Lane line extraction method and system based on unmanned aerial vehicle image, medium and equipment

Through the 3D point cloud processing of drone image data and transformer semantic segmentation neural network, the problem of low lane line detection efficiency is solved, and fast and flexible lane line extraction is achieved, which is suitable for large-scale high-precision map updates.

CN119027833APending Publication Date: 2024-11-26WUHAN ZHONGHAITING DATA TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410918080.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-10
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

In the prior art, lane line detection is mostly based on vehicle-mounted camera data, which has problems with limited imaging range and low efficiency. Satellite remote sensing images cannot quickly and flexibly obtain lane-level information, making it difficult to apply to large-scale high-precision map production and updates.

Method used

Using the lane line extraction method based on drone images, the 3D point cloud data collected by drone is obtained, and the lane line information is extracted by using transformer semantic segmentation neural network to achieve rapid Flexible lane line detection.

Benefits of technology

It realizes fast and flexible access to lane-level information, which is suitable for large-scale high-precision map production and updates, improves lane line extraction efficiency and accuracy, and avoids the problems of inefficiency and incomplete information in traditional methods.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119027833A_ABST
    Figure CN119027833A_ABST
Patent Text Reader

Abstract

The invention discloses a lane line extraction method and system based on an unmanned aerial vehicle image, a medium and equipment. The method comprises the following steps: acquiring 3D point cloud data acquired by an unmanned aerial vehicle on a to-be-processed road; performing fusion slicing on the 3D point cloud data, and segmenting the 3D point cloud data into a plurality of point cloud blocks with the same length; projecting each point cloud block into a two-dimensional world coordinate system to obtain a corresponding two-dimensional projection drawing; extracting lane line information in each two-dimensional projection drawing by using a trained transform semantic segmentation neural network, and back-projecting the extracted lane line information to a three-dimensional world coordinate system to obtain corresponding three-dimensional lane line information; constructing a corresponding lane line area according to the three-dimensional lane line information in each point cloud block, and sequentially performing vectorization connection on the lane line areas corresponding to all the point cloud blocks to obtain road surface lane line information of the to-be-processed road; according to the invention, lane level information can be rapidly and flexibly obtained, and lane line detection and extraction are carried out.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of lane line detection and extraction, and in particular to a lane line extraction method, system, medium and equipment based on drone images. Background Art

[0002] 3D high-precision electronic maps play an important auxiliary role in unmanned driving and intelligent driving. Lane lines are an important part of the lane-level road network of high-precision maps. In the past, lane line detection in high-precision maps was mostly based on on-board camera data, which had efficiency problems caused by limited imaging range, perspective transformation, and multiple stitching. The method of detecting lane lines based on on-board images is not suitable for large-scale high-precision map production and updating, and the satellite remote sensing image method is limited by spatial resolution and update cycle, and cannot quickly and flexibly obtain lane-level information. Summary of the invention

[0003] The present invention provides a lane line extraction method, system, medium and equipment based on drone images, which can quickly and flexibly obtain lane-level information and perform lane line detection and extraction.

[0004] In a first aspect, a lane line extraction method based on drone images is provided, comprising the following steps:

[0005] Obtain the 3D point cloud data collected by the drone on the road to be processed;

[0006] The 3D point cloud data is merged and sliced ​​into a plurality of point cloud blocks of the same length;

[0007] Projecting each of the point cloud blocks into a two-dimensional world coordinate system to obtain a corresponding two-dimensional projection image;

[0008] Using the trained transformer semantic segmentation neural network to extract lane line information in each of the two-dimensional projection images, and back-projecting the extracted lane line information into a three-dimensional world coordinate system to obtain corresponding three-dimensional lane line information;

[0009] A corresponding lane line area is constructed according to the three-dimensional lane line information in each point cloud block, and the lane line areas corresponding to all point cloud blocks are vectorized and connected in sequence to obtain the road surface lane line information of the road to be processed.

