Lane line virtual and real classification method and system based on geometric features

Abstract

This invention discloses a method and system for classifying lane lines as solid or void based on geometric features. It collects lane line data and unifies the lane line representation as a broken line formed by connecting scattered points extending along the lane line direction. Then, based on the angle formed between the lines connecting lane lines in the preceding and following segments and the lane lines in the following segment, it finds the preceding and following adjacent lines of each lane line. Finally, it determines the solid or void classification of the lane line based on the ratio of the lane line's length to its distance from its adjacent lines. This effectively solves the problem of classifying the solid or void attributes of lane lines in crowdsourcing scenarios with limited precision.

Description

Technical Field

[0001] This invention relates to high-precision map production technology, and in particular to a method and system for classifying lane lines as solid or void based on geometric features. Background Technology

[0002] In the field of autonomous driving, navigation technology based on electronic maps relies heavily on the mapping results of various map elements around the vehicle. Lane lines are one of the most common and crucial elements for navigation. However, in crowdsourced scenarios, the lane line information collected may lack or have inaccurate virtual / real attributes, leading to a decrease in map quality and further affecting the implementation of navigation functions such as lane changing and turning.

[0003] Lane detection is crucial for autonomous driving. Accurate lane line positioning can prevent rear-end collisions and other traffic accidents, and also aids in vehicle positioning. Currently, due to advancements in deep learning research, there are many research results on lane detection in the field of autonomous driving, but research on the classification of lane line attributes (real and virtual) is relatively limited. Recent research on lane detection and classification, such as Fabip Pizzati, Marco Allodi, et al.'s "Lane Detection and Classification using Cascaded CNNs" published in July 2019, focuses on training a classifier using cascaded convolutional neural networks and labeled datasets, achieving an accuracy of 95.24% at a frame rate of 59.93. However, the TuSimple dataset they used has limited capacity, and the image-based processing method may not be suitable for crowdsourced mapping. Summary of the Invention

[0004] The purpose of this invention is to overcome the above-mentioned technical deficiencies and propose a lane line virtual-real classification method and system based on geometric features, so as to solve the problem that virtual-real lane lines cannot be distinguished in existing crowdsourcing scenarios.

[0005] To achieve the above-mentioned technical objectives, the first aspect of the technical solution of the present invention provides a method for classifying lane lines as solid or void based on geometric features, which includes the following steps:

[0006] Collect lane line data and standardize the lane line representation to a broken line formed by connecting scattered points extending along the lane line direction;

[0007] Based on the angle formed between the lines connecting the lane lines in the front and rear segments and the lane lines in the rear segment, find the adjacent lines of each lane line.

[0008] The classification of lane lines as solid or dashed is determined by the ratio of the length of the lane line itself to its distance from the adjacent line.

[0009] A second aspect of the present invention provides a lane line real / virtual classification system based on geometric features, which includes the following functional modules:

[0010] The format unification module is used to collect lane line data and unify the lane line representation to a broken line form formed by connecting scattered points extending along the lane line direction.

[0011] The adjacent line lookup module is used to find the adjacent lines of each lane line based on the angle formed by the line connecting the lane lines located in the front and rear segments and the lane lines located in the rear segment.

[0012] The virtual / real lane marking module is used to determine the virtual / real lane marking classification based on the ratio of the length of the lane marking itself to its distance from the adjacent lane marking.

[0013] A third aspect of the present invention provides a server, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described method for classifying lane lines based on geometric features.

[0014] A fourth aspect of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for classifying lane lines based on geometric features.

[0015] Compared with existing technologies, the lane line virtual-real classification method and system based on geometric features described in this invention collects lane line data and unifies the lane line representation as a broken line formed by connecting scattered points extending along the lane line direction; then, based on the angle formed by the lines connecting the lane lines in the preceding and following segments and the lane lines in the following segment, the preceding and following adjacent lines of each lane line are found; finally, the virtual-real classification of the lane line is determined according to the ratio of the length of the lane line body to its distance from its adjacent lines; thus, it can effectively solve the problem of lane line virtual-real attribute classification in crowdsourcing scenarios with limited accuracy. Attached Figure Description

[0016] Figure 1 This is a flowchart of the lane line virtual-real classification method based on geometric features according to an embodiment of the present invention;

[0017] Figure 2 yes Figure 1 The flowchart of step S2;

[0018] Figure 3 This is a schematic diagram showing that the lane lines are rectangular in shape.

