Road network generation method and device, equipment and storage medium

Abstract

The invention discloses a road network generation method and device, equipment and a storage medium. The method comprises the following steps: extracting a network topology structure of a road center line segment in a preset area from a road image; gradually superposing the road center lines from the longest road center line to form a road network based on the network topology structure according to the line segment length of the road center lines, and calculating the structural difference of the road network before and after each superposition; and when the structural difference is smaller than an iteration termination threshold value, stopping the superposition operation of the road center lines, and taking a road network obtained by the last superposition operation as a road center line network. According to the embodiment of the invention, the road center line network is constructed step by step according to the line segment length of the road center line and the network difference, the completeness and accuracy of the road are well balanced, the computing power requirement is reduced, and the method is suitable for large-area engineering application.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a road network generation method, device, equipment and storage medium. Background Art

[0002] Roads are an important part of a national or regional transportation system and an important basic geographic information data. How to quickly and accurately identify the centerline of a road is of great significance to the management, maintenance and application of basic transportation databases. Since the 21st century, with the popularization of high-resolution remote sensing images, domestic and foreign scholars have conducted in-depth research on road extraction based on high-resolution remote sensing images.

[0003] In the prior art, the road centerline network extraction method is a road analysis method that integrates multi-source data: by integrating or integrating multi-level road features or other geographic data (such as lidar data, synthetic aperture radar data, vehicle trajectory data, etc.), the fused road extraction results are obtained. This method requires the fusion of multi-source data, has high requirements on computing power, and is not suitable for large-area engineering applications. Summary of the invention

[0004] Based on this, the present invention provides a road network generation method, device, equipment and storage medium, which can gradually construct a road centerline network according to the line segment length and network differences of the road centerline, thereby achieving a good balance between road integrity and accuracy. It only needs to superimpose the road centerlines extracted from the image, and there is no need to process multi-source data, which reduces computing power requirements and is suitable for large-area engineering applications.

[0005] To achieve the above object, an embodiment of the present invention provides a road network generation method, comprising:

[0006] Extracting the network topology of the road centerline in a preset area from the road image;

[0007] Based on the network topology, the road centerline is gradually superimposed on the road network in the order of line segment length from large to small, and the structural difference of the road network before and after each superposition is calculated;

[0008] When the structural difference is less than the iteration termination threshold, the superposition operation of the road centerline is stopped, and the road network obtained by the last superposition operation is used as the road centerline network.

[0009] As an improvement to the above solution, an optimization operation is performed on the road centerline network, and the optimization operation includes at least one of the following methods:

[0010] Deleting open line segments of the road centerline network;

[0011] Deleting a small closed loop of the road centerline network; wherein the road centerline network includes a large closed loop and a small closed loop, and the road length of the small closed loop is less than the road length of the adjacent large closed loop;

[0012] Eliminate burrs in the road centerline network.

[0013] As an improvement of the above solution, the calculation of the structural difference of the road network before and after each superposition includes:

[0014] Calculate the length of the road network before and after each overlay;

[0015] For each superposition operation, the difference between the length of the road network after superposition and the length of the road network before superposition is divided by the length of the road network before superposition to obtain the structural difference of the road network before and after each superposition.

[0016] As an improvement of the above solution, the iteration termination threshold is determined by the following method:

[0017] Before each superposition, the iteration termination threshold is obtained by subtracting the current cumulative probability from one; wherein the current cumulative probability is equal to the length of the current road network divided by the sum of the lengths of all road centerlines of the network topology.

[0018] As an improvement of the above solution, the method of extracting the network topology structure of the road centerline in the preset area from the road image includes:

[0019] Using several different models to extract road center lines in a preset area from the road image to obtain several initial network structures;

[0020] After all the initial network structures are fused, the Hessian matrix and Steger algorithm are used to extract the road centerline to obtain the network topology.

[0021] As an improvement of the above scheme, after all the initial network structures are fused, and before the Hessian matrix and Steger algorithm are used to extract the fused road centerline, it also includes: performing Gaussian smoothing on the fused initial network structure.

