A method for matching heterogeneous point pairs based on local topological features of triangles

Abstract

The application discloses a heterogeneous point pair matching method based on triangular local topological features and belongs to the field of geographic information processing. The method comprises the following steps: obtaining two semantic point sets; constructing local topological feature point subsets and non-topological feature point subsets of the two point sets; constructing triangular topological features according to the local topological feature point sets to obtain topological features of the two semantic point sets; establishing feature description sub-matching pairs to obtain a topological matching rule set of the topological features of the to-be-matched point set; performing frequency statistics on matching points in the matching relationship of each point in the to-be-matched point set, and the matching point with the highest frequency is the final matching point of the point; updating attribute and position information of each point to obtain an updated to-be-matched point set. The application can extract local semantic topological information with good robustness and can effectively support geographic point matching under different scale spaces and point position error conditions.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of geographic information processing, and particularly relates to a method for matching heterogeneous point pairs based on local topological features of triangles. BACKGROUND

[0002] With the rapid development of geospatial information technology, in order to more conveniently collect multi-temporal information for the same kind of ground object or geographic entity, multi-means are usually adopted to collect data of entities with the same geographic semantics. In order to effectively utilize the multi-means information, the fusion of heterogeneous data is often faced.

[0003] The information of different data sources has differences in time resolution, spatial resolution and error, and even the information representing the same kind of ground object element has great differences in the results presented by different means, which brings many difficulties to the fusion of heterogeneous data. At the same time, the improvement of the edge computing capability of existing devices makes it possible to efficiently extract geographic entity semantic information, which provides a new source of geographic entity data acquisition. Therefore, in the face of the demand for multi-source data fusion, how to design a relatively stable matching strategy according to the characteristics of different data sources has become an important research direction. SUMMARY

[0004] The application aims to provide a method for matching heterogeneous point pairs based on local topological features of triangles, which can extract local semantic topological information with better robustness and effectively support the matching of geographic points under different scales of space and point error conditions.

[0005] In order to achieve the above-mentioned purpose, the technical scheme adopted by the application is as follows:

[0006] A method for matching heterogeneous point pairs based on local topological features of triangles, comprising the following steps:

[0007] Step 1, obtaining two heterogeneous geographic information semantic point sets of the same geographic region and wherein, has accurate attribute information and geographic position information, is a point set to be matched which lacks accurate attribute information and geographic position information;

[0008] Step 2, selecting the minimum point in the plane coordinates of the two point sets and and

[0009] Step 3, finding the minimum point in the plane coordinates of the point sets and and​​ the nearest point of and

[0010] Step 4, constructing two point sets and the local topological feature point subsets of and Initially:

[0011]

[0012]

[0013] At the same time, construct the non-topological feature point subsets of and and

[0014]

[0015]

[0016] Step 5, respectively according to the points in the feature point subsets and to form line segments and and find the nearest point of the midpoint of and in the corresponding non-topological feature point subsets and and update the feature point subsets and

[0017]

[0018]

[0019]

[0020]

[0021] Step 6, respectively based on the points in the local topological feature point subsets and to form triangular topological features and

[0022]

[0023] ​​

[0024] in, They are respectively from The three points form the largest and smallest interior angles of a triangle. They represent respectively by The minimum and second minimum normalized side lengths of the triangle formed by the three points;

[0025] Step 7, let Repeat steps 5 through 7 until... and If the set is empty, then all the topological features of the triangles obtained are the semantic point set. and Topological features and

[0026]

[0027]

[0028] in, They are respectively from The three points form the largest and smallest interior angles of a triangle. They represent respectively by The smallest and second smallest normalized side lengths of the triangle formed by the three points, f * Let f be the number of topological triangles, and f d ≤f m , Representing the topological features of a triangle;

[0029] Step 8, for the set of points to be matched Topological features Topological features of each triangle In topological features Find the most similar triangle topological features Constructing feature descriptor matching pairs

[0030] Step 9, matching pairs for each feature descriptor Included and Representing point sets respectively and The three points in the diagram are used to obtain the matching rule set for feature descriptor matching pairs based on the angle matching principle.