[0019] Figure 4 This is a schematic diagram of the adjacent lines connecting the dashed lane markings;

[0020] Figure 5 This is a block diagram of the lane line virtual-real classification system based on geometric features as described in an embodiment of the present invention. Detailed Implementation

[0021] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0022] like Figure 1 As shown, embodiments of the present invention provide a lane line solid / dummy classification method based on geometric features, which includes the following steps:

[0023] S1. Collect lane line data and standardize the lane line representation to a broken line formed by connecting scattered points extending along the lane line direction.

[0024] When the lane line data consists of sampling points distributed along the lane line direction, adjacent sampling points with a spacing less than a set threshold are connected to form multiple sequentially connected polylines. The set threshold spacing is an empirical distance value between adjacent dashed lines or dashed-solid lines. If the spacing between adjacent sampling points is less than the set threshold spacing, they are determined to be on the same lane line; otherwise, they are not on the same lane line.

[0025] When the lane line data is in the form of polygons, specifically taking a quadrilateral whose outline is close to a rectangle as an example, compare the lengths of the two adjacent sides of the first index point, take the midpoint of the shorter side and the midpoint of another shorter side set opposite to the shorter side, and take the equidistant points of the connecting line to split the connecting line into multiple sequentially connected polylines.

[0026] For example: Figure 3 As shown, in the case of a quadrilateral, the lane line outline can be approximated as a rectangle. The quadrilateral point sequence of the vehicle outline can be either clockwise or counterclockwise. It is necessary to determine the relationship between the distances of points 1 and 2 and the distances of points 1 and 4 to confirm the point sequence. If the distance between points 1 and 2 is greater than the distance between points 1 and 4, it means that the distance between points 1 and 2 corresponds to the longer side of the rectangle, and the point sequence is clockwise. Figure 3 As shown in (a), otherwise it means that the distances between points 1 and 4 correspond to the longer side of the rectangle, and the point order is counterclockwise, as shown in (a). Figure 3 As shown in (b). Take the midpoint A of the short side of the rectangle where point 1 is located, the midpoint E of the other short side of the rectangle, the center point C of the rectangle, the midpoint B of AC, and the midpoint D of CE as the broken line of the lane line.

[0027] S2. Based on the angle formed by the line connecting the lane lines located in the front and rear segments and the lane lines located in the rear segment, find the adjacent lines of each lane line.

[0028] Specifically, the angle formed between the connection between the lane lines located in the front and rear segments and the lane lines located in the rear segment is the angle formed between the line connecting the end point of the preceding lane line and the starting point of the preceding lane line and the broken line where the starting point of the preceding lane line is located.

[0029] Specifically, such as Figure 2 As shown, step S2 includes the following sub-steps:

[0030] S21. Using the start and end points of the current lane line as the center, search for adjacent lane lines within a preset radius threshold range; the adjacent lane lines within the preset radius threshold range include all adjacent lane lines within the preset radius threshold range and those intersecting with the preset radius threshold range.

[0031] S22. Connect the end points of adjacent lane lines within a preset radius threshold range centered on the starting point to the starting point of the current lane line to form the first connection point.

[0032] S23. Connect the starting points of adjacent lane lines within a preset radius threshold range centered on the end point to the end point of the current lane line to form a second connection point.

[0033] S24. Calculate the cosine of the first angle between each first point connection and the broken line where the current lane line starts, and the cosine of the second angle between each second point connection and the broken line where the current lane line ends. Based on the comparison results of the first and second angle cosine values ​​with the preset adjacent connection cosine thresholds, respectively, select and obtain the preparatory front adjacent connection set and the rear adjacent connection set.

[0034] S25. Select the pre-adjacent connection with the shortest first point connection between its end point and the start point of the current lane line from the pre-prepared pre-adjacent connection set as the pre-adjacent connection of the current lane line. Select the pre-adjacent connection with the shortest first point connection between its start point and the end point of the current lane line from the pre-prepared post-adjacent connection set as the post-adjacent connection of the current lane line.