[0022] To achieve the above object, an embodiment of the present invention further provides a road network generation device, comprising:

[0023] A topology acquisition module, used to extract the network topology of the road centerline in a preset area from the road image;

[0024] A road network construction module, for gradually superimposing the road centerline into the road network in descending order of line segment length based on the network topology, and calculating the structural difference of the road network before and after each superposition;

[0025] The road network generation module is used to stop the superposition operation of the road centerline when the structural difference is less than the iteration termination threshold, and use the road network obtained by the last superposition operation as the road centerline network.

[0026] As an improvement of the above solution, a road network optimization module is further included, which is used to perform an optimization operation on the road centerline network, and the optimization operation includes at least one of the following methods:

[0027] Deleting open line segments of the road centerline network;

[0028] Deleting a small closed loop of the road centerline network; wherein the road centerline network includes a large closed loop and a small closed loop, and the road length of the small closed loop is less than the road length of the adjacent large closed loop;

[0029] Eliminate burrs in the road centerline network.

[0030] To achieve the above objectives, an embodiment of the present invention further provides a road network generation device, comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the road network generation method as described in any of the above embodiments is implemented.

[0031] To achieve the above objectives, an embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the road network generation method as described in any of the above embodiments.

[0032] Compared with the prior art, the road network generation method, device, equipment and storage medium disclosed in the embodiments of the present invention first extract the network topology structure of the road centerline segments in a preset area from the road image; then, based on the network topology structure, according to the line segment length of the road centerline, starting from the longest road centerline, the road centerlines are gradually superimposed to form a road network, and the structural difference of the road network before and after each superposition is calculated; when the structural difference is less than the iteration termination threshold, the superposition operation of the road centerline is stopped, and the road network obtained by the last superposition operation is used as the road centerline network. It can be seen that the embodiments of the present invention achieve a good balance between road integrity and accuracy by gradually constructing a road centerline network according to the line segment length and network differences of the road centerline, and only the road centerlines extracted from the image need to be superimposed, and there is no need to obtain the road extraction result by fusing multi-source data, which reduces the computing power requirements and is suitable for large-area engineering applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the technical solution of the present invention, the drawings required for use in the implementation mode will be briefly introduced below. Obviously, the drawings described below are only some implementation modes of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0034] Figure 1 It is a flowchart of a road network generation method provided by an embodiment of the present invention;

[0035] Figure 2 is a schematic diagram of road network topology optimization provided by an embodiment of the present invention;

[0036] Figure 3 This is a schematic diagram of a network topology structure of a road centerline provided by an embodiment of the present invention;

[0037] Figure 4 It is a schematic diagram of a process of a road network generating device provided by an embodiment of the present invention;

[0038] Figure 5 It is a flow chart of a road network generation device provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0039] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0040] See also Figure 1 , is a flow chart of a road network generation method provided by an embodiment of the present invention. Specifically, the road network generation method includes steps S1 to S3:

[0041] S1, extracting the network topology structure of the road centerline in a preset area from the road image;

[0042] S2. Based on the network topology, the road centerline is gradually superimposed onto the road network in descending order of line segment length, and the structural difference of the road network before and after each superposition is calculated;

[0043] S3. When the structural difference is less than the iteration termination threshold, the superposition operation of the road centerline is stopped, and the road network obtained by the last superposition operation is used as the road centerline network.

[0044] It is worth noting that the network topology refers to the structure formed by the road centerlines extracted directly from the road image, which belongs to the original topology. The road network refers to the topology formed by gradually superimposing the road centerlines in the network topology. The road centerline network is the final topology obtained after the road centerline superposition operation is completed.

[0045] Specifically, (1) the network topology of the road centerlines in the preset area is extracted from the remote sensing image, and (2) the lengths of the road centerlines are arranged in descending order from long to short. The set of arranged road centerlines is

[0046] ∪L i ,i=1,2,…,n(L i represents the length of the centerline of the ith road, n is the number of road centerlines, L1>L2>…>L n ), (3) then gradually superimpose them into the road network according to the length of the road centerline from large to small; when a new road segment is added to the road network, calculate the difference (structural difference) between the two road networks caused by adding the road centerline and not adding the road centerline; (4) when the difference d is greater than or equal to the iteration termination threshold y, repeat steps (2) and (3); when the difference d is less than the iteration termination threshold y, the iteration is terminated and the road centerline network construction is completed.

