Lane line marking method and device and computer equipment
By acquiring and processing three-dimensional point cloud data, masking height information to generate two-dimensional projection coordinates, and lane line labeling, the problems of low efficiency and high cost of manual labeling in the prior art are solved, and efficient and low-cost lane line labeling are achieved.
Patent Information
- Application Number
- CN202411733663.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-29
- Publication Date
- 2025-05-02
AI Technical Summary
The existing lane line labeling methods rely on manual labeling, which is inefficient and costly, making it difficult to reduce labeling costs without losing accuracy.
By obtaining the initial three-dimensional point cloud data, blocking the height information of the target point, obtaining the two-dimensional projection coordinates, inputting the annotation tool for labeling, and determining the height information of the target point based on the initial three-dimensional point cloud data, obtaining the three-dimensional target point containing the annotation results.
It realizes the reduction of lane line labeling costs without losing accuracy, improves labeling efficiency, and makes lane line labeling simpler and easier.
Smart Images
Figure CN119919903A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of autonomous driving technology, and more specifically, to a lane marking method, device and computer equipment. Background Art
[0002] Lane annotation is a very important step in the construction of high-precision maps or the training of lane extraction models. The known lane annotation method can annotate lane lines based on 3D point clouds. However, the manual annotation process of this method is time-consuming and labor-intensive, and has low efficiency. Therefore, how to improve the efficiency of lane annotation has become a technical problem that needs to be solved urgently. Summary of the invention
[0003] The present application provides a lane marking method, device and computer equipment to reduce the lane marking cost without losing accuracy. The specific technical solution is as follows.
[0004] In a first aspect, a lane marking method provided by an embodiment of the present application includes:
[0005] Acquire initial three-dimensional point cloud data corresponding to the road to be annotated; the initial three-dimensional point cloud data at least includes three-dimensional coordinate information of each target point;
[0006] Shielding the height information of each target point in the initial three-dimensional point cloud data to obtain the two-dimensional projection coordinates of each target point;
[0007] Inputting the two-dimensional projection coordinates of each target point into a labeling tool, displaying a target image including each target point, and receiving a labeling result input for the target image; the labeling result at least includes a lane line category;
[0008] For each of the target points in the target image, the height information of each of the target points is determined according to the initial three-dimensional point cloud data to obtain each of the three-dimensional target points including the labeling results.
[0009] Optionally, after obtaining the initial three-dimensional point cloud data corresponding to the road to be annotated, a lane line marking method provided in an embodiment of the present application further includes:
[0010] According to the preset road segment length, the initial three-dimensional point cloud data is processed into road segments to obtain initial three-dimensional point cloud data corresponding to each road segment.
[0011] Optionally, the three-dimensional point cloud data further includes: the driving direction of the vehicle; before inputting the two-dimensional projection coordinates of each target point into the annotation tool, the method further includes:
[0012] Determining a lane line direction according to the vehicle driving direction in the initial three-dimensional point cloud data;
[0013] The step of inputting the two-dimensional projection coordinates of each target point into the annotation tool to display a target image including each target point comprises:
[0014] The two-dimensional projection coordinates of each target point are input into the annotation tool, and a target image including each target point is displayed according to the lane line direction.
[0015] Optionally, the steps of inputting the two-dimensional projection coordinates of each target point into a labeling tool, displaying a target image including each target point, and receiving a labeling result input for the target image include:
[0016] Input the 2D projection coordinates of each target point into the annotation tool, select each target point located in the middle of the lane line according to the direction of the lane line, and display the target image including each target point;
[0017] The marking tool is used to mark the lane line category for each target point in the target image so that the target points of the same category automatically form a lane line, and the marking result for the target image input fed back from the marking tool is received.
