Road marking detection method, vehicle, equipment and product

By performing local partial detection of road area images, the position and type of longitudinal deceleration markings are determined, which solves the problem of map missing caused by irregular shapes of road markings, and improves the accuracy of maps and the effectiveness of simulation tests.

CN119091410AInactive Publication Date: 2024-12-06BEIJING JIDU TECH CO LTD
View PDF 10 Cites 0 Cited by

Patent Information

Application Number
CN202411562551.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-11-04
Publication Date
2024-12-06
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

When building high-precision maps, the shape of the road markings is irregular, especially the longitudinal deceleration markings are difficult to accurately detect, resulting in the missing road elements in the actual scene in the map, affecting the effectiveness of simulation tests.

Method used

By obtaining the road area image of the lane line, classifying and detecting the local road area image, determining the starting point position, end point position and marking type of the longitudinal deceleration mark, combining vehicle positioning data to build a target map for automatic driving simulation testing.

Benefits of technology

It improves the detection accuracy and consistency of longitudinal deceleration marks, ensures the authenticity and accuracy of the map, and improves the reliability of simulation tests.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119091410A_ABST
    Figure CN119091410A_ABST
Patent Text Reader

Abstract

The invention provides a road marking detection method, a vehicle, equipment and a product, and relates to the technical field of image processing, and the method comprises the steps: obtaining a road region image containing a lane line; performing classification detection on each local road area image of the road area image, and determining a classification detection result of the local road area images; the local road area images are arranged in the road area image in the lane line direction; and according to the classification detection result of each local road area image, determining the starting point position and the ending point position of the overall longitudinal deceleration marking line and the marking line type of the overall longitudinal deceleration marking line. According to the embodiment of the invention, the information of the overall longitudinal deceleration marking line of the whole road area can be accurately identified, the information comprises the starting position and the ending position of the overall longitudinal deceleration marking line and the marking line type of the overall longitudinal deceleration marking line, and the conditions that the longitudinal deceleration marking line is difficult to identify and extract and the identification position is not accurate are effectively reduced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present disclosure relates to the field of image processing technology, and in particular to a road marking detection method, vehicle, equipment and product. Background Art

[0002] In the process of vehicle driving, in order to facilitate vehicle planning and control, most of the time, maps are constructed to assist in controlling the vehicle. For high-precision maps, they mainly focus on position accuracy, that is, the absolute position deviation from the actual target.

[0003] In addition to assisting in controlling vehicles, HD maps are mostly used to build large-scale simulation scenarios and for simulation testing of autonomous driving. This requires that in addition to ensuring the position accuracy, the elements of HD maps also need to fully identify various road elements in the actual scene, so as to ensure the consistency between the map and the actual scene, which helps to ensure the authenticity of the simulation results. However, when building HD maps, some road markings have irregular shapes. For example, the longitudinal deceleration markings are long and jagged, which makes it difficult to extract the shape. This may result in the absence of road elements in the actual scene in the HD map, thus affecting the effect of the simulation test. Summary of the invention

[0004] The embodiments of the present disclosure at least provide a road marking detection method, vehicle, equipment and product.

[0005] The present disclosure provides a road marking detection method, the method comprising: Obtain a road area image including lane lines; Performing classification detection on each partial road area image of the road area image, determining a classification detection result of the partial road area image, the classification detection result indicating whether there is a longitudinal deceleration marking in the partial road area image, and the marking type of the longitudinal deceleration marking, the marking type indicating whether there are longitudinal deceleration markings on the left and right sides of the lane line respectively; the partial road area images are arranged along the lane line direction in the road area image; According to the classification detection results of each of the local road area images, the starting position and the end position of the overall longitudinal speed reduction marking line, and the marking type of the overall longitudinal speed reduction marking line are determined.

[0006] In an optional implementation manner, the acquiring of a road area image including lane lines includes: Acquire continuous multi-frame road area images taken when the vehicle is traveling on the target road; A lane line detection model is used to perform lane line detection on the multiple frames of road area images respectively to obtain each frame of road area image containing a lane line.

[0007] In an optional implementation manner, before classifying and detecting each local road area image of the road area image, the method further includes: The road area image is cut into a plurality of local road area images along the lane line direction according to a preset distance interval.

[0008] In an optional implementation manner, the step of cutting the road area image into a plurality of local road area images along a lane line direction at preset distance intervals includes: Determine a plurality of target points on the lane line of the road area image at preset distance intervals, and crop a local road area image of a preset size with each target point as the center; The starting point and the end point of the overall longitudinal deceleration marking are the positions of the target points.

[0009] In an optional implementation manner, determining the starting point position and the end point position of the overall longitudinal deceleration marking line and the marking type of the overall longitudinal deceleration marking line according to the classification detection results of each of the local road area images includes: Determine a road area image corresponding to the same lane line from each frame of road area image; According to the classification detection results of each of the local road area images in the road area image corresponding to the same lane line, the starting position and the end position of the overall longitudinal deceleration marking line and the marking type of the overall longitudinal deceleration marking line are determined.

[0010] In an optional implementation manner, determining the starting point position and the end point position of the overall longitudinal deceleration marking line and the marking type of the overall longitudinal deceleration marking line according to the classification detection results of each of the local road area images in the road area image corresponding to the same lane line includes: Determine a plurality of continuous local road area images including longitudinal speed reduction markings according to the classification detection results of the local road area images in the road area image corresponding to the same lane line; The starting point and the end point of the overall longitudinal deceleration marking are determined according to the position information corresponding to the multiple continuous local road area images, and the marking type of the overall longitudinal deceleration marking is determined according to the classification detection results of the multiple continuous local road area images.

[0011] In an optional implementation manner, the classification detection result is also used to indicate the confidence level of each marking type corresponding to the longitudinal deceleration marking; The step of determining the marking type of the overall longitudinal speed reduction marking according to the classification detection results of the plurality of continuous local road area images comprises: Determining, based on the classification detection results of the plurality of continuous partial road area images, whether, among the plurality of continuous partial road area images, the classification detection results of the partial road area images exceeding a preset proportion indicate the same type of markings, and the confidence level of the indication is greater than a preset confidence threshold; If yes, it is determined that the marking type of the overall longitudinal deceleration marking is the same marking type.

[0012] In an optional implementation manner, the method further includes: If there are no classification detection results of the partial road area images exceeding a preset ratio indicating the same marking type, and the indicated confidence is greater than a preset confidence threshold, determining the marking type of the overall longitudinal deceleration marking according to the relative positions of the lane lines corresponding to the plurality of continuous partial road area images; The lane line relative position indicates that the lane line belongs to one of a left lane line on the road, a right lane line on the road, and a middle lane line on the road.

