A scene matching method combining ground object semantic information and topological features

Abstract

The application discloses a scene matching method combining semantic information and topological features of ground objects, and belongs to the technical field of geographic information processing. The method comprises the following steps: dividing a city into a street block scene set by taking a main road as a boundary; calculating a scene centroid by using the center and area attributes of each building monomer in the city scene; establishing a polar coordinate system and calculating the coordinates of each building monomer in the city scene; dividing the polar coordinate system into sub-regions, counting the number of building monomers in each sub-region in the sub-region set, and obtaining a statistical histogram about the whole scene; obtaining a set of scenes to be matched, calculating the statistical histogram about the building monomers for each scene to be matched, calculating the similarity between the scene to be matched and any scene according to the difference degree of the statistical histogram, finding the best matching scene, and obtaining a matching scene set of the scene to be matched. The application can fully utilize the semantic structure information and spatial topological information of ground objects, and provides technical support for local scene interpretation and matching in a city and other applications.

Description

Technical Field

[0001] This invention belongs to the field of geographic information processing technology, specifically relating to a scene matching method that combines semantic information of ground features and topological features. Background Technology

[0002] With the widespread application of geographic information technology and remote sensing data, the acquisition of ground feature information has become more diversified, and its storage format is increasingly leaning towards advanced semantic information. For the same ground feature, different data acquisition methods will produce data information with different precision, frequency, and time. How to achieve matching of multi-source data is a key issue in solving the problem of multi-source data fusion and application.

[0003] In scenarios oriented towards geographic information acquisition and application, it is often necessary to combine multi-source data to study and analyze the distribution patterns of ground features within a scene, such as multi-temporal variation analysis and heterogeneous data correlation. Therefore, it is often necessary to solve the scene matching problem of heterogeneous data. Typically, scene matching tasks are based on feature extraction and are often applied to image-to-image matching. This feature-based scene matching method is extremely unfavorable for data storage in multi-source data matching applications across an entire city, greatly affecting the execution efficiency of scene matching tasks. With the development of edge computing, the computing power of edge devices has been greatly released, making real-time interpretation of scene semantic information a reality. Furthermore, for urban scene matching applications, the semantic information of ground features across the entire city is stored in the form of vector semantics. Therefore, how to utilize ground feature semantic information and spatial topological features to achieve real-time scene matching has become a key research focus. Summary of the Invention

[0004] The purpose of this invention is to provide a scene matching method that combines semantic information and topological features of ground features. This method can make full use of the semantic structural information and spatial topological information of ground features, and provide technical support for applications such as local scene interpretation and matching in cities.

[0005] The technical solution adopted in this invention is as follows:

[0006] A scene matching method combining semantic information and topological features of ground objects includes the following steps:

[0007] Step 1: Obtain a semantic map library of buildings and roads for the entire urban area, and divide the city into a set of street scenes R, using the main roads as boundaries. n The i-th scene can be represented as re i Each building in the scene contains detailed urban land use information;

[0008] Step 2, utilize urban scene re i Each building unit center and area Attribute calculation scene centroid

[0009] Step 3, using the scene centroid As the extreme point, the scene centroid The north direction is the polar axis direction, and the angle is taken as positive in the clockwise direction. Establish a polar coordinate system.

[0010] Step 4, calculate the re of the urban scene i Each building unit With the center of gravity of the scene Normalized distance D k and direction A k Then the scenario re i Each building unit in All in polar coordinate system In this context, it is represented as a point (D). k A k );

[0011] Step 5, convert the entire polar coordinate system The regions are partitioned according to equal intervals and angular intervals Δθ, and the sub-region set can be denoted as follows: It consists of multiple sub-regions constitute;

[0012] Step 6, based on the scenario re i All individual buildings in the polar coordinate system The coordinates are used to statistically analyze the sub-region sets. Each sub-region Number of individual buildings

[0013] Step 7: Based on the number of individual buildings in each sub-region, information about the entire scene can be obtained. i Statistical histogram

[0014] Step 8, obtain the set of scenarios to be matched R m The j-th scene can be represented as ra j The scene also includes information about individual buildings, but the absolute error of their geographic coordinates is large, and it only contains two attributes.

[0015] Step 9, for scene ra j Calculate the statistical histogram for each individual building using the methods in steps 2 to 7.