[0047] It is worth noting that in the process of constructing the road centerline network, the initial iteration difference is large. When the number of iterations exceeds a certain value (such as 200), the subsequent addition of road segments has little effect on the overall road network difference. When the iteration difference is small, most of the subsequent additional road segments are mistaken, which not only has no obvious effect on improving the overall accuracy, but also will lose the road network accuracy. At this time, the iteration process should be exited.

[0048] Compared with the prior art, the embodiment of the present invention proposes a road centerline network construction method based on an iterative algorithm, which gradually constructs the road centerline network according to the length of the road centerline segments and the network differences, achieves a good balance between road integrity and accuracy, and reduces computing power requirements, which is suitable for large-area engineering applications.

[0049] In one embodiment, the method further includes: performing an optimization operation on the road centerline network, wherein the optimization operation includes at least one of the following methods:

[0050] Deleting open line segments of the road centerline network;

[0051] Deleting a small closed loop of the road centerline network; wherein the road centerline network includes a large closed loop and a small closed loop, and the road length of the small closed loop is less than the road length of the adjacent large closed loop;

[0052] Eliminate burrs in the road centerline network.

[0053] Specifically, starting from the road network topology, the road centerline network is further improved, and some unreasonable road centerlines are deleted to improve the accuracy of the road centerline network. The initial road centerline network may have the following situations:

[0054] 1) The road centerline network has open segments that are not connected to the backbone network, such as Figure 2 (a)-①;

[0055] 2) There are some small closed loops, which destroy the topological strength of the road centerline network, such as Figure 2 (a)-②; The definition of a small closed loop can be set according to actual conditions. For example, when the ratio of the road length of a closed loop to that of other adjacent closed loops is less than a preset ratio threshold, the closed loop is considered to be a small closed loop; or, when the road length of a closed loop is less than a preset closure threshold, the closed loop is considered to be a small closed loop; or, when the ratio of the road length of a closed loop to other adjacent closed loops is less than a preset ratio threshold and the road length of the closed loop is less than the preset closure threshold, the closed loop is considered to be a small closed loop.

[0056] 3) For the urban road network, the road intersections are connected end to end to form a closed loop; the open curves attached to the closed loop account for a very low proportion in the entire urban road network, so the open curves attached to the closed loop can be considered as burrs, such as Figure 2 (a)-③.

[0057] In order to further improve and optimize the road centerline network topology, the above three situations are handled as follows:

[0058] (1): Detect closed loops based on the road centerline network and delete non-closed line segments (open line segments);

[0059] (2): The closed loops of the road centerline network are detected in layers. After the closed loop structure of the road centerline network is determined, the small closed loops are deleted. For example, closed loops of various levels of roads such as main roads, secondary roads, and roads within residential areas are detected in turn, and then the small closed loops are deleted.

[0060] (3): After completing the above two steps, the burrs are gradually eliminated until the road centerline network no longer changes (i.e., all burrs are removed) or the rate of change is less than a given threshold, and the iteration is terminated. Exemplarily, burrs are eliminated on roads at different levels in turn until all burrs are eliminated, or until the rate of change of the lengths of all road centerlines of the road centerline network before and after a certain burr elimination is less than a given threshold. The given threshold is set according to actual conditions.

[0061] Figure 2 (b) shows the results of road network topology optimization. It can be seen that by using topological features such as open line segments and small closed loops to remove incorrectly extracted road centerlines, while ensuring the road backbone network structure, the impact of noise (such as incorrectly extracted road segments) on the road network topology is reduced, further improving the road extraction accuracy and road network topology strength.

[0062] In one implementation, calculating the structural difference of the road network before and after each superposition includes:

[0063] Calculate the length of the road network before and after each overlay;

[0064] For each superposition operation, the difference between the length of the road network after superposition and the length of the road network before superposition is divided by the length of the road network before superposition to obtain the structural difference of the road network before and after each superposition.

[0065] Specifically, during the road network iteration process, when a new road center segment is added to the road network, the difference between the two road networks resulting from adding the road center line and not adding the road center line is calculated;

[0066]

[0067] Where d represents the difference between the two road networks, t represents the t-th iteration (t = 1, 2, .., n-1), represents the length of the road network at the tth iteration, Represents the length of the road network at the t+1th iteration.

