Parallel road intelligent identification method and system based on generative labels

Abstract

The invention relates to the technical field of road network data processing, in particular to a parallel road intelligent identification method and system based on generative labels. The target area road network features are input into a pre-trained parallel road recognition model, parallel roads of the target area road network are obtained through the parallel road recognition model, and the parallel road recognition model is trained based on training data by adopting a semi-supervised method; labels are automatically generated for the road sample data set in the training data according to a preset rule. The parallel road recognition accuracy of the model is improved by combining road network geometric features and semantic features, training data is automatically generated by using the generative labels, the labor cost is reduced, the data processing efficiency is improved, the dependence on manual annotation data is reduced, the parallel road recognition efficiency and accuracy can be improved, and the real-time performance of the model is improved. The method can also be expanded to related fields such as urban functional area division and traffic flow prediction, and has a good application prospect.

Description

Technical Field

[0001] The present invention relates to the technical field of road network data processing, and in particular to a method and system for intelligently identifying parallel roads based on generative labels. Background Art

[0002] As the skeleton of the urban road network, accurate identification of parallel roads is crucial for understanding the urban spatial structure, optimizing traffic flow distribution, and improving the accuracy of map data. However, with the acceleration of urbanization, the complexity of urban road networks continues to increase. Existing methods based on manual identification or simple geometric feature matching can no longer meet the requirements of high precision and high efficiency. In addition, there are also automated methods, but they rely on a large amount of manually annotated training data, which is not only costly, but also has limited generalization ability and is difficult to adapt to diverse urban environments. That is, the main problems of existing parallel road recognition technology can be summarized as follows: 1. Difficulty in geometric feature recognition: Urbanization leads to subtle road geometric features, and traditional recognition methods based on geometric features have difficulty in distinguishing parallel roads from non-parallel roads. 2. Complex multi-factor influence: Parallel road recognition is affected by many factors, such as road width, number of lanes, traffic flow, etc. The combined effect of these factors complicates the recognition process. 3. Dependence on manually annotated data: Supervised learning methods rely on manually annotated training data, which not only consumes a lot of manpower and time, but also limits the generalization ability of the model in different urban environments. 4. Lack of effective semantic feature utilization: Existing methods fail to fully utilize the semantic information of roads, such as road names, functions, and surrounding points of interest (POIs), which are of great value in improving recognition accuracy. 5. Limited model generalization ability: Existing models perform well on specific data sets, but their performance deteriorates on new data that have not been seen, and they lack good generalization ability. Therefore, there is an urgent need for a solution that can identify parallel roads in spatial data of different scales to meet the needs of intelligent management of urban road networks. Summary of the invention

[0003] To this end, the present invention provides a method and system for intelligently identifying parallel roads based on generative labels to solve the problems of insufficient accuracy and efficiency of existing manual or simple geometric feature matching, and the difficulty of automated methods in adapting to the needs of diverse urban environments.

[0004] According to the design scheme provided by the present invention, on the one hand, a method for intelligently identifying parallel roads based on generative labels is provided, comprising:

[0005] Extract road network features of the target area;

[0006] The target area road network features are input into a pre-trained parallel road recognition model, and the parallel road recognition model is used to obtain the parallel roads of the target area road network. The parallel road recognition model is based on training data and uses a semi-supervised method for model training. The training data automatically generates labels for road sample data sets according to preset rules.

[0007] As the parallel road intelligent identification method based on generative labels of the present invention, further, the road network features of the target area are extracted, including:

[0008] The road network of the target area is converted into a graph structure composed of road meshes, and multi-dimensional road network features are extracted. The road mesh represents a closed area surrounded by road segments. The multi-dimensional road network features include geometric features for describing the geometric morphology of the road mesh and semantic features for describing the distribution of interest points inside the road mesh.

[0009] As a parallel road intelligent identification method based on generative labels of the present invention, further, the geometric features include: width, aspect ratio, compactness, area and perimeter of the road mesh.

[0010] As the parallel road intelligent recognition method based on generative labels of the present invention, further, model training is performed based on training data and using a semi-supervised method, including:

[0011] Acquire a road sample data set, convert a road network in the road sample data set into a graph structure composed of road meshes, and extract road network features from a geometric dimension and a semantic dimension, wherein the road network features include geometric features for describing the geometric morphology of the road meshes and semantic features for describing the distribution of interest points inside the road meshes;

[0012] Automatically generate parallel road labels or non-parallel road labels for the road network in the road sample data set according to preset rules, and generate training data, wherein the preset rules are road construction standards and road construction field knowledge;

[0013] The parallel road recognition model is trained based on the training data using a semi-supervised learning method.