[0010] According to the first aspect, in a first possible implementation manner of the first aspect, the step of “merging and slicing the 3D point cloud data to divide it into a plurality of point cloud blocks of the same length” specifically includes the following steps:

[0011] Performing point cloud block fusion on the 3D point cloud data according to the OSM database and the point cloud range of the 3D point cloud data to generate a road surface related point cloud area;

[0012] Performing a sampling vehicle style virtualization on the road trajectory point information of the 3D point cloud data, and eliminating the point cloud data of non-road information in the relevant point cloud area of ​​the road surface;

[0013] The relevant point cloud area of ​​the road surface after segmentation is eliminated to obtain multiple point cloud blocks of the same length.

[0014] According to the first aspect, in a second possible implementation manner of the first aspect, the step of “constructing a corresponding lane line area according to the three-dimensional lane line information in each of the point cloud blocks, and sequentially vectorizing and connecting the lane line areas corresponding to all the point cloud blocks to obtain the road surface lane line information of the road to be processed” specifically includes the following steps:

[0015] The lane line shape points contained in the three-dimensional lane line information are connected in sequence to construct the lane line area corresponding to each point cloud block, and the lane line areas corresponding to all point cloud blocks are vectorized and connected in sequence. The lane lines with broken shapes during the lane line area connection process are subjected to point cloud fusion and point cloud error correction processing to obtain a complete lane line area.

[0016] According to the first aspect, in a third possible implementation manner of the first aspect, before the step of “merging and slicing the 3D point cloud data to divide it into a plurality of point cloud blocks of the same length”, the following steps are specifically included:

[0017] The 3D point cloud data is subjected to thinning processing.

[0018] According to the first aspect, in a fourth possible implementation manner of the first aspect, before the step of “using the trained transformer semantic segmentation neural network to extract lane line information in each of the two-dimensional projection images, and back-projecting the extracted lane line information into a three-dimensional world coordinate system to obtain corresponding three-dimensional lane line information”, the following steps are specifically included:

[0019] The local histogram equalization algorithm is used to enhance the two-dimensional projection image.

[0020] According to the first aspect, in a fifth possible implementation manner of the first aspect, before the step of “constructing a corresponding lane line area according to the three-dimensional lane line information in each of the point cloud blocks, sequentially vectorizing and connecting the lane line areas corresponding to all the point cloud blocks, and obtaining the road surface lane line information of the road to be processed”, the following steps are specifically included:

[0021] The repeatedly extracted lane line information and the non-lane line information that does not meet the lane line standard are eliminated from the three-dimensional lane line information.

[0022] In the second aspect, a lane line extraction system based on drone images is provided, comprising:

[0023] The data acquisition module is used to obtain the 3D point cloud data collected by the drone on the road to be processed;

[0024] A segmentation module, which is in communication with the data acquisition module and is used to fuse and slice the 3D point cloud data into a plurality of point cloud blocks of the same length;

[0025] A projection module, which is in communication with the segmentation module and is used to project each of the point cloud blocks into a two-dimensional world coordinate system to obtain a corresponding two-dimensional projection image;

[0026] a back-projection module, which is in communication with the projection module and is used to extract lane line information in each of the two-dimensional projection images using a trained transformer semantic segmentation neural network, and back-project the extracted lane line information into a three-dimensional world coordinate system to obtain corresponding three-dimensional lane line information; and

[0027] The sequential connection module is communicated with the back-projection module and is used to construct a corresponding lane line area according to the three-dimensional lane line information in each point cloud block, and sequentially vectorize and connect the lane line areas corresponding to all point cloud blocks to obtain the road surface lane line information of the road to be processed.

[0028] In a third aspect, a computer-readable storage medium is provided, on which a computer program is stored, characterized in that when the computer program is executed by a processor, the lane line extraction method based on drone images as described above is implemented.

[0029] In a fourth aspect, an electronic device is provided, comprising a storage medium, a processor, and a computer program stored in the storage medium and executable on the processor, wherein the processor implements the lane line extraction method based on drone images as described above when running the computer program.