[0018] Optionally, after receiving the annotation result fed back from the annotation tool for the target image input, a lane line annotation method provided in an embodiment of the present application further includes:
[0019] The lane lines automatically formed by the target points of the same category are normalized to obtain the normalized lane lines;
[0020] The lane lines processed in a standardized manner are encrypted using an encryption function according to the length of the line segments to obtain encrypted lane lines.
[0021] Optionally, the step of shielding the height information of each target point in the initial three-dimensional point cloud data to obtain the two-dimensional projection coordinates of each target point includes:
[0022] Selecting target three-dimensional point cloud data of a preset scale from the initial three-dimensional point cloud data;
[0023] The height information of each target point in the target three-dimensional point cloud data is shielded to obtain the two-dimensional projection coordinates of each target point.
[0024] Optionally, for each target point in the target image, determining the height information of each target point according to the initial three-dimensional point cloud data to obtain each three-dimensional target point including the annotation result includes:
[0025] For each target point in the target image, determine the coordinate information of each candidate point within a preset distance range with the target point as the center;
[0026] For each candidate point, in the initial three-dimensional point cloud data, a three-dimensional point whose x-coordinate and y-coordinate are the same as those of the candidate point is searched, and the average value of the height information of each three-dimensional point is used as the height information of the target point to obtain each three-dimensional target point containing the annotation results.
[0027] Optionally, after determining the coordinate information of each candidate point within a preset distance range with the target point as the center, a lane line marking method provided in an embodiment of the present application further includes:
[0028] When no candidate point is found, a preset number of adjacent points before and after the target point are searched on the lane line, and the height information of the target point is determined by linear interpolation based on the height information of each adjacent point.
[0029] In a second aspect, an embodiment of the present application provides a lane marking device including:
[0030] The acquisition module is configured to acquire initial three-dimensional point cloud data corresponding to the road to be annotated; the initial three-dimensional point cloud data at least includes three-dimensional coordinate information of each target point; the three-dimensional point cloud data also includes: the driving direction of the vehicle;
[0031] A shielding module is configured to shield the height information of each target point in the initial three-dimensional point cloud data to obtain the two-dimensional projection coordinates of each target point;
[0032] A labeling module is configured to input the two-dimensional projection coordinates of each target point into a labeling tool, display a target image including each target point, and receive a labeling result input for the target image; the labeling result at least includes a lane line category;
[0033] The determination module is configured to determine the height information of each target point in the target image according to the initial three-dimensional point cloud data, and obtain each three-dimensional target point including the marking result.
[0034] Optionally, the embodiment of the present application provides a lane marking device further comprising:
[0035] The segmentation module is configured to perform road segmentation processing on the initial three-dimensional point cloud data according to a preset road segment length to obtain initial three-dimensional point cloud data corresponding to each road segment.
[0036] Optionally, the embodiment of the present application provides a lane marking device further comprising: a direction determination module configured to determine the lane direction according to the vehicle driving direction in the initial three-dimensional point cloud data;
[0037] Optionally, the annotation module is configured as:
[0038] The two-dimensional projection coordinates of each target point are input into the annotation tool, and a target image including each target point is displayed according to the lane line direction.
[0039] Optionally, the annotation module is configured as:
[0040] Input the 2D projection coordinates of each target point into the annotation tool, select each target point located in the middle of the lane line according to the direction of the lane line, and display the target image including each target point;
[0041] The marking tool is used to mark the lane line category for each target point in the target image so that the target points of the same category automatically form a lane line, and the marking result for the target image input fed back from the marking tool is received.
[0042] Optionally, the embodiment of the present application provides a lane marking device further comprising: a processing module configured to:
[0043] The lane lines automatically formed by the target points of the same category are normalized to obtain normalized lane lines; the normalized lane lines are encrypted by using an encryption function according to the length of the line segments to obtain encrypted lane lines.
[0044] Optionally, the shielding module is configured to:
[0045] Selecting target three-dimensional point cloud data of a preset scale from the initial three-dimensional point cloud data;
[0046] The height information of each target point in the target three-dimensional point cloud data is shielded to obtain the two-dimensional projection coordinates of each target point.