[0013] In an optional implementation manner, determining the road area image corresponding to the same lane line from each frame of road area image includes: Determine, based on lane line position information in two adjacent frames of road area images, whether lane lines in two adjacent frames of road area images overlap in a lane line direction, and whether a distance in a direction perpendicular to the lane lines is less than a preset distance; If there is overlap in the lane line direction, and the distance in the direction perpendicular to the lane line is less than a preset distance, it is determined that the two adjacent frames of road area images correspond to the same lane line.

[0014] In an optional implementation manner, after determining the starting position and the end position of the overall longitudinal deceleration marking, and the marking type of the overall longitudinal deceleration marking, the method further includes: Based on the starting position and the end position of the overall longitudinal deceleration marking, the marking type of the overall longitudinal deceleration marking, and the collected road image and vehicle positioning data of the target road, a target map for simulation testing is constructed; Based on the target map, a simulation test of autonomous driving is performed on the vehicle.

[0015] The present disclosure also provides a vehicle, including a controller, wherein the controller includes: a memory for storing machine-readable instructions; and The processor is used to call and execute the machine-readable instructions stored in the memory to implement the steps in any possible implementation of the above-mentioned road marking detection method.

[0016] An embodiment of the present disclosure also provides an electronic device, including: a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate via the bus, and when the machine-readable instructions are executed by the processor, the steps in any possible implementation of the above-mentioned road marking detection method are performed.

[0017] The embodiments of the present disclosure further provide a computer program product, on which a computer program is stored. When the computer program is executed by a processor, the steps in any possible implementation of the above-mentioned road marking detection method are executed.

[0018] The road marking detection method, vehicle, equipment and product provided by the embodiments of the present disclosure can obtain a road area image containing lane lines; then classify and detect each local road area image arranged along the lane line direction in the road area image to determine the classification detection results of each local road area image, the classification detection results indicate whether there are longitudinal speed reduction markings in the local road area image, and the marking type of the existing longitudinal speed reduction markings, and the marking type indicates whether there are longitudinal speed reduction markings on the left and right sides of the lane line respectively; and then determine the starting position and end position of the overall longitudinal speed reduction marking, and the marking type of the overall longitudinal speed reduction marking according to the classification detection results of each local road area image.

[0019] In this way, for the road area image containing the lane lines, each local road area image in the road area image is classified and detected, so that refined road marking detection can be performed in a smaller range, which helps to improve the accuracy of road marking detection. Combined with the classification detection results of each local road area image, the information of the overall longitudinal speed reduction marking of the entire road area can be accurately identified, including the starting position and end position of the overall longitudinal speed reduction marking, and the marking type of the overall longitudinal speed reduction marking, which effectively reduces the difficulty in identifying and extracting the longitudinal speed reduction marking and the inaccurate identification position, so that the detected longitudinal speed reduction marking is highly consistent with the actual situation, which helps to ensure the authenticity and accuracy of the map when constructing the map later.

[0020] It should be understood that the above general description and the following detailed description are merely exemplary and explanatory, and are not intended to limit the technical solutions of the present disclosure.

[0021] In order to make the above-mentioned objectives, features and advantages of the present disclosure more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the following is a brief introduction to the drawings required for use in the embodiments. The drawings herein are incorporated into the specification and constitute a part of the specification. These drawings illustrate embodiments consistent with the present disclosure and are used together with the specification to illustrate the technical solutions of the present disclosure. It should be understood that the following drawings only illustrate certain embodiments of the present disclosure and should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can also be obtained based on these drawings without creative work.

[0023] Figure 1 A flow chart of a road marking detection method provided by an embodiment of the present disclosure is shown; Figure 2 A schematic diagram of determining a local road area image provided by an embodiment of the present disclosure is shown; Figure 3 Another schematic diagram of determining a local road area image provided by an embodiment of the present disclosure is shown; Figure 4a~Figure 4d A schematic diagram showing a classification detection result of a local road area image provided by an embodiment of the present disclosure; Figure 5 A schematic diagram of determining a road area image corresponding to the same lane line provided by an embodiment of the present disclosure is shown; Figure 6 A flow chart of a road marking detection and simulation test method provided by an embodiment of the present disclosure is shown; Figure 7 A schematic structural diagram of an electronic device provided by an embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0024] In order to make the purpose, technical scheme and advantages of the embodiments of the present disclosure clearer, the technical scheme in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only part of the embodiments of the present disclosure, rather than all of the embodiments. The components of the embodiments of the present disclosure generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present disclosure provided in the drawings is not intended to limit the scope of the present disclosure for protection, but merely represents the selected embodiments of the present disclosure. Based on the embodiments of the present disclosure, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present disclosure.

[0025] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0026] The term "and / or" herein only describes an association relationship, indicating that three relationships may exist. For example, A and / or B may represent the following three situations: A exists alone, A and B exist at the same time, and B exists alone. In addition, the term "at least one" herein represents any combination of at least two of any one or more of a plurality of. For example, including at least one of A, B, and C may represent including any one or more elements selected from the set consisting of A, B, and C.

[0027] Research has found that in addition to assisting in controlling vehicles, high-precision maps are mostly used to build large-scale simulation scenarios and for simulation testing of autonomous driving. This requires that in addition to ensuring position accuracy, the elements of high-precision maps also need to comprehensively identify various road elements in actual scenarios, thereby ensuring the consistency between the map and the actual scenario, which helps to ensure the authenticity of the simulation results. However, when constructing high-precision maps, some road markings have irregular shapes and are difficult to identify and extract. This may result in the absence of road elements in the actual scenario in the high-precision map, thus affecting the effect of the simulation test. For example, for longitudinal deceleration markings, they are long and serrated, making them difficult to detect accurately. In this way, it is very likely that there are longitudinal deceleration markings in the actual scenario, but because the longitudinal deceleration markings are not identified or are inaccurately identified, the longitudinal deceleration markings do not exist in the constructed map.

[0028] Based on the above research, the present disclosure provides a road marking detection method, which can accurately identify the longitudinal speed reduction markings in the entire road area, and can determine the starting position and end position of the longitudinal speed reduction markings, as well as the marking type of the longitudinal speed reduction markings, effectively reducing the difficulty in identifying and extracting the longitudinal speed reduction markings and the inaccurate identification position.

[0029] To facilitate understanding of this embodiment, a road marking detection method disclosed in the embodiment of the present disclosure is first introduced in detail. The execution subject of the road marking detection method provided in the embodiment of the present disclosure is generally an electronic device with certain computing capabilities. In this embodiment, the electronic device can be a server. Among them, the server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud storage, big data and artificial intelligence platforms.

[0030] In other implementations, the electronic device may also be a terminal device or other processing device, and the terminal device may be a user equipment (UE), a mobile device, a terminal, a computing device, a vehicle-mounted device, etc. Other processing devices may be devices including a processor and a memory, which are not limited here. In some possible implementations, the road marking detection method may be implemented by a processor calling a computer-readable instruction stored in a memory.

