Ground marker detection method, device, equipment, and storage medium for high-precision maps

Abstract

The present application discloses a method, device and equipment for detecting ground signs for high-precision maps. The method first extracts frames from the real-time video on the autonomous driving vehicle to obtain the image to be processed, and then performs first image preprocessing, second image preprocessing and third image preprocessing on the image to be processed, thereby detecting the target ground sign. The technical solution in this application solves the problem that traditional recognition has large edge distortion and cannot accurately identify lane lines, and provides a reference for subsequent real-time decision-making and road planning of autonomous driving vehicles. At the same time, it can provide important auxiliary functions for scene perception, autonomous positioning, real-time decision-making and route planning for unmanned driving.

Description

Technical Field

[0001] The present application relates to the field of autonomous driving technology, and in particular to ground marker detection methods, devices, equipment, and storage media for high-precision maps. Background Art

[0002] With the development of autonomous driving technology, lane line detection and ground sign recognition have gradually become important components of environmental perception in the autonomous driving process. The navigation system must plan a reasonable route so that the autonomous driving vehicle can make correct driving decisions. High-precision maps are the foundation and an important component of the autonomous driving system, providing scene perception, autonomous positioning, real-time decision-making, and route planning, which plays an important auxiliary role in providing autonomous driving.

[0003] Currently, the primary method for identifying road signs during autonomous driving involves capturing real-time road images, performing perspective transformation on these images, and then identifying the road signs in these images. This recognition process relies primarily on a recognition model, which is continuously learned to improve recognition accuracy. Traditional recognition suffers from significant edge distortion and inaccuracies, resulting in low efficiency and high error rates for new or unique road signs. This negatively impacts subsequent real-time decision-making and route planning for autonomous driving. Summary of the Invention

[0004] In view of the above problems, the embodiments of the present application provide a method, apparatus, device, and storage medium for detecting ground signs in high-precision maps. This method enables the recognition and detection of special road signs, overcoming the problem that traditional recognition methods cannot accurately identify roadside signs due to large edge distortion, especially in the presence of light interference, resulting in a high error rate in the acquisition of ground signs. The technical solution provided by this application has greatly improved the recognition efficiency of road signs, especially in situations where light is limited, and has certain practicality.

[0005] The embodiments of this application adopt the following technical solutions:

[0006] In the first aspect, an embodiment of the present application provides a ground landmark detection method for high-precision maps, the method comprising: first, performing frame extraction processing on the real-time video on the autonomous driving vehicle to obtain a first picture to be processed, performing first image preprocessing, and obtaining a second picture including at least a lane line recognition result, for use in subsequent image processing processes; then, performing second image preprocessing based on the lane line recognition result in the second picture obtained after the first preprocessing, and obtaining a third picture including at least a lane line straight line clustering result, for use in subsequent image processing processes; finally, performing third image preprocessing on the lane line straight line clustering result in the third picture obtained after the second preprocessing, and detecting the target ground landmark to provide a reference for subsequent real-time decision-making and road planning of the autonomous driving vehicle.

[0007] Preferably, the lane line recognition results obtained after performing a first image preprocessing on the real-time video frame extractor include lane line recognition results based on threshold segmentation, where the threshold segmentation is based on threshold segmentation model parameters. The lane line clustering results include lane line detection results and lane line clustering results. The purpose is to classify and determine all detected lane lines for bus lane identification.

[0008] The method of obtaining a lane line recognition result based on the first preprocessed image and then performing a second image preprocessing to obtain a third image that at least includes a lane line clustering result includes: performing lane line line detection based on the lane line recognition result based on the threshold segmentation to obtain a lane line detection result, wherein the threshold segmentation is based on the threshold segmentation model parameters; and line detection is based on a line detection algorithm, wherein the input is the image after threshold segmentation and the output is a line array, wherein the line is represented in a two-point manner.

[0009] Based on the linear relationships in the lane line detection results, a lane line cluster set is established. Based on this lane line cluster set, multiple lane line clustering results are determined. Line clustering is based on a line clustering algorithm that clusters based on the slope and intercept of a line. The detected lines are clustered and output as a line cluster list, where the length of the list is the number of categories.