[0047] Optionally, determine the module, configured as:
[0048] For each target point in the target image, determine the coordinate information of each candidate point within a preset distance range with the target point as the center;
[0049] For each candidate point, in the initial three-dimensional point cloud data, a three-dimensional point whose x-coordinate and y-coordinate are the same as those of the candidate point is searched, and the average value of the height information of each three-dimensional point is used as the height information of the target point to obtain each three-dimensional target point containing the annotation results.
[0050] Optionally, the embodiment of the present application provides a lane marking device further comprising: a search module configured to:
[0051] When no candidate point is found, a preset number of adjacent points before and after the target point are searched on the lane line, and the height information of the target point is determined by linear interpolation based on the height information of each adjacent point.
[0052] In a third aspect, an embodiment of the present application provides a computer device comprising: a memory and a processor, wherein the memory and the processor are coupled;
[0053] The memory is used to store one or more computer instructions;
[0054] The processor is used to execute the one or more computer instructions to implement the lane marking method as described in any one of claims 1-8.
[0055] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, on which one or more computer instructions are stored, and the instructions are executed by a processor to implement the lane line marking method as described in the first aspect above.
[0056] In a fifth aspect, the present application provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, it implements the lane line marking method as described in the first aspect above.
[0057] The present application provides a lane line labeling method, device and computer equipment, which obtains the initial three-dimensional point cloud data corresponding to the road to be labeled; shields the height information of each target point in the initial three-dimensional point cloud data, and obtains the two-dimensional projection coordinates of each target point; inputs the two-dimensional projection coordinates of each target point into the labeling tool, displays the target image including each target point, and receives the labeling results input for the target image; for each target point in the target image, determines the height information of each target point according to the initial three-dimensional point cloud data, and obtains each three-dimensional target point including the labeling results. The present application can realize lane line labeling only on two-dimensional images according to the needs, reduces the labeling cost without losing accuracy, and makes lane line labeling simpler and easier. The lane line labels obtained by the present application can be used for training lane line extraction models or for building high-precision maps. BRIEF DESCRIPTION OF THE DRAWINGS
[0058] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the prior art description are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings without creative work.
[0059] Figure 1 A schematic diagram of a process flow of a lane marking method provided in an embodiment of the present application;
[0060] Figure 2 A schematic diagram of the annotation interface provided in an embodiment of the present application;
[0061] Figure 3 A schematic diagram of the structure of a lane marking device provided in an embodiment of the present application.
[0062] Figure 4 A schematic diagram of the structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0063] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0064] It should be noted that the terms "including" and "having" and any variations thereof in the embodiments of the present application and the accompanying drawings are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device comprising a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units that are not listed, or may optionally include other steps or units that are inherent to these processes, methods, products or devices.
[0065] The embodiment of the present application discloses a lane line labeling method, device and computer equipment, which labels lane lines according to requirements, reduces the labeling cost without losing accuracy, and the labeled lane line labels can be used for training lane line extraction models or for building high-precision maps. The embodiment of the present application is described in detail below.
[0066] Figure 1 A schematic diagram of a lane marking method provided in an embodiment of the present application is shown. Figure 1 As shown, the purpose of this application is to predict the running trajectory of each target vehicle.
[0067] refer to Figure 1 , a lane marking method provided in this application includes:
[0068] S1, obtaining initial three-dimensional point cloud data corresponding to the road to be annotated; the initial three-dimensional point cloud data at least includes three-dimensional coordinate information of each target point; the three-dimensional point cloud data may also include: the driving direction of the vehicle.
[0069] This step can be performed by collecting environmental data from a real-world vehicle equipped with a laser radar, camera, GPS, and IMU. The camera data is processed through an image segmentation algorithm to obtain lane line related information, and then combined with the laser radar to obtain the three-dimensional point cloud data of the lane line. GPS and IMU can provide accurate location information and attitude information. The three-dimensional point cloud data includes the three-dimensional coordinate information of the surrounding environment.
