Road marking line extraction method, device, electronic device and storage medium

Abstract

The present disclosure provides a road marking line extraction method, which is applied to the field of computer technology, including: obtaining all scan lines of a target road, wherein all scan lines include road marking line contour points, performing adaptive convolution on the echo intensity of the midpoint of each scan line, obtaining the convolution result of the echo intensity of the midpoint of each scan line, extracting the road marking line contour points in each scan line based on the convolution result of each scan line, and extracting the road marking lines of the target road based on the road marking line contour points in each scan line. The present application also discloses a road marking line extraction device, electronic equipment, and storage medium, which can accurately extract road marking lines.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method, device, electronic device and storage medium for extracting road marking lines. Background Art

[0002] As a core component of autonomous driving technology, high-precision maps play an important role in assisting vehicle driving decisions, obstacle detection and avoidance, etc. To ensure the safety of autonomous driving, high-precision maps require detailed road information, among which road traffic markings (hereinafter referred to as road markings) can guide, regulate and prohibit vehicle behavior during driving.

[0003] Most current methods for extracting road marking lines are based on traditional two-dimensional optical images or three-dimensional laser point cloud data. Traditional two-dimensional optical images are easily affected by weather and light during the acquisition process and cannot directly obtain the three-dimensional coordinate information of objects, which is not conducive to the efficient and automated production of high-precision maps. The vehicle-mounted laser scanning system is not restricted by weather and light, and can quickly and efficiently obtain high-precision dense three-dimensional point cloud data of the road surface. It also records the reflection intensity information of the target object to the laser, providing a reliable data basis for road marking line detection.

[0004] The road marking line detection methods based on point cloud data can be divided into two categories: in the first category, the road marking line is extracted by generating a two-dimensional intensity image based on the point cloud intensity information. The uneven distribution of point clouds makes the strategy of extracting road marking lines based on intensity thresholds less robust and difficult to obtain complete and accurate extraction results. The road marking line extraction and classification method that combines intensity images and neural network models is complex to implement and has high requirements on computer performance. The training of the extraction model depends on a large number of manually labeled data sets. The second category of road marking line extraction methods based on discrete point clouds can avoid the error introduced by the step of converting point clouds to images and avoid the influence of image resolution on extraction accuracy, effectively improving the accuracy and completeness of road marking line extraction results. However, in the application of large-scale scene massive point cloud data, the uneven distribution of point cloud data still makes it difficult to extract marking lines accurately and completely. Although the neural network model based on discrete point clouds has a high level of automation, it still requires a large number of manually labeled data sets in the early stage, and has limitations in processing massive point cloud data.

[0005] In summary, there are three main difficulties in extracting road markings based on the vehicle-mounted laser scanning system:

[0006] 1) Changes in the distance between the scanner and the target and the laser incident angle result in uneven distribution of point cloud echo intensity and density; 2) The intensity contrast between the worn road markings and the asphalt road surface in old roads is low; 3) During the point cloud data scanning process, obstacles such as vehicles block the road surface point cloud data, resulting in missing road point cloud data. Summary of the invention

[0007] The main purpose of the present application is to provide a road marking line extraction method, device, electronic device and storage medium to solve at least one of the above-mentioned technical problems.

[0008] To achieve the above-mentioned purpose, the first aspect of the embodiment of the present application provides a road marking line extraction method, comprising:

[0009] Acquire all scan lines of the target road, wherein all scan lines include road surface marking line contour points;

[0010] Adaptively convolve the echo intensity of the scanning point in each scanning line to obtain the intensity convolution result of the scanning point in each scanning line;

[0011] Extracting the road marking line contour points in each of the scan lines based on the intensity convolution results of the scan points in each of the scan lines;

[0012] The road surface marking lines of the target road are extracted based on the road surface marking line contour points in each of the scan lines.

[0013] In one embodiment of the present application, extracting the road marking line contour points in each scanning line based on the intensity convolution result of the scanning points in each scanning line includes:

[0014] Determining a convolution result threshold for each of the scan lines;

[0015] Extracting scanning points in each of the scanning lines whose convolution results are greater than the corresponding convolution result threshold;

[0016] Based on the scanning points in each of the scanning lines, road surface marking line contour points in each of the scanning lines are determined.

[0017] In an embodiment of the present application, determining the road surface marking line contour points in each of the scan lines based on the scan points in each of the scan lines includes:

[0018] The scanning point in each of the scanning lines is determined as the road surface marking line contour point.

[0019] In an embodiment of the present application, determining the road surface marking line contour points in each of the scan lines based on the scan points in each of the scan lines includes:

[0020] Obtaining two adjacent points of each of the scanning points in the corresponding scanning line;

[0021] Calculating the intensity change value between each of the scanning points and the corresponding two adjacent points;

[0022] Extracting scanning points in each of the scanning lines whose intensity change values ​​are greater than a preset intensity threshold;

[0023] The scanning point in each of the scanning lines is determined as the road surface marking line contour point.