[0031] Step 10, based on the topological features of each triangle Corresponding matching rule set get Topological features Topology matching rule set

[0032]

[0033] Step 11, topology matching rule set Contains a point set to be matched The matching relationship of each point in The matching relationship of point is recorded as

[0034] Step 12, frequency statistics of matching points in the matching relationship of point The most frequently appearing is the final matching point of point According to the final matching point Update the attribute and location information of all points Thus, the point set to be matched lacking accurate attribute information and geographic location information is updated to a point set containing accurate attribute information and geographic location information

[0035] Further, the semantic point set in step 1 and are respectively represented as:

[0036]

[0037]

[0038] In the formula, n d and n m respectively represent the number of points in the semantic point set and n m ≥n d , id is the accurate semantic information of the jth point in , and and respectively represent the coordinate system space of the point set and , wherein represents a more accurate real geographic coordinate system.

[0039] Further, in step 8, is represented as:

[0040]

[0041]

[0042] ​where dist(i,j) represents the distance metric of topological descriptor and min() represents the minimum value of elements.

[0043] Further, the matching rule set of the feature descriptor matching pair in step 9 is :

[0044]

[0045] In the formula, A point and B point establish a corresponding matching relationship, the feature descriptor matching pair is represented as wherein and respectively contain and Each of the three points in each descriptor constitutes a triangle, and here and and and respectively represent the vertices corresponding to the largest, second largest and smallest internal angle in the triangle.

[0046] Further, the matching relationship of the point in step 11 is represented as:

[0047]

[0048] wherein l n represents the final number of matching points, which is a positive integer.

[0049] Further, the final matching point of the point in step 12 is represented as:

[0050]

[0051] wherein, represents the frequency maximum matching point in the matching relationship , Max() represents the maximum value, represents the frequency of the point ;

[0052] The updated point set is:

[0053]

[0054] The present application has the following beneficial effects:

[0055] (1) The application provides a heterogeneous point pair matching method based on triangular local topological features, and the matching of the heterogeneous point pair can be effectively extracted by constructing the spatial topological relationship of the semantic point set.

[0056] (2) The application can extract relatively stable local topological features, and provide effective support for the matching and fusion of semantic point sets of different data sources. BRIEF DESCRIPTION OF DRAWINGS

[0057] Figure 1 It is a schematic diagram of the principle of the heterogeneous point pair matching method based on triangular local topological features. DETAILED DESCRIPTION

[0058] The application will be further described in detail below in combination with the drawings and specific embodiments.

[0059] A heterogeneous point pair matching method based on triangular local topological features, comprising the following steps:

[0060] Step 1, two semantic point sets are respectively obtained by multiple means and and have accurate attribute information and geographic location information;

[0061] Step 2, the minimum point of the plane coordinates in the two point sets and is selected respectively and

[0062] Step 3, the nearest point of the minimum point of the plane coordinates and in the point set and is found respectively and

[0063] Step 4, the local topological feature point subsets of the two point sets and are constructed and Meanwhile, the non-topological feature point subsets are and

[0064] Step 5, the line segments and are formed according to the points in the point subsets and and the nearest point of the midpoint in the corresponding non-topological feature point subsets and is found and Update the idea set separately and

[0065] Step 6, based on local topological feature point subsets respectively and The points in the triangle form a topological feature. and

[0066] Step 7, let Repeat steps 5 through 7 until... and If the set is empty, then the resulting triangle has the following topological features. and For semantic point set and Topological features;

[0067] Step 8, for the set of points to be matched Topological features Each feature descriptor In topological features Find the most similar feature descriptor Constructing feature descriptor matching pairs

[0068] Step 9, matching pairs for each feature descriptor Included feature descriptors and These are semantic point sets and Based on the angle matching principle, the matching rule set for feature descriptor matching pairs can be obtained from the three points in the image.