[0016] Step 10, using scene ra j As a scenario to be matched, based on the statistical histogram difference... Calculate its relationship with the scene set R n any scene in China i Similarity;

[0017] Step 11: Find the scene ra based on the statistical histogram difference. j The best matching scenario is

[0018] Step 12, obtain the scene to be matched R m Matching scenario set R m-n ;

[0019] Furthermore, in step 1, the street scene set R n It can be represented as:

[0020] R n ={re1,re2,...,re n}

[0021] In the formula, n represents the number of all scenes in the entire city, and re i The scenario with number i can be represented as:

[0022]

[0023] in, Representing the scenario re i The k-th building entity contains t attributes, which are respectively N i This represents the number of individual buildings.

[0024] Furthermore, in step 2, the scene centroid The calculation method is as follows:

[0025]

[0026]

[0027] in, For individual buildings The center coordinates, For individual buildings The area.

[0028] Furthermore, in step 4, the normalized distance D k It can be represented as:

[0029]

[0030]

[0031] in, Represents a single building unit and scene centroid distance, and Representing the centroid of the scene respectively The coordinates are given by disMAX, which represents the maximum distance between a single building in the scene and the scene's centroid.

[0032] Secondly, direction A k It can be represented as:

[0033]

[0034]

[0035]

[0036] in, This indicates the calculation of the arctangent function.

[0037] Furthermore, in step 5, the polar coordinate system subregion It can be represented as:

[0038]

[0039]

[0040] In the formula, This represents a set of sub-regions obtained using distance and angle intervals. To represent a certain sub-region, [0.1*(t) d -1), 0.1*t d ] represents the distance range, 0.1 is the quantization distance interval, t d The separable distance interval number; [Δθ*(t)] θ -1),Δθ*t θ ] indicates the angle range, t θ The interval number is the separable angular interval number, and int() is used for integer rounding.

[0041] Furthermore, in step 6, the individual building units It is determined to belong to a certain sub-region The condition criteria are:

[0042]

[0043] Among them, (D) k A k (a single building) In polar coordinate system The coordinates.

[0044] Furthermore, in step 7, scenario re i Statistical histogram It can be represented as:

[0045]

[0046] in, To represent a two-dimensional histogram, t d ,t θ This indicates the center of the two-dimensional coordinates of the histogram.

[0047] Furthermore, in step 8, the scenario R to be matched... m It can be represented as:

[0048] R m ={ra1,ra2,...,ra m}

[0049] In the formula, m represents the number of all scenes in the entire city, and ra j The scenario with ID j can be represented as:

[0050]

[0051] in, Representing scene ra j The k-th building entity contains two attributes: the building entity center. and area N j This represents the number of individual buildings.

[0052] Furthermore, in step 9, scene ra j Statistical histogram It can be represented as:

[0053]

[0054] in, To represent a two-dimensional histogram, t d ,t θ This indicates the center of the two-dimensional coordinates of the histogram. Representing scene ra j subregion The number of individual buildings in the building.

[0055] Furthermore, in step 10, scene ra j With scene re i Statistical histogram of differences It can be represented as:

[0056]

[0057] Furthermore, in step 11, scene ra j Best matching scenario It can be represented as:

[0058]

[0059] in, Represents the scene set R n any scene in re i .

[0060] Furthermore, in step 12, R m Matching scene set R m-n It can be represented as:

[0061]

[0062] The present invention has the following beneficial effects:

[0063] (1) This invention proposes a scene matching method that combines semantic information of ground features and topological features, which can quickly construct topological information within the entire urban scene and effectively achieve scene matching of urban blocks in a large area.

[0064] (2) The method of the present invention can effectively calculate the histogram features of individual buildings in each scene. The topological features are consistent in different spatial coordinate systems, providing effective technical support for semantic extraction and scene matching at the intelligent edge. Attached Figure Description

[0065] Figure 1 This is a schematic diagram illustrating the principle of a scene matching method that combines semantic information and topological features of ground features. Detailed Implementation

[0066] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0067] A scene matching method that combines semantic information of ground features and topological features, such as... Figure 1 As shown, it includes the following steps:

[0068] Step 1: Obtain a semantic map library of buildings and roads for the entire urban area, and divide the city into a set of street scenes R, using the main roads as boundaries. n The i-th scene can be represented as re i Each building in the scene contains detailed urban land use information; the street scene set R n It can be represented as:

[0069] R n ={re1,re2,...,re n}

[0070] In the formula, n represents the number of all scenes in the entire city, and re iThe scenario with number i can be represented as:

[0071]

[0072] in, Representing the scenario re i The k-th building entity contains t attributes, which are respectively N i This represents the number of individual buildings.