[0068] In one implementation, the iteration termination threshold is determined by:

[0069] Before each superposition, the iteration termination threshold is obtained by subtracting the current cumulative probability from one; wherein the current cumulative probability is equal to the length of the current road network divided by the sum of the lengths of all road centerlines of the network topology. The length of the current road network refers to the sum of the lengths of the road centerlines currently superimposed on the road network.

[0070] Specifically, the iteration termination threshold y is defined as:

[0071]

[0072] Where n is the number of all road centerlines; t is the current iteration number.

[0073] In one embodiment, extracting the network topology structure of the road centerline segments in a preset area from the road image includes:

[0074] Using several different models to extract road center lines in a preset area from the road image to obtain several initial network structures;

[0075] After all the initial network structures are fused, the Hessian matrix and Steger algorithm are used to extract the road centerline to obtain the network topology.

[0076] In one embodiment, after all initial network structures are fused, and before the Hessian matrix and Steger algorithm are used to extract the fused road centerline, the method further includes: performing Gaussian smoothing on the fused initial network structure.

[0077] Specifically, in the prior art, most fusion algorithms require the deep learning network to output road surface data, and then fuse multiple road surface data into a single road surface, and finally identify the road centerline. If the deep learning integration model directly outputs multiple road centerlines, its fusion strategy is not suitable for fusing multiple initial network structures. Therefore, the embodiment of the present invention adopts the following fusion extraction method for multiple road centerlines:

[0078] (I) A road centerline trimming algorithm based on Gaussian filtering is used to obtain the original road.

[0079] A road centerline recognition algorithm is proposed from the perspective of Gaussian filtering. Assuming I is a road surface image, I(x,y) is a pixel in the image (x and y represent the row and column number of the pixel, respectively), the Hessian matrix of the pixel can be expressed as:

[0080]

[0081] In the formula, r xx is the second-order partial derivative of I(x,y) in the row direction, r xy 、r yx 、r yy The meaning of r xx Similar, no further description is given here. Due to the complexity of the road scene, before applying the Hessian matrix to calculate the road centerline, it is necessary to perform Gaussian smoothing on the original image (the image obtained by fusing all the initial network structures), and the kernel bandwidth in Gaussian smoothing is set to half the road width. According to the Hessian matrix, the Steger algorithm is used to extract the initial road centerline to obtain the network topology. For details, see Figure 3 As shown, Figure 3 (a) and (b) are the centerline extraction results (initial network structure) based on the road width map and road probability map, respectively; (c) is the superposition result of the two centerlines; (d) is the result after Gaussian smoothing of (c); (e) and (f) are the local magnified images of the centerline before and after Gaussian smoothing. Among them, the road width map and road probability map are road images obtained using different models.

[0082] This implementation overcomes the problem of offset between multiple road centerlines by superimposing multiple road centerline results (initial network structure) and using Gaussian filtering to correct the multiple road centerline results.

[0083] Compared with the prior art, the embodiment of the present invention first uses different deep learning models to extract multiple initial network structures; then superimposes multiple road centerline results (i.e., fuses multiple initial network structures); due to the presence of offsets and other situations in multiple road centerlines, a Gaussian filtering method is used to trim and improve multiple road centerline networks; the Hessian matrix and Steger algorithm are used to identify the road centerline from the Gaussian filtering results, and extract the initial road network information (the network topology structure of the road centerline), which is suitable for the scenario of fusion of multiple road centerlines, overcoming the limitation that the traditional fusion method is only applicable to road surfaces; according to the length of the road centerline segment , starting from the longest road centerline segment, the road centerline segments are gradually superimposed to form a network. When the network structure difference is less than a given threshold, the iteration is terminated to form a road centerline network. Based on the idea of ​​iteration, a better balance can be achieved between the integrity and accuracy of the road centerline network, which effectively reduces the impact of erroneous road extraction (such as burrs) on the accuracy; finally, the topological structure characteristics of the road centerline network are comprehensively considered, the topological structure of the road centerline network is optimized, and the accuracy of the road centerline network is further improved. It only needs to superimpose the road centerlines extracted from the image, and there is no need to obtain the road extraction results by fusing multi-source data, which reduces the computing power requirements.