[0069] Step 10, based on each feature descriptor Corresponding matching rule set Available Topological features Topology matching rule set

[0070] Step 11, Topology matching rule set Contains a set of points to be matched The matching relationship of each point in the table, for point... The matching relationship is denoted as

[0071] Step 12, point-to-point Matching relationship The most frequently occurring point in the matching points is the point that appears the most often. final matching point Then you can update the points. The attribute and location information of the points to be matched is updated to...

[0072] Furthermore, the semantic point set in step 1 and They are represented as follows:

[0073]

[0074]

[0075] In the formula, n d and n m Representing semantic point sets respectively and Number of points (n) m ≥n d ), id is The precise semantic information of the j-th point. and Representing point sets respectively and The coordinate system space, where It represents a more accurate real-world geospatial coordinate system.

[0076] Furthermore, in step 2, the point set and Minimum point of planar coordinates in and They can be represented as:

[0077]

[0078]

[0079] in, and Representing points respectively and The planar coordinates in its corresponding coordinate space.

[0080] Furthermore, the nearest neighbor in step 3 and They can be represented as:

[0081]

[0082]

[0083] Here, dis(,) represents the planar distance between two points.

[0084] Furthermore, in step 4, for the semantic point set and The local feature subset can be represented as follows:

[0085]

[0086]

[0087] The non-topological feature point subset can be represented as:

[0088]

[0089]

[0090] Further, the most adjacent points in step 5 and can be represented as:

[0091]

[0092]

[0093] wherein, and respectively represent the midpoints of the line segments and ;

[0094] The updating method of each point subset and is as follows:

[0095]

[0096]

[0097]

[0098]

[0099] Further, the triangular topological features in step 6 and can be respectively represented as:

[0100]

[0101]

[0102] wherein, are respectively the largest internal angle and the smallest internal angle of the triangle formed by three points, respectively represent the smallest normalized edge length and the second smallest normalized edge length of the triangle formed by three points.

[0103] Further, the triangle topology feature in step 7 and may be respectively expressed as:

[0104]

[0105]

[0106] wherein, are the maximum and minimum internal angles of the triangle formed by the three points, are the minimum and second minimum normalized edge lengths of the triangle formed by the three points, f * is the number of topology triangles, and f d ≤f m , denotes the triangle topology feature.

[0107] Further, the feature descriptor most similar to the feature descriptor in step 8 can be expressed as:

[0108]

[0109]

[0110] wherein dist(i,j) denotes the distance measure of the topology descriptors and , and min() denotes the minimum value of the elements.

[0111] Further, the matching rule set of the feature descriptor matching pair in step 9 is:

[0112] In the formula,

[0113] denotes that the point A and the point B establish a corresponding matching relationship, and the feature descriptor matching pair may also be expressed as wherein and contain and respectively. Each descriptor in and and and may form a triangle, and and

[0114] ​Furthermore, the topology matching rule set in step 10 It can be represented as:

[0115]

[0116] Furthermore, in step 11... Matching relationship It can be represented as:

[0117]

[0118] Among them, l n This represents the final number of the matching point, which is a positive integer.

[0119] Furthermore, in step 12... final matching point It can be represented as:

[0120]

[0121] in, Indicates the matching relationship The maximum frequency matching point in the matrix; Max() represents finding the maximum value. Point The frequency of occurrence.

[0122] Finally, point The attribute and location information are updated as follows:

[0123]

[0124] Updated set of points to be matched for:

[0125]

[0126] Here is a more specific example:

[0127] like Figure 1 As shown, a heterogeneous point pair matching method based on local topological features of a triangle includes the following steps:

[0128] Step 1: Obtain two semantic point sets using multiple methods. and and It has precise attribute information and geographic location information, and can be represented as:

[0129]

[0130]

[0131] In the formula, n d and nm Representing semantic point sets respectively and Number of points (n) m ≥n d ), id is The precise semantic information of the j-th point. and Representing point sets respectively and The coordinate system space, where It represents a more accurate real-world geospatial coordinate system.