[0031] A road marking detection method provided by an embodiment of the present disclosure is described below in conjunction with the accompanying drawings.

[0032] See also Figure 1 As shown, it is a flow chart of a road marking detection method provided by an embodiment of the present disclosure, such as Figure 1 As shown in , the road marking detection method provided by the embodiment of the present disclosure includes steps S101 to S103, wherein: S101: Acquire a road area image including lane lines.

[0033] Here, the lane line is located on the road surface of the target road and is used to assist the vehicle in traveling through the target road.

[0034] In practical applications, a vehicle is often provided with a camera, and the image of the road area can be acquired through the vehicle camera.

[0035] Specifically, a continuous multi-frame road area image taken when the vehicle is traveling on the target road can be obtained; a lane line detection model is used to perform lane line detection on the multi-frame road area images respectively to obtain each frame road area image containing the lane line.

[0036] In the above steps, when the vehicle is traveling on the target road, multiple frames of road area images can be continuously captured by the vehicle camera at preset time intervals, and then the lane lines can be detected from the multiple frames of road area images through a lane line detection model, thereby obtaining each frame of road area image containing the lane lines.

[0037] Exemplarily, the vehicle camera can capture 30 frames of road area images per second.

[0038] It can be understood that at least part of the road area images in the multiple frames of road area images contain lane lines.

[0039] Among them, the lane line detection model is used to detect whether there is a lane line in the road area image. It can be understood that if it exists, the detection result is the lane line attribute information of the lane line in the corresponding road area image, and the lane line attribute information includes but is not limited to the position and direction of the lane line. If it does not exist, the detection result is empty.

[0040] Optionally, the lane line detection model can perform lane line detection on an area within a preset range. Exemplarily, the lane line detection model can be used to extract lane lines within 50 meters in front of the vehicle.

[0041] Here, the lane line detection model may be pre-trained. Specifically, the lane line detection model may be trained through a plurality of sample road area images and lane line annotation information corresponding to each sample road area image.

[0042] Optionally, in order to ensure the model training effect, the lane line marking information can be obtained by manual marking.

[0043] In other implementations, lane lines may be detected from the road area image by other means, such as a target positioning algorithm.

[0044] In this way, continuous multi-frame road area images can be obtained by taking pictures with the camera set by the vehicle itself, and then lane line detection can be performed on the multi-frame road area images through the deep learning lane line detection model to obtain each frame of road area images containing lane lines, which helps to improve the accuracy of the road area images containing lane lines.

[0045] It should be noted that vehicles are often equipped with multiple sensor components such as cameras, radars, global positioning systems (GPS), inertial measurement units (IMU), etc. Before taking multiple consecutive frames of road area images through the vehicle camera, the multiple sensor components can be calibrated and registered in advance to obtain calibration and registration parameters. The calibration and registration parameters include the conversion relationship between the camera coordinate system and the vehicle coordinate system, and the conversion relationship between the vehicle coordinate system and the world coordinate system.

[0046] In some possible implementations, the acquired road area image may be a bird's-eye view, that is, an image under a bird's-eye-view (BEV) perspective.

[0047] In practical applications, vehicles mostly shoot the multi-frame road area images from a forward perspective, and lane lines are mostly printed on the ground. Therefore, the lane lines may be presented at an inclined angle in the multi-frame road area images. By converting to the BEV perspective, the lane lines are presented more regularly in the multi-frame road area images under the BEV perspective.

[0048] Specifically, when the perspective conversion is performed, multiple frames of road area images under the BEV perspective can be obtained based on the correspondence between the vehicle camera shooting perspective and the BEV perspective.

[0049] In this way, considering that the target type of marking is a ground printed sign, obtaining the road area image from the BEV's perspective can help focus more on the lane lines and effectively improve the accuracy of lane line detection.

[0050] Correspondingly, when the lane line detection model is obtained through training, the sample road area image may specifically be a sample road area image from a BEV perspective, and then the road area image from a BEV perspective is annotated to obtain lane line annotation information.

[0051] S102: Classify and detect each local road area image of the road area image to determine the classification detection result of the local road area image, wherein the classification detection result indicates whether there is a longitudinal speed reduction marking in the local road area image and the type of the longitudinal speed reduction marking, wherein the marking type indicates whether there are longitudinal speed reduction markings on the left and right sides of the lane line respectively; the local road area images are arranged along the lane line direction in the road area image.

[0052] Here, each partial road area image is arranged along the lane line direction in the road area image, that is, each partial road area image is obtained by dividing the road area image along the lane line direction. In the case where the road area image including the lane line is a continuous plurality of frames, each frame of the road area image can be divided into a plurality of partial road area images, and each partial road area image for classification detection is obtained by dividing a plurality of continuous frames of road area images.

[0053] In a possible implementation manner, before classifying and detecting each local road area image of the road area image, the method further includes: The road area image is cut into a plurality of local road area images along the lane line direction according to a preset distance interval.

[0054] In this step, in order to improve the detection accuracy of the longitudinal speed reduction markings, multiple local road area images can be cropped from the road area image along the lane line direction, so that each of the local road area images can be classified and detected later, and refined longitudinal speed reduction marking detection can be performed in a smaller range, which helps to improve the accuracy of longitudinal speed reduction marking detection.

[0055] Specifically, a plurality of target points may be determined on the lane lines of the road area image at preset distance intervals, and a local road area image of a preset size may be cropped with each of the target points as the center.

[0056] Here, when determining the multiple target points, optionally, starting from any end point of the lane line, the multiple target points can be determined on the lane line of the road area image according to the preset distance interval.

[0057] Optionally, a plurality of virtual lines perpendicular to the lane line may be determined according to the preset distance interval, and the intersection point of each virtual line with the lane line may be determined as the target point.

[0058] Exemplarily, the shape of the local road area image includes but is not limited to a rectangle, a square, etc.

[0059] The specific values ​​of the preset distance interval and the preset size may be determined according to actual needs for cropping the local road area image, and are not specifically limited here.

[0060] For example, please refer to Figure 2 , Figure 2 This is a schematic diagram of determining a local road area image provided by an embodiment of the present disclosure. In this example, a longitudinal deceleration marking is used as an example to illustrate the existence of longitudinal deceleration markings on both sides of the lane line. Figure 2 As shown in , the road area image 200 includes a lane line L1. According to preset distance intervals, a target point P1, a target point P2 and a target point P3 are determined on the lane line L1 of the road area image 200, and a local road area image 210 of a preset size is cropped with the target point P1 as the center, a local road area image 220 of a preset size is cropped with the target point P2 as the center, and a local road area image 230 of a preset size is cropped with the target point P3 as the center.