[0030] Compared with the prior art, the advantages of the present invention are as follows: lane line detection and extraction based on drone image data can quickly and flexibly obtain lane-level information, and perform lane line detection and extraction to be suitable for large-scale high-precision map production and updating. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a flow chart of an embodiment of a lane line extraction method based on drone images of the present invention;

[0032] Figure 2 It is the projection map after point cloud segmentation, lane line extraction map and lane line connection projection map of the present invention;

[0033] Figure 3It is a flow chart of another embodiment of a lane line extraction method based on drone images of the present invention;

[0034] Figure 4 It is a structural schematic diagram of a lane line extraction system based on drone images of the present invention. DETAILED DESCRIPTION

[0035] Reference will now be made in detail to specific embodiments of the present invention, examples of which are illustrated in the accompanying drawings. Although the present invention will be described in conjunction with specific embodiments, it will be understood that it is not intended to limit the present invention to the embodiments described. On the contrary, it is intended to cover changes, modifications and equivalents included in the spirit and scope of the present invention as defined by the appended claims. It should be noted that the method steps described herein can all be implemented by any functional block or functional arrangement, and any functional block or functional arrangement can be implemented as a physical entity or a logical entity, or a combination of the two.

[0036] In order to enable those skilled in the art to better understand the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0037] Note: The example to be introduced below is only a specific example, and is not intended to limit the embodiments of the present invention to the following specific steps, values, conditions, data, sequences, etc. Those skilled in the art can apply the concept of the present invention to construct more embodiments not mentioned in this specification by reading this specification.

[0038] See also Figure 1 As shown, an embodiment of the present invention provides a lane line extraction method based on drone images, comprising the following steps:

[0039] S100, obtaining 3D point cloud data collected by the drone on the road to be processed;

[0040] The 3D point cloud data includes the collected point cloud information and trajectory point information.

[0041] S200, fusing and slicing the 3D point cloud data to divide it into a plurality of point cloud blocks of the same length;

[0042] S300, projecting each of the point cloud blocks into a two-dimensional world coordinate system to obtain a corresponding two-dimensional projection image; see Figure 2 As shown in (a);

[0043] According to the trajectory point information, the point cloud block data is perspective projected one by one to generate a grayscale image of a fixed size, and the elevation information is retained.

[0044] S400, using the trained transformer semantic segmentation neural network to extract lane line information in each of the two-dimensional projection images, see Figure 2 As shown in (b), the extracted lane line information is back-projected into a three-dimensional world coordinate system to obtain corresponding three-dimensional lane line information;

[0045] According to the extracted lane line information, the reverse perspective projection change is performed, and combined with the previously retained elevation information, the extracted lane line point two-dimensional coordinates are converted into three-dimensional world coordinates.

[0046] S500, constructing a corresponding lane line area according to the three-dimensional lane line information in each point cloud block, and sequentially vectorizing and connecting the lane line areas corresponding to all point cloud blocks to obtain the road surface lane line information of the road to be processed, see Figure 2 (c) in the figure shows the projection image after the lane lines are connected.

[0047] Specifically, in this embodiment, since lane line detection in existing high-precision maps is mostly based on vehicle-mounted camera data, there are efficiency problems caused by limited imaging range, perspective transformation and multiple stitching. Therefore, the method of detecting lane lines based on vehicle-mounted images is not suitable for large-scale high-precision map production and updating; and the satellite remote sensing image method is limited by spatial resolution and update cycle, and cannot quickly and flexibly obtain lane-level information. Therefore, in response to the above problems, the present invention is based on drone image data for lane line detection and extraction, which can quickly and flexibly obtain lane-level information and perform lane line detection and extraction to be suitable for large-scale high-precision map production and updating.

[0048] Preferably, in another embodiment of the present application, the step of “S200, fusing and slicing the 3D point cloud data to divide it into a plurality of point cloud blocks of the same length” specifically includes the following steps:

[0049] Performing point cloud block fusion on the 3D point cloud data according to the OSM database and the point cloud range of the 3D point cloud data to generate a road surface related point cloud area;

[0050] Performing a sampling vehicle style virtualization on the road trajectory point information of the 3D point cloud data, and eliminating the point cloud data of non-road information in the relevant point cloud area of ​​the road surface;

[0051] The relevant point cloud area of ​​the road surface after segmentation is eliminated to obtain multiple point cloud blocks of the same length.

