CN111611668A - An automatic road network selection method considering geometric features and semantic information - Google Patents

Abstract

The present invention discloses a road network automatic selection method considering geometric features and semantic information, comprising: firstly, constructing a road mesh and a road stroke; then, establishing a relationship between the road mesh, the road stroke and POI data and taxi trajectory data respectively Finally, with the help of logical operations, the selection results of the mesh model and the stroke model, which comprehensively consider the semantic information, are integrated. The selection results of this method fully consider the semantic information of the road network, maintain the overall structure and local structure of the road, and the selection results are more valuable and scientific in practical applications.

Description

An automatic road network selection method considering geometric features and semantic information

technical field

The invention belongs to the technical field of cartography and geographic information systems, and relates to an automatic road network selection method considering geometric features and semantic information.

Background technique

Automatic map synthesis is a challenging problem in cartography. In the digital environment, the use of map synthesis technology to export large-scale maps in spatial databases into small-scale maps is still the focus and difficulty of theoretical research on spatial information science. As the basic geographic data, the road network is of great significance to the national economy and national security. The automatic selection of road network has always been one of the main contents of comprehensive map research.

At present, the achievements in this neighborhood are mainly divided into two categories: (1) Considering the selection method of the geometric characteristics of the road network, there are mainly: using the feature points of the road to select the road network; according to the principle of good continuity, using the "stroke" model Selecting the road network according to the importance of the road by considering the thematic attributes of the road; using the density characteristics of the road to select the road according to the sparseness of the road; In contrast, it gets rid of the disadvantage of not taking into account the connectivity and regionality in the selection work, but does not consider semantic information in the road network selection work. (2) Combined with the selection method of semantic information, on the basis of comprehensively considering the topological characteristics of the road network, geographic big data is introduced as semantic information to realize the automatic selection of the road network. This kind of method only uses a single semantic data, and the selection results are difficult to be comprehensive. Reflect road semantic information.

POI data and taxi trajectory data are important sources of geographic information. POI records the location and attribute information of important facilities within a geographic information city, and the information of facilities on both sides of the road is related to the cognition of road importance. The road traffic flow formed by the taxi trajectory data reflects the importance of the road to a certain extent. To this end, this paper takes POI data and taxi trajectory data as semantic information, and proposes an automatic road network selection method that comprehensively considers road geometric features and semantic information. By constructing the relationship model between road mesh, road stroke and POI data and taxi trajectory data respectively, this method realizes that the selection result fully maintains the overall structure and local structure of the road network while taking into account the semantic information of the road network.

SUMMARY OF THE INVENTION

Aiming at the shortcomings of the existing road network selection methods, the invention proposes a road network selection method that takes into account geometric features and semantic information. The method combines the advantages of mesh model and stroke model in road selection, and makes up for the single model in road selection. In the road selection, POI data, taxi trajectory data and space syntax attributes are added, which effectively avoids the defects caused by single semantic data in road selection. The method of the present invention includes three parts: constructing a mesh and stroke model, establishing a relationship model between the model and semantic information, and integrating the selection results of the mesh model and the stroke model. Figure 1 shows the overall flow of the road network selection method of the present invention.

The construction of the mesh model is mainly to divide a series of small areas according to the density of the road network. By determining the density threshold and using attributes such as geometry, the road sections in the area are selected. The process is shown in Figure 2. In Figure 2, the grid with a density of 0.21 is the largest among all grids. If it is greater than the threshold, the mesh with a density of 0.21 is merged with the surrounding mesh with a high density, and the middle road segment L is deleted to generate a new density of 0.08. mesh. Follow this iterative method until the mesh density of all meshes is less than the threshold.

The construction of the stroke model is mainly based on the good continuity principle in Gestalt visual perception, constructs the network topology structure for the road segment, generates the relative edges and nodes, sorts the angles generated by the edges in order, and judges whether the minimum angle is less than the threshold. If it is less than, the two road segments that generate the included angle will be merged as a stroke, as shown in Figure 3. By sorting the importance of each stroke from high to low, select the road with high importance.