[0014] As the parallel road intelligent identification method based on generative labels of the present invention, further, automatically generating parallel road labels or non-parallel road labels according to preset rules includes:

[0015] Setting parallel road recognition parameters according to road construction standards and road construction field knowledge, wherein the parallel road recognition parameters include: parallel road mesh width range, interest point distribution and aspect ratio threshold;

[0016] Based on the road network features in the road sample data set and in accordance with the parallel road identification parameters, parallel road labels or non-parallel road labels are generated for the road network in the road sample data set.

[0017] As a parallel road intelligent identification method based on generative labels of the present invention, further, the parallel road identification model is constructed based on a topological adaptive graph neural network, and the model is trained using a cross entropy loss function.

[0018] As a parallel road intelligent identification method based on generative labels of the present invention, further, the parallel road identification model includes: an input layer, a hidden layer and an output layer, wherein the hidden layer includes 3 topological adaptive graph neural network convolutional layers and 1 fully connected layer to encode the input road network graph structural features into feature vectors and generate feature classification probabilities.

[0019] In another aspect, the present invention further provides a parallel road intelligent identification system based on generative labels, comprising: a road feature extraction module and a road feature identification module, wherein:

[0020] A road feature extraction module is used to extract the road network features of the target area;

[0021] The road feature recognition module is used to input the characteristics of the target area road network into a pre-trained parallel road recognition model, and use the parallel road recognition model to obtain the parallel roads of the target area road network. The parallel road recognition model is based on training data and uses a semi-supervised method for model training. The training data automatically generates labels for road sample data sets according to preset rules.

[0022] Beneficial effects of the present invention:

[0023] The present invention utilizes generative tags to automatically generate training data, reduces labor costs, improves the efficiency of data processing, reduces dependence on manually labeled data, combines road network geometric features and semantic features to improve the model's accuracy in identifying parallel roads, uses a topological adaptive graph neural network TAGCN to build a recognition model, enhances the model's adaptability to road networks in different urban environments, introduces point of interest POI data, and makes full use of road semantic information to assist in parallel road recognition, which can not only improve the efficiency and accuracy of parallel road recognition, but also be extended to other related fields, such as urban functional area division, traffic flow prediction, etc., to provide technical support for intelligent urban management, and has good application prospects in urban planning and design, intelligent traffic management, geographic information system update, navigation system of autonomous driving vehicles, and automated analysis of urban road networks. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a schematic diagram of the intelligent identification process of parallel roads based on generative labels in the embodiment;

[0025] Figure 2 It is a schematic diagram of the construction of the graph structure of the road network in the embodiment;

[0026] Figure 3 It is a schematic diagram for explaining the characteristic items of the road network in the embodiment;

[0027] Figure 4 The figure is a schematic diagram of the superposition of vector diagrams and POIs of some regions in the embodiment, which is used to illustrate the rationality of generating labels;

[0028] Figure 5 It is a schematic diagram of a parallel road recognition framework based on TAGCN in the embodiment;

[0029] Figure 6 It is a schematic diagram of experimental data in the embodiment;

[0030] Figure 7 Schematic diagram of parallel road recognition results of experimental data in the embodiment. DETAILED DESCRIPTION

[0031] In order to make the purpose, technical solutions and advantages of the present invention clearer and more understandable, the present invention is further described in detail below in conjunction with the accompanying drawings and technical solutions.

[0032] Traditional supervised learning algorithms rely on a large amount of manually annotated data, which is not only time-consuming and labor-intensive, but also limits the generalization ability of the model. Most existing technologies focus on analyzing the geometric features of roads, while ignoring the importance of semantic information. Figure 1 As shown, a method for intelligently identifying parallel roads based on generative labels is provided, comprising:

[0033] S101, extracting road network features of the target area.

[0034] Specifically, the road network of the target area can be converted into a graph structure composed of road meshes, and multi-dimensional road network features can be extracted. The road mesh represents a closed area surrounded by road segments. The multi-dimensional road network features include geometric features for describing the geometric morphology of the road mesh and semantic features for describing the distribution of interest points inside the road mesh.

[0035] like Figure 2 and 3 As shown in the figure, a "polygonization" operation is performed on the road network to convert it into a graph structure consisting of road meshes. Each mesh represents a closed area surrounded by road segments. If two meshes share common edges or intersections, they are considered to be connected, thus constructing a graph structure of road meshes.

[0036] The features of the mesh are extracted from two dimensions: geometry and semantics. Geometric features may include the width, aspect ratio, compactness, area, and perimeter of the mesh. These features can describe the geometry of the mesh and provide basic information for recognition. Semantic features can obtain the semantic information of the mesh by analyzing the distribution of points of interest (POIs) inside the mesh. For example, based on whether there are POIs of building types inside the mesh, a feature vector is constructed to describe the functional semantics of the mesh. By introducing POI data related to the road mesh, geometric and semantic dual features are extracted to provide a richer and more accurate feature description for parallel road recognition.