[0132] Step 2, select two point sets respectively. and Minimum point of planar coordinates in and It can be represented as:

[0133]

[0134]

[0135] in, and Representing points respectively and The planar coordinates in its corresponding coordinate space.

[0136] Step 3, in the point set respectively and Find the point with the minimum plane coordinates and nearest neighbor and It can be represented as:

[0137]

[0138]

[0139] Here, dis(,) represents the planar distance between two points.

[0140] Step 4, construct two point sets and Local topological feature point subset and It can be represented as follows:

[0141]

[0142]

[0143] Non-topological feature point subset is and may be expressed as:

[0144]

[0145]

[0146] Step 5, line segments and are respectively constructed based on the points in point subsets and and the midpoints in the corresponding non-topological feature point subsets and are respectively determined as the nearest points to the midpoints of and may be expressed as:

[0147]

[0148]

[0149] wherein, and respectively represent the midpoints of line segments and ;

[0150] The updating method of each point subset and is:

[0151]

[0152]

[0153]

[0154]

[0155] Step 6, triangular topologies and are respectively constructed based on the points in local topological feature point subsets and may be respectively expressed as:

[0156]

[0157]

[0158] wherein, are respectively the maximum internal angle and the minimum internal angle of a triangle formed by three points, respectively represent the minimum normalized edge length and the second minimum normalized edge length of a triangle formed by three points.​

[0159] Step 7, let Steps 5 to 7 are executed in a loop until and is an empty set, at this time, the triangular topology feature and is a topology feature of the semantic point set and can be expressed as:

[0160]

[0161]

[0162] wherein, are the maximum and minimum internal angles of the triangle composed of three points, respectively represent the minimum and second minimum normalized edge length of the triangle composed of three points, f * is the number of topological triangles, and f d f m , indicates the triangular topology feature.

[0163] Step 8, for each feature descriptor in the topology feature of the point set to be matched find the most similar feature descriptor in the topology feature to form a feature descriptor matching pair The feature descriptor most similar to the feature descriptor can be expressed as:

[0164]

[0165]

[0166] wherein, dist(i,j) represents the distance measure of the topology descriptor and min() represents the minimum value of the elements.

[0167] Step 9, the feature descriptors and contained in each feature descriptor matching pair are three points in the semantic point sets and respectively, and the matching rule set of the feature descriptor matching pair can be obtained according to the angle matching principle It can be represented as:

[0168]

[0169] In the formula, This indicates that a corresponding matching relationship is established between point A and point B; the feature descriptor matching pair It can also be expressed as in and Each contains and Each of the three points in a descriptor can form a triangle, here and and and These represent the vertices corresponding to the largest, second largest, and smallest interior angles within the triangle, respectively.

[0170] Step 10, based on each feature descriptor Corresponding matching rule set Available Topological features Topology matching rule set It can be represented as:

[0171]

[0172] Step 11, Topology matching rule set Contains a set of points to be matched The matching relationship of each point in the table, for point... The matching relationship is denoted as It can be represented as:

[0173]

[0174] Among them, l n This represents the final number of the matching point, which is a positive integer.

[0175] Step 12, point-to-point Matching relationship The most frequently occurring point in the matching points is the point that appears the most often. final matching point It can be represented as:

[0176]

[0177] in, Indicates the matching relationship The maximum frequency matching point in the matrix; Max() represents finding the maximum value. Point The frequency of occurrence.

[0178] Updatable points The attribute and position information are:

[0179]

[0180] Updated point set to be matched The attribute and position information are:

[0181]

[0182] In summary, the application proposes a heterogeneous point pair matching method based on triangle local topological features, which can realize fast matching of heterogeneous point sets. The method first constructs a stable topological structure of the point set and realizes fast rough matching of the topological descriptor, and then uses the matching point frequency to find the final matching point, which can effectively support geographic point matching under different scale spaces, point position errors and data sources.

