the third feature line segment and the spatial features of the fourth feature line segment to obtain a fourth matching result and a fifth matching result".

2. The geographic entity generalization matching method based on spatial and semantic features according to claim 1 is characterized in that: According to the spatial features of the planar first geographic entity object and the spatial features of the planar second geographic entity object, roughly matching the planar first geographic entity object and the planar second geographic entity object to obtain a first matching result specifically includes: Determine the minimum circumscribed rectangle of the planar first geographic entity object and the planar second geographic entity object respectively; According to the planar first geographic entity object and the planar second geographic entity object corresponding to the minimum circumscribed rectangle having an intersecting relationship, constructing the planar entity matching combination including at least one planar first geographic entity object and at least one planar second geographic entity object; According to the planar first geographic entity object and the planar second geographic entity object corresponding to the minimum circumscribed rectangle that do not have an intersection relationship, the planar entity matching combination including at least one of the planar first geographic entity object and zero of the planar second geographic entity objects, or the planar entity matching combination including zero of the planar first geographic entity object and at least one of the planar second geographic entity objects is constructed.

3. The geographic entity generalization matching method based on spatial and semantic features according to claim 1 is characterized in that: The process of acquiring the first characteristic line segment or the third characteristic line segment specifically includes: constructing a geographic entity model according to the spatial relationship between the first geographic entity objects; It is judged in sequence whether each target line segment on the arc segment has a curvature feature to obtain a third judgment result, wherein the arc segment is a line segment representing the contour of the first geographic entity object in the geographic entity model, the target line segment is a line between the first point and the last point in a preset point group, the preset point group includes a preset number of points arranged in sequence on the arc segment, the points include nodes and geometric points, and of two adjacent target line segments, the last point in the preset point group of the former target line segment is used as the first point in the preset point group of the latter target line segment; When the third judgment result is yes, the first geographic entity object is determined to be a natural morphological feature feature, and the feature points are determined using the first extraction rule; Otherwise, determining that the first geographic entity object is an artificial morphological feature feature, and using a second extraction rule to determine the feature point; The first extraction rule includes: Select nodes whose number of associated arc segments is greater than or equal to the quantity threshold as feature points; The vertical distance limit method is used to extract the characteristic points on the arc segment; Taking the line between two adjacent characteristic points arranged in sequence as the first characteristic line segment or the third characteristic line segment; The second extraction rule includes: Select nodes whose number of associated arc segments is greater than or equal to the quantity threshold as feature points; The vertex of the common angle is used as a feature point, wherein the common angle refers to the angle formed by taking the common geometric points of two adjacent target line segments as vertices, and the common angle is less than or equal to a preset value; If there is a concave-convex structure in the outline of the first geographic entity object, then removing feature points in an independent concave-convex structure whose area is smaller than an area threshold, and removing feature points on a common edge of two adjacent concave-convex structures, wherein the independent concave-convex structure refers to a concave-convex structure that exists alone; A line connecting two adjacent characteristic points arranged in sequence is used as the first characteristic line segment or the third characteristic line segment.

4. The geographic entity generalization matching method based on spatial and semantic features according to claim 1 is characterized in that: Before executing the step of "according to the spatial features of the first feature line segment and the spatial features of the second feature line segment in each entity matching combination, the first feature line segment and the second feature line segment are precisely matched to obtain a second matching result and a third matching result respectively", the first matching process further includes: Processing the characteristic line segments of the target geographic entity object, wherein the target geographic entity object refers to a planar first geographic entity object or a planar second geographic entity object in a multi-planar entity matching combination, and the multi-planar entity matching combination refers to the planar entity matching combination including a plurality of planar first geographic entity objects or a plurality of planar second geographic entity objects; When the target geographic entity object is an artificial morphological feature object, the process of processing the characteristic line segments of the target geographic entity object specifically includes: Determine whether there is an adjacency relationship between the plurality of target geographic entity objects, and obtain a fourth determination result; When the fourth judgment result is yes, removing common feature line segments between the plurality of target geographic entity objects; When the fourth judgment result is no, the contour lines of the plurality of target geographic entity objects are respectively expanded according to the preset outward expansion distances to obtain an outward expansion buffer zone; Merging all of the outward expansion buffers into an overall buffer; The overall buffer zone is reduced according to a preset inward contraction distance to obtain a reduced overall buffer zone, wherein the preset inward contraction distance is greater than the preset outward expansion distance; Removing characteristic line segments intersecting with the reduced overall buffer zone from a plurality of the target geographic entity objects; The characteristic line segments retained after the removal operation are used as the final characteristic line segments of the plurality of target geographic entity objects.