[0061] In another example, you can also refer to Figure 3 , Figure 3 This is another schematic diagram of determining a local road area image provided by an embodiment of the present disclosure. In this example, a case where there are no longitudinal deceleration markings on both sides of the lane line is used as an example. Figure 3 As shown in the figure, the road area image 300 includes a lane line L2. According to preset distance intervals, target points P4, P5 and P6 are determined on the lane line L2 of the road area image 300, and a local road area image 310 of a preset size is cropped with the target point P4 as the center, a local road area image 320 of a preset size is cropped with the target point P5 as the center, and a local road area image 330 of a preset size is cropped with the target point P6 as the center.

[0062] The starting point and the end point of the overall longitudinal deceleration marking are the positions of the target points.

[0063] In this way, multiple target points can be determined on the lane line at preset distance intervals, and then a local road area image of a preset size can be cropped with each target point as the center. This not only ensures the size consistency of each local road area image, but also helps to improve the efficiency of obtaining the local road area image.

[0064] When determining each local road area image of the road area image, a marking line classification detection model may be used to perform classification detection on each local road area image to determine a classification detection result of the local road area image.

[0065] It can be understood that the classification detection results in the embodiment of the present disclosure specifically include four types. Figure 4a~Figure 4d , is a schematic diagram of the classification detection result of the local road area image provided by the embodiment of the present disclosure. Figure 4a As shown in , the classification detection result includes that only the right side of the lane line has a longitudinal deceleration marking; Figure 4b As shown in , the classification detection results include the presence of longitudinal deceleration markings on both the left and right sides of the lane line; Figure 4c As shown in , the classification detection result includes that only the left side of the lane line has a longitudinal deceleration marking; Figure 4d As shown in , the classification detection result includes that there are no longitudinal deceleration markings on the left and right sides of the lane line.

[0066] Here, the road marking classification detection model may be pre-trained. Specifically, the road marking classification detection model may be trained by using a sample local road area image cropped from a sample road area image.

[0067] Among them, the step of cropping the sample local road area image from the sample road area image is similar to the step of cropping the local road area image from the road area image. For details, please refer to the description of the aforementioned embodiment, which will not be repeated here.

[0068] In order to ensure the training effect of the road marking classification detection model, the sample local road area images also include four types of sample local road area images, and each type of sample local road area image corresponds to a classification detection result.

[0069] Optionally, when training the road marking classification detection model, the numbers of the four types of sample local road area images may be set to be the same, thereby ensuring the robustness of the road marking classification detection model.

[0070] S103: Determine the starting point and the end point of the overall longitudinal speed reduction marking, and the marking type of the overall longitudinal speed reduction marking according to the classification detection results of each of the local road area images.

[0071] In this step, the starting point and the end point of the overall longitudinal deceleration marking can be determined according to the position information of each of the local road area images, and the marking type of the overall longitudinal deceleration marking can be determined according to the classification detection results of each of the local road area images.

[0072] Here, the overall longitudinal deceleration marking is the longitudinal deceleration marking corresponding to the entire road area.

[0073] Optionally, a target map for simulation testing may be subsequently constructed according to the starting position and the ending position of the overall longitudinal deceleration marking and the marking type of the overall longitudinal deceleration marking.

[0074] In some possible implementations, the length of the overall longitudinal deceleration marking line may be determined based on the starting position and the ending position of the overall longitudinal deceleration marking line, thereby facilitating better subsequent construction of the target map.

[0075] In order to facilitate the processing of position information, before determining the starting position and end position of the overall longitudinal speed reduction marking and the marking type of the overall longitudinal speed reduction marking based on the classification detection results of each local road area image, the position information of each target point in the same coordinate system can be determined.

[0076] Specifically, the pixel coordinates of each target point in the corresponding road area image can be converted into a unified coordinate system, such as a world coordinate system, to obtain absolute precision coordinates, and the absolute precision coordinates are used as the position information of the target point.

[0077] Specifically, according to the above content, the calibration and registration parameters are obtained by pre-calibration and registration, and the calibration and registration parameters include the conversion relationship between the camera coordinate system and the vehicle coordinate system, and the conversion relationship between the vehicle coordinate system and the world coordinate system. Accordingly, the pixel coordinates of each target point in the corresponding road area image can be converted into absolute precision coordinates based on the calibration and registration parameters and the vehicle positioning position information.

[0078] The vehicle positioning position information refers to the position of the vehicle when the corresponding road area image is captured, which can be specifically determined by data collected by the GPS and IMU set on the vehicle.

[0079] The absolute precision coordinates refer to the coordinates of each target point in the world coordinate system. Optionally, the absolute precision coordinates are in the form of longitude and latitude.

[0080] When the pixel coordinates of each target point are converted into absolute precision coordinate values ​​based on the calibration and registration parameters and the vehicle positioning position information, specifically, the conversion relationship between the camera coordinate system and the vehicle coordinate system can be obtained based on the calibrated camera parameter matrix included in the calibration and registration parameters, so that the pixel coordinates can be converted into the vehicle coordinate system (for example, a coordinate system with the center of the vehicle body as the coordinate origin) to obtain the intermediate coordinates in the vehicle coordinate system. Combined with the vehicle positioning position information, the intermediate coordinates can be converted from the vehicle coordinate system to the world coordinate system to obtain the absolute precision coordinates in the world coordinate system.

[0081] In some possible implementations, determining the starting position and the end position of the overall longitudinal deceleration marking, and the marking type of the overall longitudinal deceleration marking according to the classification detection results of each of the local road area images, includes: Determine a road area image corresponding to the same lane line from each frame of road area image; According to the classification detection results of each of the local road area images in the road area image corresponding to the same lane line, the starting position and the end position of the overall longitudinal deceleration marking line and the marking type of the overall longitudinal deceleration marking line are determined.

[0082] In the above steps, considering that lane lines and longitudinal deceleration markings are generally long in actual scenes, it is difficult to include a complete lane line in one frame of the road area image, so the road area image corresponding to the same lane line can be determined from each frame of the road area image, and then the starting position and end position of the overall longitudinal deceleration marking corresponding to the entire lane line, as well as the marking type of the overall longitudinal deceleration marking, can be determined based on the classification detection results of each of the local road area images in the road area image corresponding to the same lane line. In this way, it is helpful to ensure the integrity and comprehensiveness of the longitudinal deceleration marking detection.

[0083] When determining the road area image corresponding to the same lane line from each frame of road area image, it can be determined whether the two adjacent frames of road area image correspond to the same lane line based on the overlap of the lane lines in the lane line direction and the distance in the direction perpendicular to the lane line.

[0084] Specifically, based on the lane line position information in two adjacent frames of road area images, it can be determined whether the lane lines in the two adjacent frames of road area images overlap in the lane line direction, and whether the distance in the direction perpendicular to the lane lines is less than a preset distance; if there is an overlap in the lane line direction, and the distance in the direction perpendicular to the lane line is less than the preset distance, it is determined that the two adjacent frames of road area images correspond to the same lane line.