[0024] In an embodiment of the present application, determining the convolution result threshold of each scan line includes:

[0025] Calculating the mean and variance of the convolution results of all scanning points in each of the scanning lines;

[0026] Calculating a convolution result threshold of each of the scan lines based on the mean and the variance;

[0027] Among them, T1=μ+a×σ, T1 represents the convolution result threshold, μ represents the mean of the convolution results of all scanning points in each scanning line, σ represents the variance of the convolution results of all scanning points in each scanning line, and a represents a constant.

[0028] In an embodiment of the present application, the adaptive convolution is performed on the echo intensity of the scanning point in each scanning line to obtain the intensity convolution result of the scanning point in each scanning line, including:

[0029] For each of the scan lines, obtain the scan point p in the scan line i , p i (i∈[1,n]);

[0030] Take the scanning point p i is the center of the circle, and the distance r is the radius to search for the scanning point p i The scan points in the neighborhood {p i1 , p i2 , ..., p it};

[0031] The scanning point p is generated by taking the scanning point within the distance r and the Gaussian first-order derivative as a template. i The corresponding convolution kernel The calculation formula of the convolution kernel is as follows:

[0032] x j =jt / 2(j∈[1,t])

[0033]

[0034] Where j is the scanning point p i The jth point in the neighborhood, k′ j is the convolution kernel value corresponding to the j-th point;

[0035] For the convolution kernel Perform normalization to obtain the final convolution kernel;

[0036] The scanning point p i Convolve with the corresponding final convolution kernel to obtain the convolution result of each scanning point in the scanning line.

[0037] In an embodiment of the present application, extracting the road surface marking lines of the target road based on the road surface marking line contour points in each of the scan lines includes:

[0038] Fitting all the road surface marking line contour points to obtain preliminary road surface marking lines of the target road;

[0039] Based on the preliminary road surface marking lines of the target road, determining the missing portion of the road surface marking lines of the target road;

[0040] A Kalman filter algorithm is used to fill in the missing part of the road surface marking line of the target road to obtain the road surface marking line of the target road.

[0041] A second aspect of an embodiment of the present application provides a road marking line extraction device, comprising:

[0042] The acquisition module is used to acquire all scan lines of the target road, wherein all scan lines include road surface marking line contour points.

[0043] The convolution module is used to perform adaptive convolution on the echo intensity of the scanning point in each scanning line to obtain the intensity convolution result of the scanning point in each scanning line.

[0044] The point extraction module is used to extract the road surface marking line contour points in each of the scan lines based on the convolution result of each of the scan lines.

[0045] The line extraction module is used to extract the road surface marking lines of the target road based on the road surface marking line contour points in each of the scan lines.

[0046] In one embodiment of the present application, the point extraction module includes:

[0047] A first determination submodule, used to determine a convolution result threshold of each of the scan lines;

[0048] An extraction submodule, used for extracting scanning points in each of the scanning lines whose convolution results are greater than a corresponding convolution result threshold;

[0049] The second determination submodule determines the road surface marking line contour points in each of the scan lines based on the scan points in each of the scan lines.

[0050] In one embodiment of the present application, the second determining submodule is specifically configured to determine the scanning point in each of the scanning lines as the road surface marking line contour point.

[0051] In one embodiment of the present application, the second determining submodule includes:

[0052] An acquisition unit, used for acquiring two adjacent points of each scanning point in a corresponding scanning line;

[0053] A calculation unit, used for calculating the intensity change value between each of the scanning points and the corresponding two adjacent points;

[0054] An extraction unit, used for extracting scanning points in each of the scanning lines whose intensity change values ​​are greater than a preset intensity threshold;

[0055] A determination unit is used to determine the scanning point in each of the scanning lines as the road surface marking line contour point.

[0056] In one embodiment of the present application, the first determining submodule includes:

[0057] A mean-variance calculation unit, used for calculating the mean and variance of the convolution results of all scanning points in each of the scanning lines;

[0058] A threshold calculation unit, used for calculating a convolution result threshold of each of the scan lines based on the mean and the variance;

[0059] Among them, T1=μ+a×σ, T1 represents the convolution result threshold, μ represents the mean of the convolution results of all scanning points in each scanning line, σ represents the variance of the convolution results of all scanning points in each scanning line, and a represents a constant.