[0035] by Figure 4 To illustrate the criteria for determining adjacent lines, suppose we need to find the adjacent lines before and after A2B2C2D2E2. Let the length of the broken line be L. Draw a circle with A2 and E2 as the center and 10L as the radius. All lane lines inside the circle or intersecting the circle are recorded as possible adjacent lines.

[0036] Within the circle defined by center A2, connect the end point E1 of each lane line to its starting point A2. Find... and The cosine of the included angle should fall above 0.99, i.e., within 8° or 0.14 rad. This will allow you to find the lane lines on the forward extension of A2B2C2D2E2.

[0037] Then, among these lane lines, find the lane line that is closest to A2B2C2D2E2, which is the previous neighbor of A2B2C2D2E2. There are several ways to calculate the distance here. The simplest way is to find the lane line that is closest to the starting point A2 of the current lane line among all the lane lines that meet the included angle condition found in the previous step. This is the previous neighbor of the current lane line.

[0038] Using the same method, the next adjacent connection of A2B2C2D2E2 can be found within the circle centered at point E2.

[0039] S3. Based on the ratio of the length of the lane line itself to its distance from the adjacent lane line, determine the classification of lane lines as solid or dashed. This includes:

[0040] The ratio of the length of the first connection point to the current lane line is calculated as the preceding neighbor ratio, and the ratio of the length of the second connection point to the current lane line is calculated as the following neighbor ratio. Both the preceding and following neighbor ratios are compared with a preset dashed line neighbor ratio threshold range. Based on the comparison results, the current lane line is determined to be either dashed or solid. Specifically, if either the current or following neighbor ratio falls within the preset dashed line neighbor ratio threshold range, the current lane line is considered dashed; otherwise, it is considered solid.

[0041] The difference between dashed and solid lines is that the ratio of the distance between a dashed line and its adjacent line to its own length is larger. In road safety standards, this ratio generally falls between 0.5 and 2. For solid lines, this ratio can be very small, typically below 0.1. Therefore, this ratio is calculated based on the preceding and following adjacent lines of each lane. Taking the diagram as an example, for A2B2C2D2E2, this ratio is... Let it be denoted as d1 / L, and the ratio to the next neighbor can also be obtained. Let this be denoted as d2 / L. Further considering measurement errors and combining actual test data, an empirical range can be given between (0.32, 5.91). If either d1 / L or d2 / L falls within this range, it can be considered a dashed line; otherwise, it is considered a solid line.

[0042] It's important to note that not all lane markings have readily available front and rear adjacent lines. Excluding the possibility of missed data collection, lane markings located at the ends of long dashed lines or isolated long solid lines may only have a front or rear adjacent line, or no adjacent lines at all. When adjacent lines cannot be found, the lane marking length and λL must be considered. m The size relationship is used to determine, where L mThis refers to the mode of lane lengths calculated in step 2 above. λ is a correction factor for this mode of length based on the specific measurement error; an empirical value of 1.28 is acceptable. Lanes with no neighbors and a length greater than λL are considered lanes with this mode. m If a lane line is solid, has no adjacent lane lines, and its length is less than or equal to λL, then it can be considered a solid line. m If it is a dashed line, it can be considered a dashed line; if there is only one adjacent line, it can be determined based on the d1 / L or d2 / L and length of this neighbor.

[0043] This invention collects lane line data and standardizes the representation of lane lines as a broken line formed by connecting scattered points extending along the lane line direction. Then, based on the angle formed between the lines connecting the lane lines in the preceding and following segments and the lane lines in the following segment, it finds the preceding and following adjacent lines of each lane line. Finally, based on the ratio of the length of the lane line itself to its distance from its adjacent line, it determines the virtual / real classification of the lane lines. This invention can effectively solve the problem of virtual / real attribute classification of lane lines in crowdsourcing scenarios with limited precision.

[0044] like Figure 5 As shown in the figure, this invention also discloses a lane line real / virtual classification system based on geometric features, which includes the following functional modules:

[0045] The format unification module 10 is used to collect lane line data and unify the lane line representation as a broken line formed by connecting scattered points extending along the lane line direction.