[0085] Considering that the vehicle camera has a high shooting frequency, generally 30 frames of road area images can be captured per second, if two adjacent frames of road area images correspond to the same lane line, most of the lane lines in the two adjacent frames of road area images overlap in the lane line direction. Therefore, in this step, it can be determined whether the lane lines in the two adjacent frames of road area images overlap in the lane line direction.

[0086] Considering that the road area image is captured by the vehicle camera while the vehicle is driving on the target road, in actual applications, due to vehicle shaking or vehicle camera shaking, the lane lines in two adjacent frames of road area images may have a small deviation in the direction perpendicular to the lane lines and may not completely overlap. Therefore, in this step, it can be determined whether the distance between the lane lines in the direction perpendicular to the lane lines in two adjacent frames of road area images is less than the preset distance.

[0087] When the lane lines in two adjacent frames of road area images overlap in the lane line direction and the distance perpendicular to the lane line direction is less than a preset distance, it can be determined that the two adjacent frames of road area images correspond to the same lane line.

[0088] The specific value of the preset distance may be determined according to the actual lane line judgment needs and is not specifically limited here.

[0089] For example, please refer to Figure 5 , Figure 5 A schematic diagram of determining a road area image corresponding to the same lane line provided by an embodiment of the present disclosure. Figure 5 As shown in , for two adjacent frames of road area images, the first frame of the road area image includes lane line L3, lane line L5, and lane line L7, and the second frame of the road area image includes lane line L4, lane line L6, and lane line L8.

[0090] In actual scenes, lane lines are often narrow. In order to more accurately determine whether lane lines in two adjacent frames of road area images overlap in the lane line direction and whether the distance in the direction perpendicular to the lane lines is less than the preset distance, Figure 5 It can be seen that the lane lines in two adjacent frames of road area images can be width-expanded to obtain lane lines after width expansion. Through the lane lines after width expansion in two adjacent frames of road area images, it can be clearly determined whether the lane lines in the two adjacent frames of road area images overlap in the lane line direction, and whether the distance in the direction perpendicular to the lane lines is less than the preset distance.

[0091] In this example, combined with Figure 5It can be seen that lane line L3 and lane line L4 correspond to the same lane line, lane line L5 and lane line L6 correspond to the same lane line, and lane line L7 and lane line L8 correspond to the same lane line.

[0092] It should be noted that in this example, for ease of understanding, Figure 5 In the image, the direction of the lane line is indicated by an arrow. In actual scenes, there may or may not be an arrow on the lane line in the road area image.

[0093] In this way, whether two adjacent frames of road area images correspond to the same lane line can be judged based on whether there is overlap in the lane line direction and whether the distance in the direction perpendicular to the lane line is less than a preset distance, thereby helping to quickly and accurately determine the road area image corresponding to the same lane line.

[0094] After determining the road area image corresponding to the same lane line, the starting point and end point of the overall longitudinal deceleration marking, as well as the marking type of the overall longitudinal deceleration marking can be determined based on the classification detection results of each of the local road area images in the road area image corresponding to the same lane line.

[0095] Specifically, a plurality of continuous local road area images including longitudinal speed reduction markings may be determined according to the classification detection results of the local road area images in the road area image corresponding to the same lane line; The starting point and the end point of the overall longitudinal deceleration marking are determined according to the position information corresponding to the multiple continuous local road area images, and the marking type of the overall longitudinal deceleration marking is determined according to the classification detection results of the multiple continuous local road area images.

[0096] It can be understood that if the classification detection results of multiple consecutive local road area images all indicate the existence of longitudinal speed reduction markings, the types of longitudinal speed reduction markings in the multiple consecutive local road area images are often consistent in actual scenarios, and it is basically impossible for one lane line to correspond to two types of longitudinal speed reduction markings at the same time, because the lane line is generally located in the middle of the road, or on both sides of the road (the left side of the road or the right side of the road). When the lane line is located in the middle of the road, there are often longitudinal speed reduction markings on both the left and right sides of the lane line. When the lane line is located on both sides of the road, there are often longitudinal speed reduction markings on the left or right side of the lane line.

[0097] Therefore, the determined plurality of continuous local road area images containing longitudinal deceleration markings correspond to an overall longitudinal deceleration marking.

[0098] In this way, based on the classification detection results of each local road area image in the road area image corresponding to the same lane line, multiple continuous local road area images whose classification detection results are not empty (i.e., containing longitudinal speed reduction markings) can be determined. Then, based on the position information and classification detection results of the multiple continuous local road area images, the starting and ending positions and marking types of the overall longitudinal speed reduction markings can be determined, which helps to ensure the integrity and comprehensiveness of the longitudinal speed reduction marking detection.

[0099] In combination with the foregoing content, it can be known that before determining the starting position and the end position of the overall longitudinal speed reduction marking, as well as the marking type of the overall longitudinal speed reduction marking based on the classification detection results of each of the local road area images, the position information (absolute accuracy coordinates) of each target point is obtained. Here, the position information corresponding to the multiple continuous local road area images can be represented by the position information of the target points in the local road area images.

[0100] In some possible implementations, determining the starting position and the ending position of the overall longitudinal speed reduction marking line according to the position information corresponding to the plurality of continuous local road area images includes: According to the position information corresponding to the multiple continuous local road area images, the position of the target point in the local road area image where the first classification detection result indicates that there is a longitudinal speed reduction marking is determined as the starting position of the overall longitudinal speed reduction marking, and the position of the target point in the local road area image where the first classification detection result indicates that there is no longitudinal speed reduction marking is determined as the end position of the overall longitudinal speed reduction marking.

[0101] Specifically, the classification detection results of the multiple consecutive local road area images can be traversed according to the lane line direction; when the classification detection results corresponding to a preset number of consecutive local road area images all indicate the existence of a longitudinal speed reduction marking, the position of the target point in the local road area image whose first classification detection result indicates the existence of a longitudinal speed reduction marking is determined as the starting position of the overall longitudinal speed reduction marking; when the classification detection results corresponding to a preset number of consecutive local road area images all indicate the absence of a longitudinal speed reduction marking, the position of the target point in the local road area image whose first classification detection result indicates the absence of a longitudinal speed reduction marking is determined as the end position of the overall longitudinal speed reduction marking.

[0102] The specific value of the preset number may be determined according to the actual length of the longitudinal speed reduction marking, and is not specifically limited here.

[0103] In some possible implementations, the classification detection result is also used to indicate the confidence level of each marking type corresponding to the longitudinal deceleration marking.

[0104] The sum of the confidences of the marking types is 1, and the confidence of the marking type indicated by the classification detection result is the largest among the confidences of the marking types.