[0010] Preferably, the lane line clustering result is subjected to a third image preprocessing to detect target ground markers, specifically including:

[0011] The lane line clustering result obtained by preprocessing the third image is a line list whose length is the number of line classifications. The first target lane line and the second target lane line are determined from the number of line classifications.

[0012] A perspective transformation is performed based on multiple endpoints of the first target lane line and the second target lane line to obtain a fourth image including at least two lane lines, wherein the perspective transformation is performed by taking appropriate points on the obtained first and second straight lines, taking two points on each straight line, and performing a perspective transformation on the four points to obtain an image after perspective.

[0013] Preferably, determining the first target lane line and the second target lane line based on the lane line clustering results in the third image includes: determining lane line groupings based on polar angle sorting based on the lane line clustering results in the third image, calculating the average slope k for each line group in the line clustering, and sorting them by 1 / k to obtain the two rightmost groups, namely, bus lanes. Obtaining the first target lane line and the second target lane line based on the lane line groupings, i.e., obtaining bus lane lines, is used for subsequent ground sign recognition processing.

[0014] Preferably, the detection of the target ground marking includes: according to the fourth picture including at least two lane lines, after a preset image change, obtaining the target area in the two lane lines, wherein the preset image transformation refers to related image processing performed to enhance the recognition effect of the landmark marking, including but not limited to image grayscale processing, image histogram equalization processing, image corrosion processing, image expansion processing, image edge detection processing, etc.

[0015] Identify the text in the target area and determine the type of the ground sign. The text in the target area includes but is not limited to the word "bus lane".

[0016] Preferably, the preset image transformation includes at least one or more of image grayscale, image histogram equalization, image erosion, image dilation, and image edge detection processing, and the target area in the two lane lines includes a binary image of the target area in the lane lines.

[0017] Preferably, the first image to be processed is obtained by extracting frames from a real-time road video.

[0018] According to the second aspect of the present application, an embodiment of the present application also provides a ground marker detection device for high-precision maps, the device comprising: a first image preprocessing unit, for performing a first image preprocessing on a first image to be processed to obtain a second image including at least a lane line recognition result; a second image preprocessing unit, for performing a second image preprocessing on the lane line recognition result in the second image to obtain a third image including at least a lane line straight line clustering result; a third image preprocessing unit, for performing a third image preprocessing on the lane line straight line clustering result in the third image to detect the target ground marker.

[0019] According to a third aspect of the present application, an embodiment of the present application further provides a ground marker detection device for a high-precision map, the ground marker detection device for a high-precision map comprising: a memory, a processor, and a lane ground marker detection program for a high-precision map stored and running on the processor, wherein the lane ground marker detection program for a high-precision map, when executed by the processor, implements any one of the following methods:

[0020] First, a picture is directly extracted from the vehicle video. The lane line recognition model based on threshold segmentation is input with the model parameters consisting of the extracted picture and the weight threshold. The output is the picture after threshold segmentation.

[0021] Then, the image segmented by the threshold is processed for line detection. The image segmented by the threshold is input and a line array is output. Then, a clustering algorithm is used to cluster the lines according to their slope and intercept, and a line cluster list is output. The length of the list is the number of categories, and the line list obtained by the line clustering is sorted by polar angle. Finally, a perspective transformation is performed on the polar angle-sorted image to obtain the text in the target area.

[0022] Finally, the text in the obtained target area is processed at least one or more of image grayscale, image histogram equalization, image erosion, image expansion, and image edge detection to obtain a ground mark.

[0023] According to another aspect of the present application, an embodiment of the present application further provides a computer-readable storage medium, on which a lane ground marking detection program for a high-precision map is stored. When the lane ground marking detection program for a high-precision map is executed by a processor, any of the following methods is implemented:

[0024] First, a picture is directly extracted from the vehicle video. The lane line recognition model based on threshold segmentation is input with the model parameters consisting of the extracted picture and the weight threshold. The output is the picture after threshold segmentation.

[0025] Then, the image segmented by the threshold is processed for line detection. The image segmented by the threshold is input and a line array is output. Then, a clustering algorithm is used to cluster the lines according to their slope and intercept, and a line cluster list is output. The length of the list is the number of categories. The line list obtained by the line clustering is sorted by polar angle sorting. Polar angle sorting is sorting by the inverse of the slope of the line. Finally, a perspective transformation is performed on the polar angle sorted image to obtain the text in the target area.