[0070] After acquiring the initial three-dimensional point cloud data, the present application performs road segmentation processing on the initial three-dimensional point cloud data according to a preset road segment length to obtain initial three-dimensional point cloud data corresponding to each road segment.
[0071] The length of the road segments can be segmented starting from the intersection and can be adjusted according to the actual situation. During the entire segmentation process, the length of the road segments is consistent.
[0072] S2, shielding the height information of each target point in the initial three-dimensional point cloud data to obtain the two-dimensional projection coordinates of each target point;
[0073] After obtaining the initial 3D point cloud data, the application shields the height information and obtains the 2D projection coordinates of the target points. The 2D projection coordinates projected onto the same plane will be more concentrated on or near the lane line, so if the target points on the lane lines of the same category can form a lane line.
[0074] S3, inputting the two-dimensional projection coordinates of each target point into a labeling tool, displaying a target image including each target point, and receiving a labeling result input for the target image; the labeling result at least includes a lane line category;
[0075] Before the two-dimensional projection coordinates of each target point are input into the annotation tool, the method further includes: determining the lane line direction according to the vehicle driving direction in the initial three-dimensional point cloud data.
[0076] The S3 step of the present application can input the two-dimensional projection coordinates of each target point into the labeling tool, and display a target image including each target point according to the lane line direction.
[0077] In order to speed up the annotation process, when generating an image based on the 2D projection coordinates, only one-tenth of the 2D projection coordinates are randomly selected for visualization. This can save time to the greatest extent without affecting the shape of the lane line. Since the lane line has a direction, it is necessary to mark the points according to the direction of the lane line when annotating. This application marks the direction of the lane line when visualizing the lane line image, as shown above. Figure 2 The arrows in the figure show the lane line directions generated based on the driving trajectory of the true value vehicle.
[0078] S4, for each target point in the target image, determine the height information of each target point according to the initial three-dimensional point cloud data, and obtain each three-dimensional target point including the labeling result.
[0079] It is worth noting that the final three-dimensional target points of this application contain height information and are marked with lane line categories.
[0080] As an optional implementation of the present application, as an optional implementation of the present application, the step of shielding the height information of each target point in the initial three-dimensional point cloud data and obtaining the two-dimensional projection coordinates of each target point includes:
[0081] Selecting target three-dimensional point cloud data of a preset scale from the initial three-dimensional point cloud data;
[0082] The height information of each target point in the target three-dimensional point cloud data is shielded to obtain the two-dimensional projection coordinates of each target point.
[0083] As an optional implementation of the present application, the steps of inputting the two-dimensional projection coordinates of each target point into a labeling tool, displaying a target image including each target point, and receiving a labeling result input for the target image include:
[0084] Input the 2D projection coordinates of each target point into the annotation tool, select each target point located in the middle of the lane line according to the direction of the lane line, and display the target image including each target point;
[0085] The marking tool is used to mark the lane line category for each target point in the target image so that the target points of the same category automatically form a lane line, and the marking result for the target image input fed back from the marking tool is received.
[0086] After running the annotation tool code, the 2D projection coordinates will pop up. First, click the mouse to select the annotation point. Figure 2 , the marking point should be selected in the middle of the lane line (gray part) as much as possible, and the marking point needs to be marked in the direction indicated by the arrow. To ensure the integrity of the lane line, marking points must be added at both ends. You can choose manual marking. After marking, press the combination of 0 to 9 on the keyboard to assign the lane line category, such as dotted line, virtual lane line, connecting line single solid line, double solid line, etc.). Finally, press the keyboard n to complete the marking, and the marked points will automatically connect to form a lane line.