The method of the present invention proposes a new road network selection method while taking into account the road geometric characteristics and taking into account the road semantic information. First, the road mesh and road stroke are constructed from the road network data. Secondly, establish the relationship model between mesh model, stroke model and space syntax, traffic flow and POI data. Finally, perform logical operations on the results selected by the mesh model and the stroke model to obtain the final selection result. The experimental results show that the present invention can retain the main road, maintain the overall structure and local structure of the road, better take into account the geometric characteristics and semantic information of the road, and the selection result has a high consistency with the reference benchmark road.

Description of drawings

Figure 1 Basic framework

Figure 2 Mesh density construction method

Figure 3 Construction of the road stroke process

Detailed ways

In order to describe in detail the technical content, the selection model, the achieved purpose and the achieved effect of the present invention, the following detailed description is given in conjunction with the specific embodiments.

Step 1: Construct the road mesh and determine the mesh density threshold. The grid area is generated from the road network, and the grid is the road mesh. Through the statistics of the mesh density of the basic road network and the reference road network, it is found that there is a correlation between the two, and the appropriate mesh density is determined to be 0.006m/m ² .

Step 2: Sort in reverse order according to the road mesh density. Then, according to the sorting result, it is judged whether the mesh density is less than the set threshold in turn, and if it is less than the set threshold, the mesh is selected.

Step 3: If the road mesh density is greater than the set threshold, sort the mesh density in descending order. Select the largest mesh in the sorted set to merge with the adjacent mesh with the highest density, and determine whether the combined threshold is greater than the set threshold. If it is greater than the set threshold, continue to merge with the adjacent mesh with the largest density until the combined threshold is less than Set the threshold.

Step 4: Calculate the number of POIs in the road mesh, and sort them in descending order according to the POI number value. According to the sorted order, select the meshes that meet the threshold in turn.

Step 5: Build a road stroke based on the principle of geometry perception.

Step 6: Introduce taxi trajectory data, and use taxi trajectory data to characterize road traffic flow to assist road network selection. The road network is selected according to the principle of space syntax and the road importance score formed by the road traffic flow.

Step 7: Integrate the mesh model selection result and the stroke model selection result through logical operations to obtain the final selection result.

To sum up, the present invention is an automatic road network selection method considering geometric features and semantic information. First, the road mesh and the road stroke are constructed from the road network data. Secondly, establish the relationship model between mesh model, stroke model and space syntax, traffic flow and POI data. Finally, perform logical operations on the results selected by the mesh model and the stroke model to obtain the final selection result. The present invention is effective and feasible in road network selection, and its main advantages are:

(1) Combining the advantages of mesh model and stroke model in road selection, it makes up for the shortcomings of a single model in road selection;

(2) POI data, taxi trajectory data and space syntax attributes are added to road selection, which effectively avoids the defects caused by single semantic data in road selection.

(3) A new approach is provided for the semantic measurement of road network selection. Meanwhile, POI data and taxi trajectory data are time-sensitive, and the road network selected by the present invention is more valuable and scientific in practical application.

Claims (3)

1. A road network automatic selection method taking into account geometric features and semantic information, comprising two parts: road network selection of a mesh model taking into account semantic information and road network selection of a stroke model taking into account semantic information: Road network selection of mesh model considering semantic information: Step 1: Construct road mesh and select it according to the road mesh density threshold; Step 2: Calculate the number of POIs in the road mesh, and sort them in descending order according to the POI number value. Select the meshes that meet the threshold in turn according to the sorted order. Road network selection of stroke model considering semantic information: Step 3: Build a road stroke based on the principle of geometric perception; Step 4: The road stroke needs to be selected within the buffer range of important POI facilities. Step 5: Introduce taxi trajectory data, and use taxi trajectory data to characterize road traffic flow to assist road network selection. The road network is selected according to the principle of space syntax and the road importance score formed by the road traffic flow. Step 6: Integrate the selection result of the mesh model and the selection result of the stroke model through logical operations to obtain the final selection result. 2. A kind of road network automatic selection method considering geometrical feature and semantic information according to claim 1, is characterized in that, in Step2, utilizes the mesh model that Step1 step generates to establish semantic relation model. 3. according to claim 1 or claim 2, a kind of road network automatic selection method taking into account geometric feature and semantic information is characterized in that, in step Step6, completes the selection of road network, also comprises: as taking into account semantics The road network selection results of the mesh model with information and the road network selection results of the stroke model considering semantic information.

Images