[0037] S102. Input the target area road network characteristics into a pre-trained parallel road recognition model, and use the parallel road recognition model to obtain the parallel roads of the target area road network. The parallel road recognition model is based on training data and uses a semi-supervised method to perform model training. The training data automatically generates labels for road sample data sets according to preset rules.

[0038] Specifically, the model training is performed based on the training data and using a semi-supervised method, which can be designed to include:

[0039] Acquire a road sample data set, convert a road network in the road sample data set into a graph structure composed of road meshes, and extract road network features from a geometric dimension and a semantic dimension, wherein the road network features include geometric features for describing the geometric morphology of the road meshes and semantic features for describing the distribution of interest points inside the road meshes;

[0040] Automatically generate parallel road labels or non-parallel road labels for the road network in the road sample data set according to preset rules, and generate training data, wherein the preset rules are road construction standards and road construction field knowledge;

[0041] The parallel road recognition model is trained based on the training data using a semi-supervised learning method.

[0042] The automatic generation of parallel road labels or non-parallel road labels according to preset rules includes:

[0043] Setting parallel road recognition parameters according to road construction standards and road construction field knowledge, wherein the parallel road recognition parameters include: parallel road mesh width range, interest point distribution and aspect ratio threshold;

[0044] Based on the road network features in the road sample data set and in accordance with the parallel road identification parameters, parallel road labels or non-parallel road labels are generated for the road network in the road sample data set.

[0045] like Figure 4As shown, labels for training data are automatically generated based on road construction standards and domain knowledge. Width range: According to road design specifications, the width range of the parallel lines road PLR mesh is set, such as 7.5 meters to 15 meters. POI distribution: The distribution of POIs inside the mesh is used to assist in determining whether the mesh is a PLR. Aspect ratio: A threshold is set to further determine whether the mesh is a PLR based on the aspect ratio of the mesh. Through the above criteria, PLR meshes and non-PLR meshes are automatically labeled to generate training data with generative labels. Through rules based on road construction standards and domain knowledge, training data with generative labels are automatically generated, which significantly reduces the workload of manual labeling while improving the diversity and coverage of training samples.

[0046] The parallel road recognition model can be constructed based on the topological adaptive graph neural network TAGCN, and the model can be trained using the cross entropy loss function. TAGCN can adapt to the complex and changeable connection conditions of the graph node degree, and is suitable for processing the graph structure data of the road network in the embodiment of this case. TAGCN is used to construct the recognition model. The network can adapt to the graph data with complex and changeable node connections, and compared with other graph neural networks, it shows better performance when processing such data. The use of TAGCN improves the model's recognition ability for complex road networks. Among them, the model structure can be composed of an input layer, a hidden layer, and an output layer. The hidden layer contains 3 TAGCN convolutional layers and 1 fully connected layer, which is responsible for encoding the input graph data into a feature vector and outputting the node classification probability. Activation function: Leaky ReLU is selected as the activation function to alleviate the gradient vanishing problem. Loss function: The cross entropy loss function is selected to increase the model's attention to minority categories. Optimizer: The Adam optimizer is selected for model training.

[0047] During the training process, the model performance is optimized by adjusting hyperparameters (such as learning rate, number of filter layers, number of hidden layer channels, etc.). After training, the model is applied to road meshes in other regions to identify parallel roads, and the recognition effect is evaluated based on indicators such as accuracy, precision, recall, and F1 value.

[0048] Furthermore, based on the above method, an embodiment of the present invention also provides a parallel road intelligent identification system based on generative labels, comprising: a road feature extraction module and a road feature identification module, wherein:

[0049] A road feature extraction module is used to extract the road network features of the target area;

[0050] The road feature recognition module is used to input the characteristics of the target area road network into a pre-trained parallel road recognition model, and use the parallel road recognition model to obtain the parallel roads of the target area road network. The parallel road recognition model is based on training data and uses a semi-supervised method for model training. The training data automatically generates labels for road sample data sets according to preset rules.

[0051] In order to verify the effectiveness of this solution, the following is a further explanation based on experimental data:

[0052] like Figure 5 and 6 As shown in Figure 2, experiments were conducted on real road datasets in City A and City B to verify the effectiveness of this solution. Figure 7 As shown in the figure, the experimental results show that the scheme in this case has shown good performance in parallel road recognition tasks in different regions, with an accuracy rate of more than 93%. Compared with the unsupervised method, the scheme in this case has significantly improved the recognition accuracy. Even compared with supervised methods (such as SVM and GCN), the model in this case has advantages in performance and does not need to rely on labeled data.