[0087] As an optional implementation manner of the present application, after receiving the annotation result for the target image input fed back from the annotation tool, the method further includes:
[0088] The lane lines automatically formed by the target points of the same category are normalized to obtain the normalized lane lines;
[0089] The lane lines processed in a standardized manner are encrypted using an encryption function according to the length of the line segments to obtain encrypted lane lines.
[0090] After receiving the annotation results for the target image input from the annotation tool, the application will normalize the lane line, encrypt the lane line using a function according to the length of the line segment, and then annotate the next lane line. If the marking is wrong or the category is wrong, press the keyboard z to undo the annotation of the previous lane line.
[0091] As an optional implementation of the present application, for each target point in the target image, determining the height information of each target point according to the initial three-dimensional point cloud data, and obtaining each three-dimensional target point including the annotation result includes:
[0092] For each target point in the target image, determine the coordinate information of each candidate point within a preset distance range with the target point as the center;
[0093] For each candidate point, in the initial three-dimensional point cloud data, a three-dimensional point whose x-coordinate and y-coordinate are the same as those of the candidate point is searched, and the average value of the height information of each three-dimensional point is used as the height information of the target point to obtain each three-dimensional target point containing the annotation results.
[0094] Because the lane lines are annotated in the target image generated by the two-dimensional projection coordinates, the annotated lane line labels do not have height information, and the height information needs to be restored in combination with the original three-dimensional point cloud information. For example, for each annotated point, take the point as the center and find the two-dimensional coordinate point 20cm around the point; then find the corresponding point in the original three-dimensional point cloud, get its height information and take the average value as the height information of the point.
[0095] As an optional implementation manner of the present application, after determining the coordinate information of each candidate point within a preset distance range with the target point as the center, the method further includes:
[0096] When no candidate point is found, a preset number of adjacent points before and after the target point are searched on the lane line, and the height information of the target point is determined by linear interpolation based on the height information of each adjacent point.
[0097] If no candidate point is found, that is, if there are no other points near the point, a preset number of adjacent points before and after the target point can be found on the lane line. Then, based on the height information of each adjacent point, linear interpolation is used to assign its height to determine the height information of the target point.
[0098] like Figure 3 As shown, the present application provides a lane marking device comprising:
[0099] The acquisition module 31 is configured to acquire initial three-dimensional point cloud data corresponding to the road to be annotated; the initial three-dimensional point cloud data at least includes three-dimensional coordinate information of each target point;
[0100] A shielding module 32 is configured to shield the height information of each target point in the initial three-dimensional point cloud data to obtain the two-dimensional projection coordinates of each target point;
[0101] The annotation module 33 is configured to input the two-dimensional projection coordinates of each target point into an annotation tool, display a target image including each target point, and receive an annotation result input for the target image; the annotation result at least includes a lane line category;
[0102] The determination module 34 is configured to determine the height information of each target point in the target image according to the initial three-dimensional point cloud data, and obtain each three-dimensional target point including the labeling result.
[0103] Optionally, the embodiment of the present application provides a lane marking device further comprising:
[0104] The segmentation module is configured to perform road segmentation processing on the initial three-dimensional point cloud data according to a preset road segment length to obtain initial three-dimensional point cloud data corresponding to each road segment.
[0105] Optionally, the embodiment of the present application provides a lane marking device further comprising: a direction determination module configured to determine the lane direction according to the vehicle driving direction in the initial three-dimensional point cloud data;
[0106] Optionally, the marking module 33 is configured to:
[0107] The two-dimensional projection coordinates of each target point are input into the annotation tool, and a target image including each target point is displayed according to the lane line direction.
[0108] Optionally, the marking module 33 is configured to:
[0109] Input the 2D projection coordinates of each target point into the annotation tool, select each target point located in the middle of the lane line according to the direction of the lane line, and display the target image including each target point;
[0110] The marking tool is used to mark the lane line category for each target point in the target image so that the target points of the same category automatically form a lane line, and the marking result for the target image input fed back from the marking tool is received.