[0084] See also Figure 4 The embodiment of the present invention further provides a road network generation device, comprising:

[0085] A topology structure acquisition module 11 is used to extract the network topology structure of the road centerline segment in a preset area from the road image;

[0086] A road network construction module 12 is used to gradually superimpose the road centerline segments into the road network in descending order of line segment length based on the network topology, and calculate the structural difference of the road network before and after each superposition;

[0087] The road network generation module 13 is used to stop the superposition operation of the road centerline segments when the structural difference is less than the iteration termination threshold, and use the road network obtained by the last superposition operation as the road centerline network.

[0088] In one embodiment, a road network optimization module is further included, which is used to perform an optimization operation on the road centerline network, and the optimization operation includes at least one of the following methods:

[0089] Deleting open line segments of the road centerline network;

[0090] Deleting a small closed loop of the road centerline network; wherein the road centerline network includes a large closed loop and a small closed loop, and the road length of the small closed loop is less than the road length of the adjacent large closed loop;

[0091] Eliminate burrs in the road centerline network.

[0092] In one implementation, calculating the structural difference of the road network before and after each superposition includes:

[0093] Calculate the length of the road network before and after each overlay;

[0094] For each superposition operation, the difference between the length of the road network after superposition and the length of the road network before superposition is divided by the length of the road network before superposition to obtain the structural difference of the road network before and after each superposition.

[0095] In one implementation, the iteration termination threshold is determined by:

[0096] Before each superposition, the iteration termination threshold is obtained by subtracting the current cumulative probability from one; wherein the current cumulative probability is equal to the length of the current road network divided by the sum of the lengths of all road centerlines of the network topology.

[0097] In one implementation, the topology structure acquisition module is specifically used to:

[0098] Using several different models to extract road center lines in a preset area from the road image to obtain several initial network structures;

[0099] After all the initial network structures are fused, the Hessian matrix and Steger algorithm are used to extract the road centerline to obtain the network topology.

[0100] In one embodiment, the topology structure acquisition module is further used to: after all initial network structures are fused, and before using the Hessian matrix and Steger algorithm to extract the fused road centerline, perform Gaussian smoothing on the fused initial network structure.

[0101] It is worth noting that the specific working process of the road network generation device can refer to the working process of the road network generation method described in the above embodiment, which will not be repeated here.

[0102] Compared with the prior art, the road network generation device disclosed in the embodiment of the present invention first extracts the network topology structure of the road centerline segments in a preset area from the road image; then, based on the network topology structure, according to the line segment length of the road centerline, starting from the longest road centerline, the road centerlines are gradually superimposed to form a road network, and the structural difference of the road network before and after each superposition is calculated; when the structural difference is less than the iteration termination threshold, the superposition operation of the road centerline is stopped, and the road network obtained by the last superposition operation is used as the road centerline network. It can be seen that the embodiment of the present invention achieves a good balance between road integrity and accuracy by gradually constructing a road centerline network according to the line segment length and network differences of the road centerline, and reduces the computing power requirements, which is suitable for large-area engineering applications.

[0103] See also Figure 5 The embodiment of the present invention further provides a road network generation device, including a processor 21, a memory 22, and a computer program stored in the memory 22 and configured to be executed by the processor 21. When the processor 21 executes the computer program, the steps in the above-mentioned road network generation method embodiment are implemented, for example Figure 1 or, the processor 21 implements the functions of each module in the above-mentioned device embodiments when executing the computer program.

[0104] Exemplarily, the computer program may be divided into one or more modules, which are stored in the memory 22 and executed by the processor 21 to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of completing specific functions, which are used to describe the execution process of the computer program in the road network generation device. For example, the computer program may be divided into multiple modules, and the specific functions of each module are as follows:

[0105] A topology structure acquisition module, used to extract the network topology structure of the road centerline segment in a preset area from the road image;

[0106] A road network construction module, for gradually superimposing the road centerline segments into the road network in descending order of segment length based on the network topology, and calculating the structural difference of the road network before and after each superposition;

[0107] The road network generation module is used to stop the superposition operation of the road centerline segments when the structural difference is less than the iteration termination threshold, and use the road network obtained by the last superposition operation as the road centerline network.