5. The geographic entity generalization matching method based on spatial and semantic features according to claim 1 is characterized in that: Before executing the step of "according to the spatial features of the first feature line segment and the spatial features of the second feature line segment in each entity matching combination, the first feature line segment and the second feature line segment are precisely matched to obtain a second matching result and a third matching result respectively", the first matching process further includes: Processing the characteristic line segments of the target geographic entity object, wherein the target geographic entity object refers to a planar first geographic entity object or a planar second geographic entity object in a multi-planar entity matching combination, and the multi-planar entity matching combination refers to the planar entity matching combination including a plurality of planar first geographic entity objects or a plurality of planar second geographic entity objects; When the target geographic entity object is a natural morphological feature, the process of processing the characteristic line segments of the target geographic entity object specifically includes: Determine whether there is an adjacency relationship between the plurality of target geographic entity objects, and obtain a fourth determination result; When the fourth judgment result is yes, removing common feature line segments between the plurality of target geographic entity objects; When the fourth judgment result is no, the contours of the plurality of target geographic entity objects are respectively expanded according to the preset outward expansion distances to obtain an outward expansion buffer zone; Merging all of the outward expansion buffers into an overall buffer; The overall buffer zone is reduced according to a preset inward contraction distance to obtain a reduced overall buffer zone, wherein the preset inward contraction distance is greater than the preset outward expansion distance; The intersection points of the reduced overall buffer zone and the plurality of target geographic entity objects are used as newly added feature points; Removing a newly added feature line segment, wherein the newly added feature line segment is a line connecting two adjacent newly added feature points on the same target geographic entity object; The characteristic line segments retained after the removal operation are used as the final characteristic line segments of the plurality of target geographic entity objects.

6. The geographic entity generalization matching method based on spatial and semantic features according to claim 1 is characterized in that: According to the spatial features of the first characteristic line segment and the spatial features of the second characteristic line segment in each entity matching combination, the first characteristic line segment and the second characteristic line segment are precisely matched to obtain a second matching result, specifically including: For each entity matching combination, the planar first geographic entity object in the entity matching combination is used as a target geographic entity object, and the corresponding planar second geographic entity object is used as a matching geographic entity object; According to different pattern recognition conditions, the matching pattern between the target feature line segment and the corresponding matching feature line segment is identified, and the feature matching combination corresponding to the different matching patterns is determined, wherein the target feature line segment is each of the first feature line segments of the planar first geographic entity object in any of the entity matching combinations, and the corresponding matching feature line segment is each of the second feature line segments of the planar second geographic entity object in the corresponding entity matching combination, and the matching patterns include a basic one-to-one pattern, a derived one-to-one pattern, a one-to-many pattern, and a many-to-many pattern; Calculating the maximum projection distance and relative offset rate between the target feature line segment and the matching feature line segment in the feature matching combination; When the maximum projection distance is greater than the distance threshold, and the corresponding relative offset rate is greater than the relative offset rate threshold, it is determined that the corresponding target feature line segment and the matching feature line segment in the feature matching combination are matched successfully; The successfully matched feature matching combination is used as the second matching result.