[0111] Optionally, the embodiment of the present application provides a lane marking device further comprising: a processing module configured to:
[0112] The lane lines automatically formed by the target points of the same category are normalized to obtain normalized lane lines; the normalized lane lines are encrypted by using an encryption function according to the length of the line segments to obtain encrypted lane lines.
[0113] Optionally, the shielding module 32 is configured as follows:
[0114] Selecting target three-dimensional point cloud data of a preset scale from the initial three-dimensional point cloud data;
[0115] The height information of each target point in the target three-dimensional point cloud data is shielded to obtain the two-dimensional projection coordinates of each target point.
[0116] Optionally, the determination module 34 is configured to:
[0117] For each target point in the target image, determine the coordinate information of each candidate point within a preset distance range with the target point as the center;
[0118] For each candidate point, in the initial three-dimensional point cloud data, a three-dimensional point whose x-coordinate and y-coordinate are the same as those of the candidate point is searched, and the average value of the height information of each three-dimensional point is used as the height information of the target point to obtain each three-dimensional target point containing the annotation results.
[0119] Optionally, the embodiment of the present application provides a lane marking device further comprising: a search module configured to:
[0120] When no candidate point is found, a preset number of adjacent points before and after the target point are searched on the lane line, and the height information of the target point is determined by linear interpolation based on the height information of each adjacent point.
[0121] Next, a computer device provided by an embodiment of the present application is introduced. Figure 4 , Figure 4 A schematic diagram of a computer device provided in an embodiment of the present application, the computer device comprising:
[0122] one or more processors 40;
[0123] The processor 40 is coupled to a storage device 41, and the storage device 41 is used to store one or more programs.
[0124] When the one or more programs are executed by the one or more processors 40, the electronic device implements the following Figure 1 to Figure 2 A technical solution for lane marking.
[0125] The present application also provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the following is achieved: Figure 1 A technical solution for the lane marking method.
[0126] The present application provides a lane line labeling method, device and computer equipment, which obtains the initial three-dimensional point cloud data corresponding to the road to be labeled; shields the height information of each target point in the initial three-dimensional point cloud data, and obtains the two-dimensional projection coordinates of each target point; inputs the two-dimensional projection coordinates of each target point into the labeling tool, displays the target image including each target point, and receives the labeling results input for the target image; for each target point in the target image, determines the height information of each target point according to the initial three-dimensional point cloud data, and obtains each three-dimensional target point including the labeling results. The present application can realize lane line labeling only on two-dimensional images according to the needs, reduces the labeling cost without losing accuracy, and makes lane line labeling simpler and easier. The lane line labels obtained by the present application can be used for training lane line extraction models or for building high-precision maps.
[0127] The present application provides a computer program product, the computer program product comprising a computer program, the computer program being executed by a processor to implement the following Figure 1 A technical solution for the lane marking method.
[0128] Those skilled in the art can understand that the accompanying drawings are only schematic diagrams of one embodiment, and the modules or processes in the accompanying drawings are not necessarily required to implement the present application.
[0129] Those skilled in the art can understand that the modules in the device in the embodiment can be distributed in the device in the embodiment according to the description of the embodiment, or can be changed accordingly and located in one or more devices different from the embodiment. The modules in the above embodiment can be combined into one module, or can be further divided into multiple sub-modules.
[0130] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.
Claims
1. A lane marking method, characterized in that: The method comprises: Acquire initial three-dimensional point cloud data corresponding to the road to be annotated; the initial three-dimensional point cloud data at least includes three-dimensional coordinate information of each target point; Shielding the height information of each target point in the initial three-dimensional point cloud data to obtain the two-dimensional projection coordinates of each target point; Inputting the two-dimensional projection coordinates of each target point into a labeling tool, displaying a target image including each target point, and receiving a labeling result input for the target image; the labeling result at least includes a lane line category; For each of the target points in the target image, the height information of each of the target points is determined according to the initial three-dimensional point cloud data to obtain each of the three-dimensional target points including the labeling results.