Claims

1. A method for matching heterogeneous point pairs based on local topological features of a triangle, characterized in that, Includes the following steps: Step 1: Obtain the semantic point sets of geographic information from two different sources within the same geographic region. and in, It has accurate attribute information and geographic location information. The set of points to be matched that lack precise attribute information and geographic location information; Step 2, select two point sets respectively. and Minimum point of planar coordinates in and Step 3, in the point set respectively and Find the point with the minimum plane coordinates and nearest neighbor and Step 4, construct two point sets and Local topological feature point subset and Initially: At the same time, build and Non-topological feature point subset and Step 5, based on the feature point subsets respectively and The points in the middle form a line segment and And find and The midpoint lies in the corresponding non-topological feature point subset. and nearest neighbor in and Update feature point subsets separately and Step 6, based on local topological feature point subsets respectively and The points in the triangle form a topological feature. and in, They are respectively from P * min ,P * ne1 ,P * ne2 The three points form the largest and smallest interior angles of a triangle. They represent P respectively * min ,P * ne1 ,P * ne2 The minimum and second minimum normalized side lengths of the triangle formed by the three points; Step 7, let Repeat steps 5 through 7 until... and If the set is empty, then all the topological features of the triangles obtained are the semantic point set. and Topological features and in, They are respectively from P * qi ,P * ri ,P * si The three points form the largest and smallest interior angles of a triangle. They represent P respectively * qi ,P * ri ,P * si The smallest and second smallest normalized side lengths of the triangle formed by the three points, f * Let f be the number of topological triangles, and f d ≤f m , Representing the topological features of a triangle; Step 8, for the set of points to be matched Topological features Topological features of each triangle In topological features Find the most similar triangle topological features Constructing feature descriptor matching pairs Step 9, matching pairs for each feature descriptor Included and Representing point sets respectively and The three points in the diagram are used to obtain the matching rule set for feature descriptor matching pairs based on the angle matching principle. Step 10, based on the topological features of each triangle Corresponding matching rule set get Topological features Topology matching rule set Step 11, Topology Matching Rule Set Contains a set of points to be matched The matching relationship of each point in the data, the points The matching relationship is denoted as Step 12, point-to-point Matching relationship The matching points in the data are counted by frequency; the point that appears most frequently is the matching point. final matching point Based on the final matching point Update all points The attribute and location information is used to match the set of points that lack precise attribute and geographic location information. Updated to a point set containing precise attribute information and geographic location information.

2. The heterogeneous point pair matching method based on local topological features of a triangle according to claim 1, characterized in that, Semantic point set in step 1 and They are represented as follows: In the formula, n d and n m Representing semantic point sets respectively and The number of points, n m ≥n d id is The precise semantic information of the j-th point. and Representing point sets respectively and The coordinate system space, where It represents a more accurate real-world geographic coordinate system.

3. The heterogeneous point pair matching method based on local topological features of a triangle according to claim 2, characterized in that, In step 8, Represented as: Where dist(i,j) represents the topological descriptor and The distance metric is min(), which finds the minimum value of the elements in the metric.

4. The heterogeneous point pair matching method based on local topological features of a triangle according to claim 3, characterized in that, Feature descriptor matching pairs in step 9 Matching rule set for: In the formula, This indicates that a corresponding matching relationship is established between point A and point B; the feature descriptor matching pair Represented as in and Each contains and The three points in each descriptor form a triangle, here and and and These represent the vertices corresponding to the largest, second largest, and smallest interior angles within the triangle, respectively.

5. The heterogeneous point pair matching method based on local topological features of a triangle according to claim 4, characterized in that, In step 11 Matching relationship Represented as: Among them, l n This represents the final number of the matching point, which is a positive integer.

6. The heterogeneous point pair matching method based on local topological features of a triangle according to claim 5, characterized in that, Midpoint of step 12 final matching point Represented as: in, Indicates the matching relationship The maximum frequency matching point in the matrix; Max() represents finding the maximum value. Point The frequency of occurrence; Updated point set for:

Images