[0046] The adjacent line lookup module 20 is used to find the adjacent lines of each lane line based on the angle formed between the line connecting the lane lines located in the front and rear segments and the lane lines located in the rear segment.

[0047] The virtual / real lane line determination module 30 is used to determine the virtual / real lane line classification based on the ratio of the length of the lane line itself to its distance from the adjacent line.

[0048] The execution method of the lane line solid-virtual classification system based on geometric features in this embodiment is basically the same as the lane line solid-virtual classification method based on geometric features described above, so it will not be described in detail.

[0049] In this embodiment, the server is a device that provides computing services, typically referring to a computer with high computing power that is provided to multiple consumers via a network. The server in this embodiment includes a memory, a processor, and a system bus. The memory includes executable programs stored thereon. Those skilled in the art will understand that the terminal device structure of this embodiment does not constitute a limitation on the terminal device and may include more or fewer components than illustrated, or combine certain components, or have different component arrangements.

[0050] The memory can be used to store software programs and modules. The processor executes various terminal functions and data processing by running the software programs and modules stored in the memory. The memory can mainly include a program storage area and a data storage area. The program storage area can store the operating system, application programs required for at least one function (such as sound playback, image playback, etc.), etc.; the data storage area can store data created according to the use of the terminal (such as audio data, phonebook, etc.). In addition, the memory can include high-speed random access memory, and can also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0051] An executable program containing a geometric feature-based lane line classification method is stored in memory. This executable program can be divided into one or more modules / units, which are stored in the memory and executed by a processor to complete the information acquisition and implementation process. Each module / unit can be a series of computer program instruction segments capable of performing a specific function, describing the execution process of the computer program on the server. For example, the computer program can be divided into a form unification module 10, an adjacent line search module 20, and a lane line classification module 30.

[0052] The processor is the control center of the server, connecting various parts of the terminal device through various interfaces and lines. It performs various terminal functions and processes data by running or executing software programs and / or modules stored in memory, and by calling data stored in memory, thereby providing overall monitoring of the terminal. Optionally, the processor may include one or more processing units; preferably, the processor may integrate an application processor and a modem processor, wherein the application processor mainly handles the operating system, applications, etc., and the modem processor mainly handles wireless communication. It is understood that the modem processor may also not be integrated into the processor.

[0053] The system bus connects various functional components within a computer, transmitting data, address, and control information. Types of system buses include PCI, ISA, and VESA. Processor instructions are transmitted to memory via the bus, and memory sends data back to the processor. The system bus handles the data and instruction exchange between the processor and memory. Of course, the system bus can also connect to other devices, such as network interfaces and display devices.

[0054] The server should include at least a CPU, chipset, memory, and disk system; other components will not be described in detail here.

[0055] In this embodiment of the invention, the executable program executed by the processor included in the terminal is specifically: a lane line virtual-real classification method based on geometric features, which includes the following steps:

[0056] Collect lane line data and standardize the lane line representation to a broken line formed by connecting scattered points extending along the lane line direction;

[0057] Based on the angle formed between the lines connecting the lane lines in the front and rear segments and the lane lines in the rear segment, find the adjacent lines of each lane line.

[0058] The classification of lane lines as solid or dashed is determined by the ratio of the length of the lane line itself to its distance from the adjacent line.

[0059] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0060] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0061] Those skilled in the art will recognize that the modules, units, and / or method steps of the various embodiments described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0062] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A lane line solid / dummy classification method based on geometric features, characterized in that, Includes the following steps: Collect lane line data and standardize the lane line representation to a broken line formed by connecting scattered points extending along the lane line direction; Based on the angle formed between the lines connecting the lane lines in the front and rear segments and the lane lines in the rear segment, find the adjacent lines of each lane line. The classification of lane lines as solid or dashed is determined based on the ratio of the length of the lane line itself to its distance from the adjacent line. The step of finding the adjacent connection points of each lane line based on the angle formed by the lines connecting the lane lines in the preceding and following segments and the lane lines in the following segment includes: Using the start and end points of the current lane line as the center, search for adjacent lane lines within a preset radius threshold range; Connect the end points of adjacent lane lines within a preset radius threshold range centered on the starting point to the starting point of the current lane line to form the first connection point; Connect the starting points of adjacent lane lines within a preset radius threshold range centered on the end point to the end point of the current lane line to form a second connection point. Calculate the cosine of the first angle between each first point connection and the broken line where the current lane line starts, and the cosine of the second angle between each second point connection and the broken line where the current lane line ends. Based on the comparison results of the first and second angle cosine values ​​with the preset adjacent connection cosine thresholds, the preparatory front adjacent connection set and the rear adjacent connection set are obtained respectively. The first-shortest pre-adjacent connection from the set of pre-prepared adjacent connections is selected as the pre-adjacent connection of the current lane line, and the first-shortest pre-rear connection from the set of pre-prepared adjacent connections is selected as the rear connection of the current lane line.