[0052] Specifically, in this embodiment, the drone point cloud tiles are first fused according to the point cloud range and the OSM (OpenStreetMap) database to generate a point cloud area related to the road surface; then the road trajectory of the point cloud data is virtualized in the style of the collection vehicle, and the point cloud data of the surrounding non-road information in the point cloud is eliminated; finally, the point cloud is re-divided into point cloud blocks of fixed length.

[0053] Sampling vehicle style virtualization refers to expressing the road trajectory point information of the collected 3D point cloud data in accordance with the collection vehicle data format, including the trajectory information traceinfo.db and the degree band number centerMedia and other information.

[0054] Preferably, in another embodiment of the present application, the step of "S500, constructing a corresponding lane line area according to the three-dimensional lane line information in each point cloud block, and sequentially vectorizing and connecting the lane line areas corresponding to all the point cloud blocks to obtain the road surface lane line information of the road to be processed" specifically includes the following steps:

[0055] The lane line shape points contained in the three-dimensional lane line information are connected in sequence to construct the lane line area corresponding to each point cloud block, and the lane line areas corresponding to all point cloud blocks are vectorized and connected in sequence. The lane lines with broken shapes during the lane line area connection process are subjected to point cloud fusion and point cloud error correction processing to obtain a complete lane line area.

[0056] Specifically, in this embodiment, the extracted lane line information vectors are independent in each point cloud block. In order to obtain complete lane line information on the traffic road, they need to be connected, which involves point cloud fusion and point cloud correction operations.

[0057] Preferably, in another embodiment of the present application, before the step of “S200, fusing and slicing the 3D point cloud data to divide it into a plurality of point cloud blocks of the same length”, the following steps are specifically included:

[0058] The 3D point cloud data is subjected to thinning processing.

[0059] Specifically, in this embodiment, the collected point cloud is thinned out, which can be understood as replacing the voxel area with the average value of a single voxel, thereby thinning the point cloud.

[0060] Preferably, in another embodiment of the present application, before the step of “S400, using the trained transformer semantic segmentation neural network to extract lane line information in each of the two-dimensional projection images, and back-projecting the extracted lane line information into a three-dimensional world coordinate system to obtain corresponding three-dimensional lane line information”, the following steps are specifically included:

[0061] The local histogram equalization algorithm is used to enhance the two-dimensional projection image.

[0062] Specifically, in this embodiment, if the enhancement processing is not performed, the generated projection image is not very clear or is missing, which will cause the subsequently extracted lane line information to be incomplete or even unable to be extracted. However, using the local histogram equalization image makes the extracted lane lines clearer and more complete, thereby improving the accuracy of lane line information extraction.

[0063] Preferably, in another embodiment of the present application, before the step of "S500, constructing a corresponding lane line area according to the three-dimensional lane line information in each of the point cloud blocks, and sequentially vectorizing and connecting the lane line areas corresponding to all the point cloud blocks to obtain the road surface lane line information of the road to be processed", the following steps are specifically included:

[0064] The repeatedly extracted lane line information and the non-lane line information that does not meet the lane line standard are eliminated from the three-dimensional lane line information.

[0065] Specifically, in this embodiment, based on the geometric relationship of the lane lines and the lane line production standards, the non-lane line information that is repeatedly extracted and does not meet the lane line standards is eliminated.

[0066] The embodiment of the present invention provides a method for extracting lane lines based on drone images. For specific steps, see Figure 3 shown.

[0067] The beneficial effects of the present invention are as follows:

[0068] 1. The present invention uses the perspective projection transformation of the three-dimensional point cloud to a two-dimensional projection map to extract lane line information, which can not only meet the 5 cm error requirement for high-precision map production, but also improve the efficiency of lane line extraction.

[0069] 2. The collected point clouds are thinned to reduce the number of point clouds to be processed and improve the efficiency of lane line information extraction.

[0070] 3. The present invention performs enhancement processing on the projection image. If the enhancement processing is not performed, the generated projection image is not very clear or is missing, which will cause the subsequently extracted lane line information to be incomplete or even unable to be extracted. The use of local histogram equalization images makes the extracted lane lines clearer and more complete, thereby improving the accuracy of lane line information extraction.