[0026] Finally, the text in the obtained target area is processed at least one or more of image grayscale, image histogram equalization, image erosion, image expansion, and image edge detection to obtain a ground mark.

[0027] At least one of the above technical solutions adopted in the embodiments of the present application can achieve the following beneficial effects:

[0028] By extracting video frames in real time and performing multiple image preprocessing on the extracted images, the text in the target area is obtained. Then, the text in the current area is subjected to preset image processing detection and recognition to obtain ground signs. Because this application uses multiple image processing methods, ground signs can be accurately obtained, and even if the quality of the obtained images is poor due to lighting conditions, ground signs can still be clearly identified. This improves the recognition efficiency of ground signs and enhances the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The drawings described herein are used to provide further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute improper limitations on the present application.

[0030] Figure 1 This is a flow chart of the ground marker detection method for high-precision maps in this application example;

[0031] Figure 2 This is a schematic diagram of the first image preprocessing process of the ground marker detection method for high-precision maps in an example of this application;

[0032] Figure 3 This is a schematic diagram of the second image preprocessing process of the ground marker detection method for high-precision maps in the example of this application;

[0033] Figure 4 This is a schematic diagram of the third image preprocessing process of the ground marker detection method for high-precision maps in the example of this application;

[0034] Figure 5 This is a schematic diagram of the structure of a ground marker detection device for high-precision maps in an example of this application;

[0035] Figure 6 This is a schematic diagram of perspective transformation of the ground marker detection method for high-precision maps in this application example;

[0036] Figure 7 This is a schematic diagram of the binary result after the preset image transformation of the ground marker detection method for high-precision maps in the example of this application;

[0037] Figure 8 This is a schematic diagram of the structure of the ground marker detection equipment used for high-precision maps in this application example. DETAILED DESCRIPTION

[0038] To make the purpose, technical solutions, and advantages of this application more clear, the technical solutions of this application will be clearly and completely described below in conjunction with the specific embodiments of this application and the corresponding drawings. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0039] The following describes in detail the technical solutions provided by various embodiments of the present application in conjunction with the accompanying drawings.

[0040] The embodiment of the present application provides a method for detecting ground signs, which is applied to image processing, machine vision, high-precision maps, and autonomous driving, especially the detection of ground signs in bus lanes. Figure 1 As shown, a flow chart of an embodiment of the present application is provided, and the method at least includes the following steps S101 to S103.

[0041] Step S101 : performing first image preprocessing on a first image to be processed to obtain a second image including at least a lane line recognition result.

[0042] In some embodiments of the present application, an autonomous vehicle captures real-time ground video while driving. The ground sign detection equipment in the present application extracts frames from the real-time video according to a predetermined program method to obtain a first image to be processed. Since the image captured by the real-time video frame extraction is obtained by the video capture device installed in the autonomous vehicle, there is geometric distortion caused by the shooting posture. Therefore, the first image needs to be corrected. This is mainly geometric correction. First, a mathematical model for geometric correction is established; second, the model parameters are determined using known conditions; and finally, the image is geometrically corrected according to the model. Image correction is also called image dedistortion. The overall idea is to convert each pixel point on the undistorted surface to a normal plane; then distort and project it onto the original image, and obtain the pixel value of the corresponding position in the original image as the pixel value of the target image point. The obtained image after image correction is processed for lane line recognition to obtain lane line information such as the number, color, and whether the lane line is curved in the current frame for subsequent processing.

[0043] Step S102: performing a second image preprocessing on the lane line recognition result in the second image to obtain a third image that at least includes a lane line clustering result.

[0044] The image after lane line recognition undergoes further processing. First, line detection is performed. This involves using, but is not limited to, the Hough line detection algorithm, transforming the line detection problem in image space into a point detection problem in parameter space. Simple cumulative statistics are performed in the parameter space to complete the detection task. Following line detection, line clustering is performed, resulting in an image containing the lane line clustering results for subsequent processing.

[0045] In step S103, a third image preprocessing is performed on the lane line clustering result in the third image to detect the target ground marker.

[0046] The image of the line clustering result is subjected to, including but not limited to, perspective transformation to obtain the target area in the target lane line. The target area is then subjected to multiple image preprocessing processes including but not limited to image grayscale conversion and image histogram equalization to detect the target ground mark.