[0060] In one embodiment of the present application, the convolution module includes:

[0061] The acquisition submodule is used to acquire the scanning point p in each scanning line. i , p i (i∈[1,n]);

[0062] The search submodule is used to scan the point p i is the center of the circle, and the distance r is the radius to search for the scanning point p i The scan points in the neighborhood {p i1 , p i2 ,…,p it};

[0063] A generation submodule is used to generate the scanning point p using the scanning point within the distance r and the Gaussian first-order derivative as a template. i The corresponding convolution kernel The calculation formula of the convolution kernel is as follows:

[0064] x j =jt / 2(j∈[1,t])

[0065]

[0066] Where j is the scanning point p i The jth point in the neighborhood, k′ j is the convolution kernel value corresponding to the j-th point;

[0067] Normalization submodule, used to normalize the convolution kernel Perform normalization to obtain the final convolution kernel;

[0068] The convolution submodule is used to transform the scan point p i Convolve with the corresponding final convolution kernel to obtain the convolution result of each scanning point in the scanning line.

[0069] In one embodiment of the present application, the line extraction module includes:

[0070] A fitting submodule, used for fitting all the road surface marking line contour points to obtain the preliminary road surface marking line of the target road;

[0071] A missing portion determination submodule, configured to determine a missing portion of a road surface marking line of the target road based on the preliminary road surface marking line of the target road;

[0072] The filling submodule is used to fill in the missing part of the road surface marking line of the target road by using a Kalman filter algorithm to obtain the road surface marking line of the target road.

[0073] A third aspect of the embodiments of the present application provides an electronic device, including:

[0074] A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the road marking line extraction method provided in the first aspect of the embodiment of the present application when executing the program.

[0075] A fourth aspect of an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the road surface marking line extraction method provided in the first aspect of the embodiment of the present application is implemented.

[0076] From the above embodiments of the present application, it can be seen that the road marking line extraction method, device, electronic device and storage medium provided by the present application combine the scanning line with adaptive convolution to achieve automatic and high-precision extraction of road marking line points under the condition of uneven density distribution. Compared with other road marking line extraction methods, it has stronger universality, simpler implementation process and higher extraction accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without paying any creative work.

[0078] Figure 1 A schematic diagram of a process for extracting road marking lines provided in an embodiment of the present application;

[0079] Figure 2(a) to Figure 2(d) A schematic diagram of a process for filling road marking lines provided in an embodiment of the present application;

[0080] Figure 3 A schematic diagram of the structure of a road marking line extraction device provided in one embodiment of the present application;

[0081] Figure 4 A schematic diagram of the hardware structure of an electronic device is shown. DETAILED DESCRIPTION

[0082] In order to make the purpose, features, and advantages of the present application more obvious and understandable, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0083] See also Figure 1 , Figure 1 This is a flow chart of a road marking line extraction method provided in an embodiment of the present application. The method can be applied to electronic devices. The method mainly includes the following operations:

[0084] S101. Acquire all scan lines of a target road, wherein all scan lines include road surface marking line contour points.

[0085] S102, performing adaptive convolution on the echo intensity of each scanning point in the scanning line to obtain the intensity convolution result of each scanning point in the scanning line.

[0086] S103: extracting the road surface marking line contour points in each scanning line based on the intensity convolution result of the scanning points in each scanning line.

[0087] S104: extracting the road surface marking lines of the target road based on the road surface marking line contour points in each of the scan lines.

[0088] In the present disclosure, in operation S101, all scan lines of the acquired target road may be obtained from point cloud data of the original road surface collected by the vehicle-mounted laser scanning system, and then the data is divided into multiple scan lines according to the collection time of the road surface point cloud data.

[0089] In an embodiment of the present application, operation S102, performing adaptive convolution on the intensity of the scanning point in each scanning line, and obtaining the intensity convolution result of the scanning point in each scanning line includes the following steps:

[0090] S1. For each scan line, obtain the scan point p in the scan line. i , p i (i∈[1,n]).

[0091] S2, with the scanning point p i is the center of the circle and the distance r is the radius to search for the scanning point p i Scan points in the neighborhood (p i1 , p i2 , ..., p it}.

[0092] S3, using the scanning point within the distance r and the Gaussian first-order derivative as a template, generate the scanning point p i The corresponding convolution kernel The calculation formula of the convolution kernel is as follows:

[0093] x j =jt / 2(j∈[1,t])

[0094]

[0095] Among them, j is the scanning point p i The jth point in the neighborhood, k′ j is the convolution kernel value corresponding to the j-th point;

[0096] S4, for the convolution kernel Normalization is performed to obtain the final convolution kernel.

[0097] Among them, the final convolution kernel The specific calculation formula is as follows:

[0098]

[0099] In the formula, k j is the normalized convolution kernel value corresponding to the jth point.

[0100] S5, the scanning point p i Convolve with the corresponding final convolution kernel to obtain the convolution result of each scanning point in the scan line.

[0101] The convolution formula is as follows:

[0102]

[0103] In the formula, c i For the scan point p i The convolution result, I j For the scan point p i The echo strength value of the scan point in the neighborhood.