2. The lane line solid / dummy classification method based on geometric features according to claim 1, characterized in that, When the lane line data is in the form of a quadrilateral, compare the lengths of the two adjacent sides of the first index point, take the midpoint of the shorter side and the midpoint of the other shorter side set opposite to the shorter side, and take the equidistant points of the connecting line to split the connecting line into multiple sequentially connected broken lines.

3. The lane line solid / dummy classification method based on geometric features according to claim 1, characterized in that, The angle formed between the connection between the lane lines located in the front and rear segments and the lane lines located in the rear segment includes: The angle formed between the line connecting the end point of the preceding lane line and the starting point of the current lane line, and the broken line where the starting point of the current lane line is located.

4. The lane line solid / dummy classification method based on geometric features according to claim 1, characterized in that, The adjacent lane lines within the preset radius threshold range include all adjacent lane lines that are within the preset radius threshold range and those that intersect with the preset radius threshold range.

5. The lane line solid / dummy classification method based on geometric features according to claim 1, characterized in that, The process of classifying lane lines as solid or dashed based on the ratio of their length to the distance to their adjacent lines includes: The length ratio of the first connection point to the current lane line is calculated as the preceding neighbor ratio, and the length ratio of the second connection point to the current lane line is calculated as the following neighbor ratio. The preceding neighbor ratio and the following neighbor ratio are compared with the preset dashed line neighbor ratio threshold range. Based on the comparison results, the dashed and solid results of the current lane line are determined.

6. The lane line solid / dummy classification method based on geometric features according to claim 5, characterized in that, If the ratio of the current adjacent ratio to the next adjacent ratio falls within the preset adjacent ratio threshold range, then the current lane line is determined to be a dashed line; otherwise, it is a solid line.

7. A lane line solid / virtual classification system based on geometric features, characterized in that, Includes the following functional modules: The format unification module is used to collect lane line data and unify the lane line representation to a broken line form formed by connecting scattered points extending along the lane line direction. The adjacent line lookup module is used to find the adjacent lines of each lane line based on the angle formed by the line connecting the lane lines located in the front and rear segments and the lane lines located in the rear segment. The virtual / real lane marking module is used to determine the virtual / real lane marking classification based on the ratio of the length of the lane marking itself to its distance from the adjacent lane marking. The adjacent line lookup module is specifically used for: Using the start and end points of the current lane line as the center, search for adjacent lane lines within a preset radius threshold range; Connect the end points of adjacent lane lines within a preset radius threshold range centered on the starting point to the starting point of the current lane line to form the first connection point; Connect the starting points of adjacent lane lines within a preset radius threshold range centered on the end point to the end point of the current lane line to form a second connection point. Calculate the cosine of the first angle between each first point connection and the broken line where the current lane line starts, and the cosine of the second angle between each second point connection and the broken line where the current lane line ends. Based on the comparison results of the first and second angle cosine values ​​with the preset adjacent connection cosine thresholds, the preparatory front adjacent connection set and the rear adjacent connection set are obtained respectively. The first-shortest pre-adjacent connection from the set of pre-prepared adjacent connections is selected as the pre-adjacent connection of the current lane line, and the first-shortest pre-rear connection from the set of pre-prepared adjacent connections is selected as the rear connection of the current lane line.

8. A server comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the lane line virtual-real classification method based on geometric features as described in any one of claims 1 to 6.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the lane line virtual-real classification method based on geometric features as described in any one of claims 1 to 6.

Images