[0108] The specific working process of each module can refer to the working process of the road network generation device described in the above embodiment, which will not be repeated here.

[0109] The road network generation device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The road network generation device may include, but is not limited to, a processor 21 and a memory 22. Those skilled in the art may understand that the road network generation device may also include input and output devices, network access devices, buses, etc.

[0110] The processor 21 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc. The processor 21 is the control center of the road network generation device, and uses various interfaces and lines to connect various parts of the entire road network generation device.

[0111] The memory 22 can be used to store the computer program and / or module. The processor 21 realizes various functions of the road network generation device by running or executing the computer program and / or module stored in the memory 22 and calling the data stored in the memory 22. The memory 22 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as an image playback function, etc.), etc.; the data storage area can store data created according to the use of the mobile phone, etc. In addition, the memory 22 can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (SmartMedia Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0112] Wherein, if the module integrated in the road network generation device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor 21, the steps of the above-mentioned method embodiments can be implemented. Wherein, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium.

[0113] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A road network generation method, characterized in that: include: Extracting the network topology of the road centerline in a preset area from the road image; Based on the network topology, the road centerline is gradually superimposed on the road network in the order of line segment length from large to small, and the structural difference of the road network before and after each superposition is calculated; When the structural difference is less than the iteration termination threshold, the superposition operation of the road centerline is stopped, and the road network obtained by the last superposition operation is used as the road centerline network.

2. The road network generation method according to claim 1, characterized in that: Also includes: An optimization operation is performed on the road centerline network, wherein the optimization operation includes at least one of the following methods: Deleting open line segments of the road centerline network; Deleting a small closed loop of the road centerline network; wherein the road centerline network includes a large closed loop and a small closed loop, and the road length of the small closed loop is less than the road length of the adjacent large closed loop; Eliminate burrs in the road centerline network.

3. The road network generation method according to claim 1, characterized in that: The calculation of the structural difference of the road network before and after each superposition includes: Calculate the length of the road network before and after each overlay; For each superposition operation, the difference between the length of the road network after superposition and the length of the road network before superposition is divided by the length of the road network before superposition to obtain the structural difference of the road network before and after each superposition.

4. The road network generation method according to claim 1, characterized in that: The iteration termination threshold is determined by: Before each superposition, the iteration termination threshold is obtained by subtracting the current cumulative probability from one; wherein the current cumulative probability is equal to the length of the current road network divided by the sum of the lengths of all road centerlines of the network topology.

5. The road network generation method according to any one of claims 1 to 4, characterized in that: The method of extracting the network topology structure of the road centerline in the preset area from the road image includes: Using several different models to extract road center lines in a preset area from the road image to obtain several initial network structures; After all the initial network structures are merged, the Hessian matrix and Steger algorithm are used to extract the road centerline to obtain the network topology structure.

6. The road network generation method according to claim 5, characterized in that: After all the initial network structures are fused, and before the Hessian matrix and Steger algorithm are used to extract the fused road centerline, the method further includes: performing Gaussian smoothing on the fused initial network structure.

7. A road network generation device, characterized in that: include: A topology acquisition module, used to extract the network topology of the road centerline in a preset area from the road image; A road network construction module, for gradually superimposing the road centerline into the road network in descending order of line segment length based on the network topology, and calculating the structural difference of the road network before and after each superposition; The road network generation module is used to stop the superposition operation of the road centerline when the structural difference is less than the iteration termination threshold, and use the road network obtained by the last superposition operation as the road centerline network.

8. The road network generation device according to claim 7, characterized in that: It also includes a road network optimization module, which is used to perform an optimization operation on the road centerline network, wherein the optimization operation includes at least one of the following methods: Deleting open line segments of the road centerline network; Deleting a small closed loop of the road centerline network; wherein the road centerline network includes a large closed loop and a small closed loop, and the road length of the small closed loop is less than the road length of the adjacent large closed loop; Eliminate burrs in the road centerline network.

9. A road network generation device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor implements the road network generation method according to any one of claims 1 to 6 when executing the computer program.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the road network generation and providing method according to any one of claims 1 to 6.

Images