[0047] In summary, the traditional method of detecting ground signs is to perform a simple perspective transformation on the image obtained by the video acquisition device and then identify the ground signs. The method is relatively simple and crude, and the error rate of ground sign recognition is high. The correct identification of ground signs is related to the real-time decision-making and road planning of subsequent intelligent vehicles. The technical solution of the present application adopts a variety of image processing methods, which can accurately obtain ground signs, and even in the case of poor light and differences in shooting posture resulting in poor image quality, the ground signs can be clearly identified, thereby improving the recognition efficiency of ground signs and enhancing the user experience.

[0048] In some embodiments of the present application, the first image preprocessing includes: step S201, i.e., image correction, and step S202, i.e., lane line recognition processing. Figure 2 As shown in the figure, because the images captured by real-time video frame extraction are from a video capture device installed in an autonomous vehicle, they are subject to geometric distortion caused by the shooting posture. Therefore, the first image needs to be corrected. This primarily involves geometric correction. First, a mathematical model for geometric correction is established; second, the model parameters are determined using known conditions; and finally, the image is geometrically corrected based on the model. The corrected image is then processed for lane line recognition, obtaining lane line information such as the number, color, and whether the lane line is curved in the current frame for subsequent processing.

[0049] In some embodiments of the present application, the second image preprocessing includes: step S301, i.e., image line detection processing, and step S302, i.e., image line clustering processing. Figure 3 As shown, the input of the image line detection processing is the image after lane line recognition, and the output is a line array. The lines in the array can be represented in multiple ways. In this example, the line is represented in two-point form.

[0050] This algorithm clusters lines based on their slope and intercept. It takes as input an array of detected lines and outputs a list of line categories, with the length of the list being the number of categories. This image processing approach ensures accurate targeting of the target area and subsequent ground landmark recognition.

[0051] In some embodiments of the present application, Figure 4 As shown in the figure, the third image preprocessing includes polar angle sorting, image perspective transformation, and preset image transformation. Specifically, polar angle sorting calculates the average slope k for each line group in the line cluster list and sorts them by 1 / k to obtain the rightmost line, i.e., the bus lane line; perspective transformation selects appropriate points on the target line, two points for each line, and then performs perspective transformation on the target area consisting of the four points obtained on the two lines to obtain a fourth image including at least two lane lines, as shown in the figure. Figure 6 Finally, the fourth image is subjected to a preset image transformation process to obtain the target area in the two lane lines, and the text in the target area is recognized to determine the ground marking. The preset image transformation process includes at least one or more of image grayscale, image histogram equalization, image erosion, image dilation, and image edge detection, thereby obtaining a binary image of the lane line target area, as shown in FIG. Figure 7 As shown in the figure, the resulting binary text image is then processed for text recognition to obtain ground signs. Through a series of image processing steps, ground signs can be accurately captured, even in poor lighting conditions resulting in poor image quality. This improves ground sign recognition efficiency and enhances the user experience.

[0052] Figure 5 A ground marker detection device 500 for high-precision maps according to an embodiment of the present application is shown. The device 500 includes at least: a first image preprocessing unit 501, a second image preprocessing unit 502, and a third image preprocessing unit 503.

[0053] The first image preprocessing unit 501 is configured to perform first image preprocessing on an image of the road surface captured by a video capture device installed in the autonomous vehicle using real-time video frame extraction, thereby obtaining a second image that includes at least lane line recognition results. Because the images captured by the real-time video frame extraction are geometrically distorted by the camera's shooting posture, the first image requires image correction. This primarily involves geometric correction. First, a mathematical model for geometric correction is established; second, model parameters are determined using known conditions; and finally, the image is geometrically corrected based on the model. Image correction, or image dedistortion, involves transforming each pixel on the undistorted image to a normal plane. The resulting pixel is then distorted and projected onto the original image, obtaining the pixel value at the corresponding position in the original image as the pixel value at that point in the target image. Lane line recognition processing is then performed on the resulting image after image correction to obtain lane line information, such as the number, color, and whether the lane lines are curved, for subsequent processing.