[0053] The above data show that this solution reduces the need for manual labeling and reduces manpower and time costs. By combining geometric features and semantic information, the accuracy of parallel road recognition is improved. The TAGCN model and semi-supervised learning framework are used to enhance the model's adaptability to road networks in different urban environments. The efficiency of model training and prediction is improved by using automatically generated training data and optimized model structure. By automatically generating training data labels, manual intervention is reduced and the impact of subjective bias on model performance is reduced, making this solution not only suitable for parallel road recognition, but also can be extended to other geographic information processing tasks, such as urban functional area division, traffic flow prediction, etc.

[0054] Unless otherwise specifically stated, the relative steps, numerical expressions and values ​​of the components and steps set forth in these embodiments do not limit the scope of the present invention.

[0055] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0056] The units and method steps of each example described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. A person of ordinary skill in the art may use different methods to implement the described functions for each specific application, but such implementation is not considered to be beyond the scope of the present invention.

[0057] Those skilled in the art will appreciate that all or part of the steps in the above method can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, such as a read-only memory, a disk or an optical disk. Optionally, all or part of the steps in the above embodiment can also be implemented using one or more integrated circuits, and accordingly, each module / unit in the above embodiment can be implemented in the form of hardware or in the form of software function modules. The present invention is not limited to any specific form of combination of hardware and software.

[0058] Finally, it should be noted that the above-described embodiments are only specific implementations of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The protection scope of the present invention is not limited thereto. Although the present invention is described in detail with reference to the above-described embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-described embodiments within the technical scope disclosed by the present invention, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A method for intelligently identifying parallel roads based on generative labels, characterized in that: Include: Extract road network features of the target area; The target area road network features are input into a pre-trained parallel road recognition model, and the parallel road recognition model is used to obtain the parallel roads of the target area road network. The parallel road recognition model is based on training data and uses a semi-supervised method for model training. The training data automatically generates labels for road sample data sets according to preset rules.

2. The method for intelligently identifying parallel roads based on generative labels according to claim 1 is characterized in that: Extract the road network features of the target area, including: The road network of the target area is converted into a graph structure composed of road meshes, and multi-dimensional road network features are extracted. The road mesh represents a closed area surrounded by road segments. The multi-dimensional road network features include geometric features for describing the geometric morphology of the road mesh and semantic features for describing the distribution of interest points inside the road mesh.

3. The method for intelligently identifying parallel roads based on generative labels according to claim 2 is characterized in that: The geometric features include: width, aspect ratio, compactness, area and perimeter of the road mesh.

4. The method for intelligently identifying parallel roads based on generative labels according to claim 1, characterized in that: Model training is performed based on training data using a semi-supervised approach, including: Acquire a road sample data set, convert a road network in the road sample data set into a graph structure composed of road meshes, and extract road network features from a geometric dimension and a semantic dimension, wherein the road network features include geometric features for describing the geometric morphology of the road meshes and semantic features for describing the distribution of interest points inside the road meshes; Automatically generate parallel road labels or non-parallel road labels for the road network in the road sample data set according to preset rules, and generate training data, wherein the preset rules are road construction standards and road construction field knowledge; The parallel road recognition model is trained based on the training data using a semi-supervised learning method.

5. The method for intelligently identifying parallel roads based on generative labels according to claim 4 is characterized in that: Automatically generate parallel road labels or non-parallel road labels based on preset rules, including: Setting parallel road recognition parameters according to road construction standards and road construction field knowledge, wherein the parallel road recognition parameters include: parallel road mesh width range, interest point distribution and aspect ratio threshold; Based on the road network features in the road sample data set and in accordance with the parallel road identification parameters, parallel road labels or non-parallel road labels are generated for the road network in the road sample data set.

6. The method for intelligently identifying parallel roads based on generative labels according to claim 1 or 4, characterized in that: The parallel road recognition model is constructed based on a topological adaptive graph neural network, and the model is trained using a cross entropy loss function.

7. The method for intelligently identifying parallel roads based on generative labels according to claim 6 is characterized in that: The parallel road recognition model includes: an input layer, a hidden layer and an output layer, wherein the hidden layer includes three topology adaptive graph neural network convolutional layers and one fully connected layer to encode the input road network graph structural features into feature vectors and generate feature classification probabilities.

8. A parallel road intelligent identification system based on generative labels, characterized in that: Contains: road feature extraction module and road feature recognition module, in, A road feature extraction module is used to extract the road network features of the target area; The road feature recognition module is used to input the characteristics of the target area road network into a pre-trained parallel road recognition model, and use the parallel road recognition model to obtain the parallel roads of the target area road network. The parallel road recognition model is based on training data and uses a semi-supervised method for model training. The training data automatically generates labels for road sample data sets according to preset rules.

9. An electronic device, characterized in that: include: at least one processor, and a memory coupled to the at least one processor; The memory stores a computer program, and the computer program can be executed by the at least one processor to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed, the method according to any one of claims 1 to 7 can be implemented.

Images