[0073] Step 2, utilize urban scene re i Each building unit center and area Attribute calculation scene centroid The calculation method is as follows:

[0074]

[0075]

[0076] in, For individual buildings The center coordinates, For individual buildings The area.

[0077] Step 3, using the scene centroid As the extreme point, the scene centroid The north direction is the polar axis direction, and the angle is taken as positive in the clockwise direction. Establish a polar coordinate system.

[0078] Step 4, calculate the re of the urban scene i Each building unit With the center of gravity of the scene Normalized distance D k and direction A k Normalized distance D k It can be represented as:

[0079]

[0080]

[0081] in, Represents a single building unit and scene centroid distance, and Representing the centroid of the scene respectively The coordinates are given by disMAX, which represents the maximum distance between a single building in the scene and the scene's centroid.

[0082] Secondly, direction A k It can be represented as:

[0083]

[0084]

[0085]

[0086] in, This indicates the calculation of the arctangent function.

[0087] Scene re i Each building unit in Both can be represented as a point in the polar coordinate system (D). k A k ).

[0088] Step 5, convert the entire polar coordinate system The regions are partitioned according to equal intervals and angular intervals Δθ, and the sub-region set can be denoted as follows: It consists of multiple sub-regions Composition; polar coordinate system subregion It can be represented as:

[0089]

[0090]

[0091] In the formula, This represents a set of sub-regions obtained using distance and angle intervals. To represent a certain sub-region, [0.1*(t) d -1), 0.1*t d ] represents the distance range, 0.1 is the quantization distance interval, t d The separable distance interval number; [Δθ*(t)] θ -1),Δθ*t θ ] indicates the angle range, t θ The interval number is the separable angular interval number, and int() is used for integer rounding.

[0092] Step 6, based on the scenario re i All individual buildings in the polar coordinate system The coordinates in the data are used to statistically analyze the sub-region sets. Each sub-region Number of individual buildings Building unit It is determined to belong to a certain sub-region The condition criteria are:

[0093]

[0094] Among them, (D) k A k (a single building) In polar coordinate system The coordinates.

[0095] Step 7: Based on the number of individual buildings in each sub-region, information about the entire scene can be obtained. i Statistical histogram It can be represented as:

[0096]

[0097] in, To represent a two-dimensional histogram, t d ,t θ This indicates the center of the two-dimensional coordinates of the histogram. Representing the scenario re i subregion The number of individual buildings in the building.

[0098] Step 8, obtain the set of scenarios to be matched R m , can be represented as:

[0099] R m ={ra1,ra2,...,ra m}

[0100] In the formula, m represents the number of all scenes in the entire city.

[0101] The j-th scene can be represented as ra j This scene also includes information about individual buildings, but the absolute error of their geographic coordinates is large, and it only contains two attributes, which can be represented as:

[0102]

[0103] in, Representing scene ra j The k-th building entity contains two attributes: the building entity center. and area N j This represents the number of individual buildings.

[0104] Step 9, for scene ra j The statistical histogram of a building can be calculated using the methods in steps 2 through 7. It can be represented as:

[0105]

[0106] in, To represent a two-dimensional histogram, t d ,t θ This indicates the center of the two-dimensional coordinates of the histogram. Representing scene ra j subregion The number of individual buildings in the building.

[0107] Step 10, using scene ra j As a scenario to be matched, based on the statistical histogram difference... Calculate its relationship with the scene set R n any scene in China i Similarity, scene ra j With scene re i Statistical histogram of differences It can be represented as:

[0108]

[0109] Step 11: Find the scene ra based on the statistical histogram difference. j The best matching scenario is It can be represented as:

[0110]

[0111] in, Represents the scene set R n any scene in re i .

[0112] Step 12, finally, the scene to be matched, R, can be obtained. m Matching scenario set R m-n , can be represented as:

[0113]

[0114] In summary, the scene matching method proposed in this invention combines semantic information and topological features of ground objects. This method constructs the topological structural features of ground objects in the scene and uses histogram matching to find the best matching scene. This method is not affected by the scale of the coordinate systems of the two data sources, and provides technical support for convenient, fast and accurate urban block scene matching.