[0054] The second image preprocessing unit 502 is configured to perform a second image preprocessing on the lane recognition results in the second image to obtain a third image comprising at least lane line clustering results. Specifically, the lane line recognition image is further processed by first performing image line detection, including but not limited to using the Hough line detection algorithm to convert the line detection problem in image space into a point detection problem in parameter space, completing the detection task by performing simple cumulative statistics in the parameter space. After image line detection, line clustering is performed to obtain an image-processed image comprising the lane line clustering results for subsequent processing.

[0055] The third image preprocessing unit 503 is configured to perform third image preprocessing on the lane line clustering result in the third image to detect the target ground marker. Specifically, the image of the line clustering result is subjected to a third image preprocessing process, including but not limited to a perspective transformation, to obtain a target region within the target lane line. The target region is then subjected to multiple image preprocessing processes, including but not limited to image grayscale conversion and image histogram equalization, to detect the target ground marker.

[0056] Figure 8 FIG1 shows a schematic diagram of a ground marker detection device for high-precision maps according to an embodiment of the present application. Figure 8 As shown, the ground marker detection device 800 includes: a processor 801; and a memory 802 arranged to store computer executable instructions. When the executable instructions are executed by the processor 801, the above method is implemented.

[0057] In summary, the technical solution of the present application obtains a pre-processed image by extracting frames from a real-time video; a lane line recognition model based on threshold segmentation is performed on the pre-processed image, the model inputs the pre-processed image and the model parameters composed of the weight threshold, and the output is the image after threshold segmentation, and then a straight line detection algorithm is performed, the input is the image after threshold segmentation, and the output is a straight line array, and then the clustering algorithm clustering by the slope and intercept of the straight line outputs a list, the length of the list is the number of categories, and then the obtained cluster list is sorted by polar angle and perspective transformation and image preset image transformation, so as to obtain ground signs. The examples of the present application improve the recognition accuracy and recognition efficiency of ground signs, especially in the case of interference from shooting posture and light interference, the recognition efficiency of ground signs is significantly improved, and the problem that the traditional recognition method cannot accurately recognize lane lines due to large edge distortion is solved, which provides a reliable guarantee for subsequent real-time decision-making and road selection of autonomous driving, improves safety, and optimizes user experience.

[0058] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0059] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0060] These computer program instructions for detecting ground markers for high-precision maps may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0061] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0062] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0063] Memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

[0064] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can be implemented using any method or technology to store information. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change RAM (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media such as modulated data signals and carrier waves.

[0065] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not preclude the presence of additional identical elements in the process, method, commodity, or apparatus that includes the element.

[0066] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0067] The foregoing is merely an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should all be included within the scope of the claims of the present application.

[0068] The various component embodiments of the present application can be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. Those skilled in the art will appreciate that a microprocessor or digital signal processor (DSP) can be used in practice to implement the embodiments.

[0069] According to some or all functions of some or all components of the ground identification detection device for high-precision maps according to an embodiment of the present application. The present application can also be implemented as a device or apparatus program (for example, a computer program and a computer program product) for performing part or all of the methods described herein. Such a program implementing the present application can be stored on a computer-readable medium, or can be in the form of one or more signals. Such a signal can be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.

[0070] For example, Figure 8The present invention shows an intention of a ground marker detection device for high-precision maps according to an embodiment of the present application. The ground marker detection device 800 includes a processor 801 and a memory 802 arranged to store computer-executable instructions (computer-readable program code). The memory 802 can be an electronic memory such as a flash memory, an EEPROM (electrically erasable programmable read-only memory), an EPROM, a hard disk, or a ROM. The memory 802 has a storage space 8021 for storing a computer-readable program code 80211 for executing any method step in the above method. For example, the storage space 8021 for storing computer-readable program code may include individual computer-readable program codes 80211 for respectively implementing various steps in the above method. The computer-readable program code 8021 can be read from or written to one or more computer program products. These computer program products include program code carriers such as a hard disk, a compact disk (CD), a memory card, or a floppy disk. The computer-readable program code 60211 can be compressed in an appropriate form.

[0071] It should be noted that the above embodiments illustrate rather than limit the present application, and that a person skilled in the art may devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference symbols placed between brackets should not be construed as limiting the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present application may be implemented by means of hardware comprising several different elements and by means of appropriately programmed computers. In a unit claim enumerating several means, several of these means may be embodied by the same item of hardware. The use of the words first, second, and third etc. does not indicate any order. These words may be interpreted as names.