2. The method according to claim 1, characterized in that After obtaining the initial three-dimensional point cloud data corresponding to the road to be annotated, the method further includes: According to the preset road segment length, the initial three-dimensional point cloud data is processed into road segments to obtain initial three-dimensional point cloud data corresponding to each road segment.
3. The method according to claim 2, characterized in that The three-dimensional point cloud data also includes: the driving direction of the vehicle; before inputting the two-dimensional projection coordinates of each target point into the annotation tool, the method also includes: Determining a lane line direction according to the vehicle driving direction in the initial three-dimensional point cloud data; The step of inputting the two-dimensional projection coordinates of each target point into the annotation tool to display a target image including each target point comprises: The two-dimensional projection coordinates of each target point are input into the annotation tool, and a target image including each target point is displayed according to the lane line direction.
4. The method according to claim 3, characterized in that The steps of inputting the two-dimensional projection coordinates of each target point into a labeling tool, displaying a target image including each target point, and receiving a labeling result input for the target image include: Input the 2D projection coordinates of each target point into the annotation tool, select each target point located in the middle of the lane line according to the direction of the lane line, and display the target image including each target point; The marking tool is used to mark the lane line category for each target point in the target image so that the target points of the same category automatically form a lane line, and the marking result for the target image input fed back from the marking tool is received.
5. The method according to claim 4, characterized in that After receiving the annotation result fed back from the annotation tool for the target image input, the method further includes: The lane lines automatically formed by the target points of the same category are normalized to obtain the normalized lane lines; The lane lines processed in a standardized manner are encrypted using an encryption function according to the length of the line segments to obtain encrypted lane lines.
6. The method according to claim 1, characterized in that The step of shielding the height information of each target point in the initial three-dimensional point cloud data to obtain the two-dimensional projection coordinates of each target point comprises: Selecting target three-dimensional point cloud data of a preset scale from the initial three-dimensional point cloud data; The height information of each target point in the target three-dimensional point cloud data is shielded to obtain the two-dimensional projection coordinates of each target point.
7. The method according to claim 1, characterized in that The step of determining the height information of each target point in the target image according to the initial three-dimensional point cloud data to obtain each three-dimensional target point including the annotation result comprises: For each target point in the target image, determine the coordinate information of each candidate point within a preset distance range with the target point as the center; For each candidate point, in the initial three-dimensional point cloud data, a three-dimensional point whose x-coordinate and y-coordinate are the same as those of the candidate point is searched, and the average value of the height information of each three-dimensional point is used as the height information of the target point to obtain each three-dimensional target point containing the annotation results.
8. The method according to claim 7, characterized in that After determining the coordinate information of each candidate point within a preset distance range with the target point as the center, the method further includes: When no candidate point is found, a preset number of adjacent points before and after the target point are searched on the lane line, and the height information of the target point is determined by linear interpolation based on the height information of each adjacent point.
9. A lane marking device, characterized in that: The device comprises: An acquisition module is configured to acquire initial three-dimensional point cloud data corresponding to the road to be annotated; the initial three-dimensional point cloud data at least includes three-dimensional coordinate information of each target point; A shielding module is configured to shield the height information of each target point in the initial three-dimensional point cloud data to obtain the two-dimensional projection coordinates of each target point; A labeling module is configured to input the two-dimensional projection coordinates of each target point into a labeling tool, display a target image including each target point, and receive a labeling result input for the target image; the labeling result at least includes a lane line category; The determination module is configured to determine the height information of each target point in the target image according to the initial three-dimensional point cloud data, and obtain each three-dimensional target point including the marking result.
10. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are coupled; The memory is used to store one or more computer instructions; The processor is used to execute the one or more computer instructions to implement the lane marking method as described in any one of claims 1-8.