[0105] Exemplarily, the classification detection result indicates that the type of longitudinal deceleration marking that exists is that the longitudinal deceleration marking exists only on the left side of the lane line, and the confidence level that the longitudinal deceleration marking exists only on the left side of the lane line is 90%, and the confidence levels of the other three types of markings are 5%, 3% and 2%, respectively.

[0106] Accordingly, determining the marking type of the overall longitudinal speed reduction marking according to the classification detection results of the plurality of continuous local road area images includes: Determining, based on the classification detection results of the plurality of continuous partial road area images, whether, among the plurality of continuous partial road area images, the classification detection results of the partial road area images exceeding a preset proportion indicate the same type of markings, and the confidence level of the indication is greater than a preset confidence threshold; If yes, it is determined that the marking type of the overall longitudinal deceleration marking is the same marking type.

[0107] In the above steps, since the multiple continuous local road area images correspond to the overall longitudinal deceleration marking, as long as the classification detection results of the local road area images exceeding a preset proportion indicate the same marking type, and the confidence indicating the same marking type is greater than a preset confidence threshold, it is considered that the marking type of the overall longitudinal deceleration marking is the same marking type.

[0108] The specific values ​​of the preset ratio and the preset confidence threshold can be determined according to the actual marking type and are not specifically limited here.

[0109] Exemplarily, the preset ratio may be 50%, and the preset confidence threshold may be 90%.

[0110] In this way, the marking type of the overall longitudinal deceleration marking can be comprehensively determined based on the marking types and confidence levels indicated by each classification detection result in multiple continuous local road area images, which helps to improve the efficiency of determining the marking type of the overall longitudinal deceleration marking.

[0111] In other possible implementations, due to reasons such as wear of the longitudinal deceleration marking in actual scenarios, blurred road area image shooting, errors in classification detection, etc., there may be a situation where there is no classification detection result of more than a preset proportion of local road area images indicating the same marking type, and the confidence of the indication is greater than a preset confidence threshold. In this case, the marking type of the complete longitudinal deceleration marking can be determined according to the default rules.

[0112] Accordingly, the method further includes: If there are no classification detection results of the partial road area images exceeding a preset ratio indicating the same marking type, and the indicated confidence is greater than a preset confidence threshold, determining the marking type of the overall longitudinal deceleration marking according to the relative positions of the lane lines corresponding to the plurality of continuous partial road area images; The lane line relative position indicates that the lane line belongs to one of a left lane line on the road, a right lane line on the road, and a middle lane line on the road.

[0113] Here, when the relative position of the lane line indicates that the lane line belongs to the left lane line of the road, the marking type of the overall longitudinal deceleration marking is determined as a longitudinal deceleration marking only on the left side of the lane line; when the relative position of the lane line indicates that the lane line belongs to the right lane line of the road, the marking type of the overall longitudinal deceleration marking is determined as a longitudinal deceleration marking only on the right side of the lane line; when the relative position of the lane line indicates that the lane line belongs to the middle lane line of the road, the marking type of the overall longitudinal deceleration marking is determined as a longitudinal deceleration marking on both the left and right sides of the lane line.

[0114] In this way, when there are no classification detection results of local road area images indicating the same marking type exceeding a preset proportion, and the indicated confidence is greater than a preset confidence threshold, the marking type of the overall longitudinal deceleration marking can be determined based on the relative positions of the lane lines corresponding to multiple consecutive local road area images, which helps to ensure that the marking type of the overall longitudinal deceleration marking can be identified.

[0115] After determining the starting position and the end position of the overall longitudinal deceleration marking, as well as the marking type of the overall longitudinal deceleration marking, the starting position, the end position, and the marking type of the overall longitudinal deceleration marking can be verified to determine whether to adopt the determined starting position, the end position, and the marking type of the overall longitudinal deceleration marking for subsequent map construction.

[0116] Specifically, the road point cloud data scanned by the vehicle radar may be obtained, and the starting position and the end position of the overall longitudinal speed reduction marking and the marking type may be verified based on the road point cloud data.

[0117] Optionally, manual verification can be performed by processing the road point cloud data and the road area image through a target tool to map the point cloud onto the road area image, so that it can be intuitively seen whether the starting position, end position and marking type of the overall longitudinal deceleration marking are deviated from the road point cloud data.

[0118] Optionally, automatic verification can also be performed. Specifically, the road point cloud data scanned by the vehicle radar can be obtained and converted into a point cloud image; the lane line detection model is used to detect the first coordinate range of the image area where the lane line is located in the world coordinate system from the point cloud image; according to the overlap between the second coordinate range of the image area where the lane line is located in the world coordinate system detected from the road area image and the first coordinate range, it is determined whether to adopt the determined starting position, end position and marking type of the overall longitudinal deceleration marking.

[0119] Here, the road point cloud data includes a plurality of sampling points, each of which carries reflection intensity.

[0120] Specifically, based on the reflection intensity carried by each sampling point and the mapping relationship between the reflection intensity and the pixel value, the road point cloud data is subjected to overhead projection processing to obtain a grayscale image corresponding to the road point cloud data as the point cloud image.

[0121] The vehicle radar may be a laser radar or other radar capable of collecting point cloud data, which is not specifically limited here.

[0122] Here, the specific process of using the lane line detection model to detect the image area where the lane line is located from the point cloud image is similar to the process of using the lane line detection model in the aforementioned embodiment to detect the lane line from the road area image. The specific description can refer to the aforementioned embodiment and will not be repeated here.

[0123] In the embodiment of the present disclosure, when the degree of overlap between the second coordinate range and the first coordinate range is greater than a preset degree of overlap threshold, it can be determined that the starting point position, the end point position and the marking type of the overall longitudinal deceleration marking can be adopted; otherwise, it is determined that the starting point position, the end point position and the marking type of the overall longitudinal deceleration marking cannot be adopted.

[0124] Here, since the point cloud image is obtained by mapping the reflection intensity to the pixel value, this will cause a large deviation in the reflection intensity of two adjacent pixel points themselves, but a smaller deviation after conversion to pixel values, which will result in lower boundary clarity in the point cloud image. Therefore, when the overlap between the second coordinate range and the first coordinate range is greater than the preset overlap threshold, the starting point position, end point position and marking type of the overall longitudinal deceleration marking determined by the road area image taken by the vehicle camera shall be used as the basis, and the starting point position, end point position and marking type of the overall longitudinal deceleration marking shall be used for subsequent map construction.

[0125] The specific value of the preset overlap threshold value may be determined according to actual information data usage requirements and is not specifically limited here.

[0126] In this way, the road point cloud data can be used to verify whether the starting position, end position and marking type of the overall longitudinal speed reduction marking are accurate for map construction, so as to determine whether the determined starting position, end position and marking type of the overall longitudinal speed reduction marking should be used for map construction in the future. This helps to ensure the authenticity and accuracy of the road map when the longitudinal speed reduction markings in the road map are constructed through the starting position, end position and marking type of the overall longitudinal speed reduction marking.