Claims

1. A method for detecting ground markers in high-precision maps, characterized in that: The method comprises: Performing a first image preprocessing on the first image to be processed to obtain a second image including at least lane line recognition results; wherein the first image preprocessing includes image correction, and the image correction includes establishing a mathematical model for geometric correction; determining model parameters using known conditions; performing geometric correction on the image according to the model; and performing lane line recognition processing on the obtained image after image correction to obtain lane line information including the number, color, and whether the lane lines are curved in the current frame; performing a second image preprocessing on the lane recognition result in the second image to obtain a third image including at least a lane line clustering result; wherein the second image preprocessing includes image line detection, converting the line detection problem in image space to the point detection problem in parameter space, and completing the detection task by performing simple cumulative statistics in the parameter space; performing a third image preprocessing on the lane line clustering result in the third image to detect a target ground marker; The performing third image preprocessing on the lane line clustering result in the third image to detect target ground markers includes: Determine a first target lane line and a second target lane line according to the lane line clustering result in the third image; A perspective transformation is performed on multiple endpoints of the first target lane line straight line and the second target lane line straight line to obtain a fourth image including at least two lane lines; a preset image transformation is performed on the fourth image to obtain a target area within the two lane lines, and text in the target area is recognized to determine the ground marking.

2. The method according to claim 1, wherein: The lane line recognition result includes: a lane line recognition result based on threshold segmentation; the lane line straight line clustering result includes: a lane line straight line detection result and a lane line straight line clustering result; The performing second image preprocessing according to the lane line recognition result in the second image to obtain a third image including at least a lane line clustering result includes: Performing lane line straight line detection based on the lane line recognition result based on the threshold segmentation to obtain a lane line straight line detection result; Establishing a lane line cluster set based on the linear relationship in the lane line detection results; Determine a plurality of lane line clustering results according to the lane line clustering set.

3. The method according to claim 1, wherein: Determining a first target lane line and a second target lane line according to the lane line clustering result in the third image includes: Determining the lane line groups according to the lane line clustering result in the third image and sorting by polar angle; The first target lane line straight line and the second target lane line straight line are obtained according to the grouping of the lane lines.

4. The method according to claim 1, wherein: The preset image transformation includes at least one or more of image grayscale, image histogram equalization, image erosion, image dilation, and image edge detection processing, and the target area in the two lane lines includes a binary image of the target area in the lane lines.

5. The method according to claim 1, wherein: The first image to be processed is obtained by extracting frames from a real-time road video.

6. A ground marker detection device for high-precision maps, characterized in that: The device comprises: a first image preprocessing unit configured to perform first image preprocessing on the first image to be processed to obtain a second image including at least lane line recognition results; wherein the first image preprocessing includes image correction, which includes establishing a mathematical model for geometric correction; determining model parameters using known conditions; performing geometric correction on the image according to the model; and performing lane line recognition processing on the obtained image after image correction to obtain lane line information indicating the number, color, and whether the lane lines are curved in the current frame; a second image preprocessing unit, configured to perform second image preprocessing on the lane recognition result in the second image to obtain a third image including at least a lane line clustering result; wherein the second image preprocessing includes image line detection, converting the line detection problem in image space to the point detection problem in parameter space, and completing the detection task by performing simple cumulative statistics in the parameter space; a third image preprocessing unit, configured to perform third image preprocessing on the lane line clustering result in the third image to detect a target ground marker; The performing third image preprocessing on the lane line clustering result in the third image to detect target ground markers includes: Determine a first target lane line and a second target lane line according to the lane line clustering result in the third image; A perspective transformation is performed on multiple endpoints of the first target lane line straight line and the second target lane line straight line to obtain a fourth image including at least two lane lines; a preset image transformation is performed on the fourth image to obtain a target area within the two lane lines, and text in the target area is recognized to determine the ground marking.

7. A ground marker detection device for high-precision maps, characterized in that: The ground marking detection device for high-precision maps includes: a memory, a processor, and a lane ground marking detection program for high-precision maps stored and running on the processor. When the lane ground marking detection program for high-precision maps is executed by the processor, it implements the method described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a lane ground marking detection program for a high-precision map, and when the lane ground marking detection program for a high-precision map is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

Images