7. The geographic entity generalization matching method based on spatial and semantic features according to claim 6 is characterized in that: The pattern recognition condition of the basic one-to-one pattern correspondence includes that the target feature line segment and the matching feature line segment have a coincidence relationship, or that the target feature line segment and the matching feature line segment have an overlapping area after mutual projection; The expression of the pattern recognition condition for the derived one-to-one pattern correspondence is: Among them, l ij The overlapping area S ij The length of angle[B j B j+1 ] is a continuous matching feature line segment B with right-angle features j B j+1 The angle between them; α is the angle threshold; d ij The overlapping area S ij The average projection distance of Indicates the target feature line segment A1A2 and the matching feature line segment B j B j+1 The overlapping areas between them intersect, or the shared edge of the overlapping area is the target feature line segment or the matching feature line segment; Case II means that when there are multiple continuous matching feature line segments and the matching feature line segments have right-angle features, condition 1 is used for identification; Case III means that when two matching feature line segments in the same data set are discontinuous and the length of the longest overlapping area is much longer than the length of any other overlapping area, condition 2 is used for identification; Case IV means that when two matching feature line segments in the same data set are discontinuous and the length of the overlapping area is close to the length of the longest overlapping area, condition 3 is used for identification; The expression of the pattern recognition condition corresponding to the one-to-many model is: in, Indicates that in the first data set and / or in the second data set, the distance between any geometric points of the feature line segment is less than the distance threshold Δ; Case V indicates that among multiple continuous matching feature line segments, when the matching feature line segment has a right-angle feature and the corner angle is less than the set threshold, condition a is used for identification; Case VI indicates that among multiple continuous matching feature line segments, when the matching feature line segment has a right-angle feature and the corner angle is greater than or equal to the set threshold, condition b is selected for identification; Case VII indicates that among multiple continuous matching feature line segments, when the matching feature line segment has a curvature feature, condition c is selected for identification; Case VIII indicates that among multiple discontinuous matching feature line segments, when the matching feature line segment has a curvature feature, condition d is selected for identification; The expression of the pattern recognition condition corresponding to the many-to-many pattern is: Among them, angle[A i A i+1 ]、angle[B j B j+1 ] refer to the angles between the feature line segments in the first data set and the second data set respectively; case IX means that when multiple continuous feature line segments in the first data set and the second data set have right-angle features and the angle between any feature line segments in the same data set is greater than or equal to the set threshold value α, condition aa is used for identification; case X means that when multiple continuous feature line segments in both data sets have curved features, condition bb is used for identification; case XI means that when multiple continuous feature line segments in the first data set and the second data set have right-angle features and the angle between feature line segments in the same data set is less than the set threshold value α, condition cc is used for identification; case XII means that when the feature line segments in the first data set are discontinuous, and multiple continuous feature line segments in the second data set have right-angle features and the angle between any feature line segments in the same data set is greater than or equal to the angle threshold value α, condition dd is used for identification; case XIII means that when the feature line segments in the first data set and the second data set are both discontinuous, condition ee is used for identification; case XIV means that when the feature line segments in the first data set are discontinuous, and multiple continuous feature line segments in the second data set have curved features, condition ff is used for identification.

8. The geographic entity generalization matching method based on spatial and semantic features according to claim 1 is characterized in that: Calculating a first semantic similarity according to the semantic features of the planar first geographic entity object and the semantic features of the planar second geographic entity object in the planar entity space matching result specifically includes: Calculate the similarity between the modified target attribute feature and the modified matching attribute feature of the same type, wherein the modified target attribute feature refers to the attribute feature of the planar first geographic entity object in the planar entity space matching result, and the modified matching attribute feature refers to the attribute feature of the planar second geographic entity object in the planar entity space matching result; Establishing a semantic connection between the modified target attribute feature and the modified matching attribute feature, and determining the weight of each of the modified target attribute feature and each of the modified matching attribute features according to an entropy method; The first semantic similarity is calculated according to all the similarities and the corresponding weights.

9. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the geographic entity generalization matching method based on spatial and semantic features as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for generalizing and matching geographic entities based on spatial and semantic features described in any one of claims 1 to 8 is implemented.

Images