[0071] 4. The present invention uses a transformer semantic segmentation neural network to extract lane line information, which can obtain lane line point coordinate information very accurately. Compared with traditional corner point detection methods, it is applicable to more and more complex scenarios and has higher accuracy.

[0072] 5. The lane line extraction model of the present invention is deployed on the server, and does not have such high requirements for the environment of the local machine. The previous models required high-end graphics cards and running memory to run on the local machine. The present invention directly calls the inference model on the server, and even a low-end computer can run normally and quickly, and at a faster speed.

[0073] 6. The present invention adds deduplication and filtering operations to the extracted lane line information, thereby avoiding confusion and redundancy in the extracted lane lines and improving the efficiency of lane line extraction.

[0074] 7. The present invention adds point cloud fusion and point cloud deviation correction operations in the lane line connection stage, making the extracted lane lines more complete and avoiding the problems of misalignment and breakage of the extracted lane lines in the previous technology.

[0075] See also Figure 4 As shown, an embodiment of the present invention provides a lane line extraction system based on drone images, comprising:

[0076] The data acquisition module is used to obtain the 3D point cloud data collected by the drone on the road to be processed;

[0077] A segmentation module, which is in communication with the data acquisition module and is used to fuse and slice the 3D point cloud data into a plurality of point cloud blocks of the same length;

[0078] A projection module, which is in communication with the segmentation module and is used to project each of the point cloud blocks into a two-dimensional world coordinate system to obtain a corresponding two-dimensional projection image;

[0079] a back-projection module, which is in communication with the projection module and is used to extract lane line information in each of the two-dimensional projection images using a trained transformer semantic segmentation neural network, and back-project the extracted lane line information into a three-dimensional world coordinate system to obtain corresponding three-dimensional lane line information; and

[0080] The sequential connection module is communicated with the back-projection module and is used to construct a corresponding lane line area according to the three-dimensional lane line information in each point cloud block, and sequentially vectorize and connect the lane line areas corresponding to all point cloud blocks to obtain the road surface lane line information of the road to be processed.

[0081] The lane line detection and extraction performed by the present invention based on drone image data can quickly and flexibly obtain lane-level information and perform lane line detection and extraction, so as to be suitable for large-scale high-precision map production and updating.

[0082] Specifically, this embodiment corresponds one-to-one to the above method embodiment, and the functions of each module have been described in detail in the corresponding method embodiment, so they will not be repeated here.

[0083] Based on the same inventive concept, an embodiment of the present application further provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, all or part of the method steps of the above method are implemented.

[0084] The present invention implements all or part of the process in the above method, and can also be completed by instructing related hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, the steps of each method embodiment can be implemented. Among them, 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 may include: any entity or device capable of carrying computer program code, recording medium, USB flash drive, mobile hard disk, 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. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electric carrier signals and telecommunication signals.

[0085] Based on the same inventive concept, an embodiment of the present application also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program running on the processor, and when the processor executes the computer program, all or part of the method steps in the above method are implemented.

[0086] The processor may be a central processing unit (CPU), 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 is the control center of a computer device, and uses various interfaces and lines to connect various parts of the entire computer device.

[0087] The memory can be used to store computer programs and / or modules. The processor realizes various functions of the computer device by running or executing the computer programs and / or modules stored in the memory, and calling the data stored in the memory. The memory 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 a sound playback function, an image playback function, etc.); the data storage area can store data created according to the use of the mobile phone (such as audio data, video data, etc.). In addition, the memory 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 (SmartMed ia 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.

[0088] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, servers or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) containing computer-usable program codes.

[0089] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), servers, and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as a combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0090] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0091] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0092] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A lane line extraction method based on drone images, characterized in that: The following steps are involved: Obtain the 3D point cloud data collected by the drone on the road to be processed; The 3D point cloud data is merged and sliced ​​into a plurality of point cloud blocks of the same length; Projecting each of the point cloud blocks into a two-dimensional world coordinate system to obtain a corresponding two-dimensional projection image; Using the trained transformer semantic segmentation neural network to extract lane line information in each of the two-dimensional projection images, and back-projecting the extracted lane line information into a three-dimensional world coordinate system to obtain corresponding three-dimensional lane line information; A corresponding lane line area is constructed according to the three-dimensional lane line information in each point cloud block, and the lane line areas corresponding to all point cloud blocks are vectorized and connected in sequence to obtain the road surface lane line information of the road to be processed.