[0127] The road marking detection method provided by the embodiment of the present disclosure, for a road area image containing lane lines, performs classification detection on each local road area image in the road area image, can perform refined road marking detection in a smaller range, and helps to improve the accuracy of road marking detection. Combined with the classification detection results of each local road area image, it can accurately identify the information of the overall longitudinal speed reduction marking of the entire road area, including the starting position and the end position of the overall longitudinal speed reduction marking, and the marking type of the overall longitudinal speed reduction marking, effectively reducing the difficulty in identifying and extracting the longitudinal speed reduction marking and the inaccurate identification position, so that the detected longitudinal speed reduction marking is highly consistent with the actual situation, which helps to ensure the authenticity and accuracy of the map when constructing the map later.

[0128] See also Figure 6 As shown, Figure 6 A flowchart of a road marking detection and simulation test method provided by an embodiment of the present disclosure, such as Figure 6 As shown in , it includes steps S601 to S605, wherein: S601: Acquire a road area image including lane lines.

[0129] S602: Classify and detect each local road area image of the road area image to determine the classification detection result of the local road area image, wherein the classification detection result indicates whether there is a longitudinal speed reduction marking in the local road area image and the type of the longitudinal speed reduction marking, and the marking type indicates whether there are longitudinal speed reduction markings on the left and right sides of the lane line respectively; the local road area images are arranged along the lane line direction in the road area image.

[0130] S603: Determine the starting point and the end point of the overall longitudinal speed reduction marking, and the marking type of the overall longitudinal speed reduction marking according to the classification detection results of each of the local road area images.

[0131] Among them, the description of step S601 to step S603 can refer to the description of step S101 to step S103, and can achieve the same technical effects and solve the same technical problems, which will not be repeated here.

[0132] S604: constructing a target map for simulation testing based on the starting position and the end position of the overall longitudinal speed reduction marking, the marking type of the overall longitudinal speed reduction marking, and the collected road image and vehicle positioning data of the target road.

[0133] In this step, the corresponding overall longitudinal speed reduction marking can be constructed based on the determined starting point and end point of the overall longitudinal speed reduction marking and the marking type of the overall longitudinal speed reduction marking, and then combined with the collected road image and vehicle positioning data of the target road, a target map for simulation testing can be constructed.

[0134] The vehicle positioning data refers to the position of the vehicle when the corresponding road image is captured, which can be specifically determined by data collected by the GPS and IMU set on the vehicle.

[0135] In practical applications, since the printing of longitudinal speed reduction markings is relatively regular and uniform, in order to improve the efficiency of map construction, a longitudinal speed reduction marking template can be used for map rendering.

[0136] S605: Based on the target map, a simulation test of autonomous driving is performed on the vehicle.

[0137] Here, a three-dimensional simulation scene corresponding to the real scene can be constructed based on the target map and other scene elements in the real scene, so that the vehicle can be simulated tested for autonomous driving in the three-dimensional simulation scene.

[0138] The above road marking detection and simulation test method, for a road area image containing lane lines, combines the classification detection results of each local road area image in the road area image, determines the starting position and end position of the overall longitudinal deceleration marking, and the marking type of the overall longitudinal deceleration marking, and then constructs a target map for simulation test based on the starting position and end position of the overall longitudinal deceleration marking, the marking type of the overall longitudinal deceleration marking, and the collected road image and vehicle positioning data of the target road, and performs a simulation test of automatic driving on the vehicle based on the target map. Through the above method, not only can the information of the overall longitudinal deceleration marking of the entire road area, including the starting position and end position of the overall longitudinal deceleration marking, and the marking type of the overall longitudinal deceleration marking, be accurately identified, and the situation of difficulty in identifying and extracting the longitudinal deceleration marking and inaccurate identification position is effectively reduced, so that the detected longitudinal deceleration marking is highly consistent with the actual situation, which helps to ensure the authenticity and accuracy of the target map, and the target map obtained by the above method is used to perform a simulation test of automatic driving on the vehicle, which helps to ensure the consistency of the longitudinal deceleration marking in the simulation scene with the real scene, thereby improving the simulation test effect and ensuring the reliability of the simulation test results.

[0139] Those skilled in the art will appreciate that, in the above method of specific implementation, the order in which the steps are written does not imply a strict execution order and does not constitute any limitation on the implementation process. The specific execution order of the steps should be determined by their functions and possible internal logic.

[0140] The disclosed embodiment also provides a vehicle, including a controller, the controller including: a memory for storing machine-readable instructions; and a processor for calling and executing the machine-readable instructions stored in the memory to implement the steps of the road marking detection method described in the above method embodiment. The vehicle may be a vehicle equipped with a variety of sensor components such as a camera, a radar, a global positioning system (GPS), an inertial measurement unit (IMU), etc.

[0141] Corresponding to the above-mentioned road marking detection method, the embodiment of the present disclosure further provides an electronic device 700, such as Figure 7 As shown in FIG. , it is a schematic diagram of the structure of an electronic device 700 provided in an embodiment of the present disclosure, including: Processor 710, memory 720, and bus 730. The memory 720 is used to store execution instructions, including internal memory 721 and external memory 722; the internal memory 721 is also called internal memory, which is used to temporarily store the operation data in the processor 710 and the data exchanged with the external memory 722 such as a hard disk. The processor 710 exchanges data with the external memory 722 through the internal memory 721.

[0142] In the embodiment of the present application, the memory 720 is specifically used to store the application code for executing the solution of the present application, and the execution is controlled by the processor 710. That is, when the electronic device 700 is running, the processor 710 communicates with the memory 720 through the bus 730, or the processor 710 communicates with the memory 720 in other ways, so that the processor 710 executes the application code stored in the memory 720, and then executes the steps of the road marking detection method described in any of the above embodiments.

[0143] Among them, the memory 720 can be, but is not limited to, random access memory (Random Access Memory, RAM), read only memory (Read Only Memory, ROM), programmable read-only memory (Programmable Read-Only Memory, PROM), erasable programmable read-only memory (Erasable Programmable Read-Only Memory, EPROM), electrically erasable read-only memory (Electric Erasable Programmable Read-Only Memory, EEPROM), etc.

[0144] The processor 710 may be an integrated circuit chip with signal processing capabilities. The above-mentioned processor may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The disclosed methods, steps and logic block diagrams in the embodiments of the present invention may be implemented or executed. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0145] It is to be understood that the structure illustrated in the embodiment of the present application does not constitute a specific limitation on the electronic device 700. In other embodiments of the present application, the electronic device 700 may include more or fewer components than shown in the figure, or combine some components, or split some components, or arrange the components differently. The components shown in the figure may be implemented in hardware, software, or a combination of software and hardware.

[0146] The embodiment of the present disclosure also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the road marking detection method described in the above method embodiment are executed. The storage medium can be a volatile or non-volatile computer-readable storage medium.