Claims

1. A scene matching method combining semantic information of ground features and topological features, characterized in that, Includes the following steps: Step 1: Obtain a semantic map library of buildings and roads for the entire urban area, and divide the city into a set of street scenes R, using the main roads as boundaries. n The i-th scene is represented as re i Each building in the scene contains detailed urban land use information; Step 2, utilize urban scene re i Each building unit center and area Attribute calculation scene centroid Step 3, using the scene centroid As the extreme point, the scene centroid The north direction is the polar axis direction, and the angle is taken as positive in the clockwise direction. Establish a polar coordinate system. Step 4, calculate the re of the urban scene i Each building unit With the center of gravity of the scene Normalized distance D k and direction A k Then the scenario re i Each building unit in All in polar coordinate system U rei In this context, it is represented as a point (D). k A k ); Step 5, convert the entire polar coordinate system The regions are divided according to equal intervals and angular intervals Δθ, and the sub-regions are denoted as follows: It consists of multiple sub-regions constitute; Step 6, based on the scenario re i All individual buildings in the polar coordinate system The coordinates in the data are used to statistically analyze the sub-region sets. Each sub-region Number of individual buildings Step 7: Based on the number of individual buildings in each sub-region, obtain information about the entire scene. i Statistical histogram Step 8, obtain the set of scenarios to be matched R m The j-th scene is denoted as ra. j The scene also includes information about individual buildings, but the absolute error of their geographic coordinates is large, and it only contains two attributes. Step 9, for scene ra j Calculate the statistical histogram for each individual building using steps 2 through 7. Step 10, using scene ra j As a scenario to be matched, based on the statistical histogram difference... Calculate its relationship with the scene set R n any scene in China i Similarity; Step 11: Find the scene ra based on the statistical histogram difference. j The best matching scenario is Step 12, obtain the scene to be matched R m Matching scenario set R m-n .

2. The scene matching method combining semantic information and topological features of ground features according to claim 1, characterized in that, Step 1 Street scene set R n Represented as: R n ={re1,re2,...,re n } In the formula, n represents the number of all scenes in the entire city, and re i Represents the scenario numbered i, re i Represented as: in, Representing the scenario re i The k-th building unit contains t attributes, which are respectively N i This represents the number of individual buildings.

3. The scene matching method combining semantic information and topological features according to claim 2, characterized in that, Scene centroid in step 2 The calculation method is as follows: in, For individual buildings The center coordinates of are x and y respectively. k y k , For individual buildings The area.

4. The scene matching method combining semantic information and topological features according to claim 3, characterized in that, The normalized distance D in step 4 k Represented as: in, Represents a single building unit and scene centroid distance, and Representing the centroid of the scene respectively The coordinates, disMAX, represent the maximum distance between a single building in the scene and the scene's centroid; Direction A k Represented as: Here, actan() represents the arctangent function.

5. The scene matching method combining semantic information and topological features of ground features according to claim 4, characterized in that, In step 5, the polar coordinate system U rei subregion Represented as: In the formula, This represents a set of sub-regions obtained using distance and angle intervals. To represent a certain sub-region, [0.1*(t) d -1), 0.1*t d ] represents the distance range, 0.1 is the quantization distance interval, t d The separable distance interval number; [Δθ*(t)] θ -1),Δθ*t θ ] indicates the angle range, t θ The interval number is the separable angular interval number, and int() is used for integer rounding.

6. The scene matching method combining semantic information and topological features of ground features according to claim 5, characterized in that, Individual buildings in step 6 It is determined to belong to a certain sub-region The condition criteria are: Among them, (D) k A k (a single building) In polar coordinate system The coordinates in the diagram.

7. The scene matching method combining semantic information and topological features of ground features according to claim 6, characterized in that, In step 7, the scene re i Statistical histogram Represented as: in, To represent a two-dimensional histogram, t d ,t θ This indicates the center of the two-dimensional coordinates of the histogram. Representing the scenario re i subregion The number of individual buildings in the building.

8. The scene matching method combining semantic information and topological features of ground features according to claim 7, characterized in that, In step 9, scene ra j Statistical histogram Represented as: in, To represent a two-dimensional histogram, t d ,t θ This indicates the center of the two-dimensional coordinates of the histogram. Representing scene ra j subregion The number of individual buildings in the building.

9. The scene matching method combining semantic information and topological features of ground features according to claim 8, characterized in that, Scene ra in step 10 j With scene re i Statistical histogram of differences Represented as:

10. The scene matching method combining semantic information of ground features and topological features according to claim 9, characterized in that, In step 11, scene ra j Best matching scenario Represented as: in, Represents the scene set R n any scene in re i .

Images