2. The lane line extraction method based on drone images as claimed in claim 1, characterized in that: The step of "merging and slicing the 3D point cloud data to divide it into a plurality of point cloud blocks of the same length" specifically includes the following steps: Performing point cloud block fusion on the 3D point cloud data according to the OSM database and the point cloud range of the 3D point cloud data to generate a road surface related point cloud area; Performing a sampling vehicle style virtualization on the road trajectory point information of the 3D point cloud data, and eliminating the point cloud data of non-road information in the relevant point cloud area of ​​the road surface; The relevant point cloud area of ​​the road surface after segmentation is eliminated to obtain multiple point cloud blocks of the same length.

3. The lane line extraction method based on drone images as claimed in claim 1, characterized in that: The step of "constructing a corresponding lane line area according to the three-dimensional lane line information in each point cloud block, and sequentially vectorizing and connecting the lane line areas corresponding to all point cloud blocks to obtain the road surface lane line information of the road to be processed" specifically includes the following steps: The lane line shape points contained in the three-dimensional lane line information are connected in sequence to construct the lane line area corresponding to each point cloud block, and the lane line areas corresponding to all point cloud blocks are vectorized and connected in sequence. The lane lines with broken shapes during the lane line area connection process are subjected to point cloud fusion and point cloud error correction processing to obtain a complete lane line area.

4. The lane line extraction method based on drone images as claimed in claim 1, characterized in that: Before the step of "merging and slicing the 3D point cloud data to divide it into a plurality of point cloud blocks of the same length", the following steps are specifically included: The 3D point cloud data is subjected to thinning processing.

5. The lane line extraction method based on drone images as claimed in claim 1, characterized in that: Before the step of "using the trained transformer semantic segmentation neural network to extract lane line information in each of the two-dimensional projection images, and back-projecting the extracted lane line information into a three-dimensional world coordinate system to obtain corresponding three-dimensional lane line information", the following steps are specifically included: The local histogram equalization algorithm is used to enhance the two-dimensional projection image.

6. The lane line extraction method based on drone images as claimed in claim 1, characterized in that: Before the step of "constructing a corresponding lane line area according to the three-dimensional lane line information in each point cloud block, and sequentially vectorizing and connecting the lane line areas corresponding to all point cloud blocks to obtain the road surface lane line information of the road to be processed", the following steps are specifically included: The repeatedly extracted lane line information and the non-lane line information that does not meet the lane line standard are eliminated from the three-dimensional lane line information.

7. A lane line extraction system based on drone images, characterized in that: include: The data acquisition module is used to obtain the 3D point cloud data collected by the drone on the road to be processed; A segmentation module, which is in communication with the data acquisition module and is used to fuse and slice the 3D point cloud data into a plurality of point cloud blocks of the same length; A projection module, which is in communication with the segmentation module and is used to project each of the point cloud blocks into a two-dimensional world coordinate system to obtain a corresponding two-dimensional projection image; A back-projection module, which is in communication with the projection module and is used to extract lane line information in each of the two-dimensional projection images using a trained transformer semantic segmentation neural network, and back-project the extracted lane line information into a three-dimensional world coordinate system to obtain corresponding three-dimensional lane line information; as well as, The sequential connection module is communicated with the back-projection module and is used to construct a corresponding lane line area according to the three-dimensional lane line information in each point cloud block, and sequentially vectorize and connect the lane line areas corresponding to all point cloud blocks to obtain the road surface lane line information of the road to be processed.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the lane line extraction method based on drone images as described in any one of claims 1 to 7 is implemented.

9. An electronic device comprising a storage medium, a processor, and a computer program stored in the storage medium and executable on the processor, characterized in that: When the processor runs the computer program, the lane line extraction method based on drone images as described in any one of claims 1 to 7 is implemented.