[0147] In addition, an embodiment of the present disclosure also provides a computer program product, which stores a computer program. When the computer program is executed by a processor, the steps of the road marking detection method provided in any of the above embodiments of the present disclosure are executed. For details, please refer to the above method embodiments, which will not be repeated here.

[0148] The computer program product may be implemented in hardware, software or a combination thereof. In an optional embodiment, the computer program product is embodied as a computer storage medium, which may be a volatile or non-volatile computer-readable storage medium. In another optional embodiment, the computer program product is embodied as a software product, such as a software development kit (SDK) and the like.

[0149] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, the specific working process of the above-described equipment and devices can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here. In the several embodiments provided in the present disclosure, it should be understood that the disclosed equipment, devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0150] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0151] In addition, each functional unit in each embodiment of the present disclosure may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0152] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present disclosure, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present disclosure. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0153] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present disclosure, which are used to illustrate the technical solutions of the present disclosure, rather than to limit them. The protection scope of the present disclosure is not limited thereto. Although the present disclosure is described in detail with reference to the above-described embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-described embodiments within the technical scope disclosed in the present disclosure, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present disclosure, and should be included in the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure shall be based on the protection scope of the claims.

Claims

1. A road marking detection method, characterized in that: The method comprises: Obtain a road area image including lane lines; Performing classification detection on each partial road area image of the road area image, determining a classification detection result of the partial road area image, the classification detection result indicating whether there is a longitudinal deceleration marking in the partial road area image, and the marking type of the longitudinal deceleration marking, the marking type indicating whether there are longitudinal deceleration markings on the left and right sides of the lane line respectively; the partial road area images are arranged along the lane line direction in the road area image; According to the classification detection results of each of the local road area images, the starting position and the end position of the overall longitudinal speed reduction marking line, and the marking type of the overall longitudinal speed reduction marking line are determined.

2. The method according to claim 1, characterized in that The step of acquiring a road area image including lane lines comprises: Acquire continuous multi-frame road area images taken when the vehicle is traveling on the target road; A lane line detection model is used to perform lane line detection on the multiple frames of road area images respectively to obtain each frame of road area image containing a lane line.

3. The method according to claim 1, before classifying and detecting each local road area image of the road area image, the method further comprises: The road area image is cut into a plurality of local road area images along the lane line direction according to a preset distance interval.

4. The method according to claim 3, characterized in that: The step of cutting the road area image into a plurality of local road area images along the lane line direction according to the preset distance intervals includes: Determine a plurality of target points on the lane line of the road area image at preset distance intervals, and crop a local road area image of a preset size with each target point as the center; The starting point and the end point of the overall longitudinal deceleration marking are the positions of the target points.

5. The method according to any one of claims 1 to 4, characterized in that The step of determining the starting point and the end point of the overall longitudinal deceleration marking and the marking type of the overall longitudinal deceleration marking according to the classification detection results of each of the local road area images comprises: Determine a road area image corresponding to the same lane line from each frame of road area image; According to the classification detection results of each of the local road area images in the road area image corresponding to the same lane line, the starting position and the end position of the overall longitudinal deceleration marking line and the marking type of the overall longitudinal deceleration marking line are determined.

6. The method according to claim 5, characterized in that The method of determining the starting point and the end point of the overall longitudinal deceleration marking line and the marking type of the overall longitudinal deceleration marking line according to the classification detection results of each of the local road area images in the road area image corresponding to the same lane line includes: Determine a plurality of continuous local road area images including longitudinal speed reduction markings according to the classification detection results of the local road area images in the road area image corresponding to the same lane line; The starting point and the end point of the overall longitudinal deceleration marking are determined according to the position information corresponding to the multiple continuous local road area images, and the marking type of the overall longitudinal deceleration marking is determined according to the classification detection results of the multiple continuous local road area images.

7. The method according to claim 6, characterized in that The classification detection result is also used to indicate the confidence level of each marking type corresponding to the longitudinal deceleration marking; The step of determining the marking type of the overall longitudinal speed reduction marking according to the classification detection results of the plurality of continuous local road area images comprises: Determining, based on the classification detection results of the plurality of consecutive partial road area images, whether, among the plurality of consecutive partial road area images, the classification detection results of the partial road area images exceeding a preset proportion indicate the same type of markings, and the confidence level of the indication is greater than a preset confidence threshold; If yes, it is determined that the marking type of the overall longitudinal deceleration marking is the same marking type.

8. The method according to claim 7, characterized in that The method further comprises: If there are no classification detection results of the partial road area images exceeding a preset ratio indicating the same marking type, and the indicated confidence is greater than a preset confidence threshold, determining the marking type of the overall longitudinal deceleration marking according to the relative positions of the lane lines corresponding to the plurality of continuous partial road area images; The lane line relative position indicates that the lane line belongs to one of a left lane line on the road, a right lane line on the road, and a middle lane line on the road.

9. The method according to claim 5, characterized in that The step of determining a road area image corresponding to the same lane line from each frame of road area image comprises: Determine, based on lane line position information in two adjacent frames of road area images, whether lane lines in two adjacent frames of road area images overlap in a lane line direction, and whether a distance in a direction perpendicular to the lane lines is less than a preset distance; If there is overlap in the lane line direction, and the distance in the direction perpendicular to the lane line is less than a preset distance, it is determined that the two adjacent frames of road area images correspond to the same lane line.

10. The method according to any one of claims 1 to 4, characterized in that After determining the starting position and the end position of the overall longitudinal deceleration marking, and the marking type of the overall longitudinal deceleration marking, the method further includes: Based on the starting position and the end position of the overall longitudinal deceleration marking, the marking type of the overall longitudinal deceleration marking, and the collected road image and vehicle positioning data of the target road, a target map for simulation testing is constructed; Based on the target map, a simulation test of autonomous driving is performed on the vehicle.

11. A vehicle, characterized in that: A controller is included, the controller comprising: a memory for storing machine-readable instructions; and A processor, configured to call and execute the machine-readable instructions stored in the memory to implement the steps of the road marking detection method according to any one of claims 1 to 10.

12. An electronic device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the machine-readable instructions are executed by the processor to perform the steps of the road marking detection method as described in any one of claims 1 to 10.

13. A computer program product, characterized in that The computer program product stores a computer program, and when the computer program is executed by a processor, the steps of the road marking detection method according to any one of claims 1 to 10 are executed.

Citation Information

Patent Citations

  • Traffic element identification method, multi-task network model and training method and device

    CN116012815A

  • Lane line detection method and device, vehicle and storage medium

    CN116092026A

  • Target detection method and device based on moving and parking integration, electronic equipment and medium

    CN116246235A

  • Method and device for detecting lane line and storage medium

    CN116311140A

  • Road traffic marking detection method based on vectorized aerial image

    CN116543321A