[0104] The above steps S1-S5 are repeated until the convolution result of each scanning point in the scanning line is obtained.

[0105] In one embodiment of the present application, operation S103, based on the intensity convolution result of the scanning points in each scanning line, extracting the contour points of the road surface marking lines in each scanning line includes: determining the convolution result threshold of each scanning line, extracting the scanning points in each scanning line whose convolution results are greater than the corresponding convolution result threshold, and determining the contour points of the road surface marking lines in each scanning line based on the scanning points in each scanning line.

[0106] In one embodiment of the present application, the above-mentioned determination of the convolution result threshold of each scan line includes: determining the mean and variance of the convolution results of all scanning points in each scan line, and calculating the convolution result threshold of each scan line based on the mean and the variance.

[0107] Among them, T1=μ+a×σ, T1 represents the convolution result threshold, μ represents the mean of the convolution results of all scanning points in each scanning line, σ represents the variance of the convolution results of all scanning points in each scanning line, and a represents a constant. The constant a is related to the ratio of the number of road marking points to the number of asphalt road surface points in the scanning line. The larger the proportion of road marking points, the smaller the constant a. The specific value is set based on experience.

[0108] In an embodiment of the present application, the determining of the road surface marking line contour point in each scanning line based on the scanning point in each scanning line includes: determining the scanning point in each scanning line as the road surface marking line contour point.

[0109] In one embodiment of the present application, the above-mentioned determination of the road surface marking line contour point in each scanning line based on the scanning point in each scanning line includes: obtaining two adjacent points of each scanning point in the corresponding scanning line, calculating the intensity change value between each scanning point and the corresponding two adjacent points, extracting the scanning points in each scanning line whose intensity change values ​​are greater than a preset intensity threshold, and determining the scanning points in each scanning line as the road surface marking line contour points.

[0110] In this disclosure, T2=|mean(I1)-mean(I2)|, median is the median calculation function, mean is the mean calculation function, I is the echo intensity value of all scanning points in a single scanning line, and T2 is the intensity change threshold.

[0111] In one embodiment of the present application, operation S104, based on the pavement marking line contour points in each of the scan lines, extracting the pavement marking lines of the target road includes: fitting all of the pavement marking line contour points to obtain preliminary pavement marking lines of the target road, determining the missing parts of the pavement marking lines of the target road based on the preliminary pavement marking lines of the target road, and using a Kalman filtering algorithm to fill in the missing parts of the pavement marking lines of the target road to obtain the pavement marking lines of the target road.

[0112] In the present disclosure, the road marking line profile can be fitted in combination with the random sample consensus algorithm (RANSAC) to achieve vector extraction of the road marking line profile. The Kalman filter algorithm is used in combination with the preliminary road marking line of the target road to obtain the road marking line of the target road, which can avoid the problem that the vehicle-mounted laser scanning system is blocked by obstacles such as vehicles during the scanning process, and the fitting result cannot fully present the original appearance of the marking line due to partial missing of the road point cloud data.

[0113] In the present disclosure, considering that the same type of road marking lines usually appear in a continuous or evenly spaced pattern, the same type of complete marking lines can provide a reference for the prediction of defective marking lines. Figure 2(a) to Figure 2(d) As shown by the dotted line mark, determine the position of the defect mark line M i , combined with M i Full identification line R in forward and backward directions f With R b , the accurate contour information of the defective identification line can be iteratively predicted. The specific filling process is as follows:

[0114] (1) As shown in Figure 2(a), the missing marking line M is determined based on the preliminary road marking line. i , search forward and backward for the complete marking line R at the defective marking line f , R b .

[0115] (2) As shown in Figure 2(b), according to the complete identification line R f Iterate from front to back and follow the complete identification line R b Iterate from back to front. In the iterative process, the center point c of the adjacent identification line i The spacing Δd between the two and the direction vector As the measured value z and the predicted value x, the iteration formula is as follows, where the first two formulas are prediction equations and the last three formulas are update equations.

[0116]

[0117]

[0118]

[0119]

[0120]

[0121] (3) According to (2), the predicted values ​​of the two-way iteration are obtained Two complete marking lines adjacent to the defective marking line The four corner points of the defect marker line are shown in Figure 2(c), and the contour positions M of the two defect marker lines are calculated. f , M b .

[0122] (4) According to the number of iterations n from the front to the back of the defect identification line f and the number of iterations from back to front n b , the contour position M of the two defect identification lines is calculated according to the following formula f , M b The weighted value is shown in Figure 2(d) to obtain the final predicted defect identification line contour position R i .

[0123]

[0124] According to the disclosed embodiment, adaptive convolution is performed in combination with the area surrounding the scanning point to achieve automatic and high-precision extraction of road marking line points under conditions of uneven density distribution. Compared with other road marking line extraction methods, this patent is more universal, the implementation process is simpler, and the extraction accuracy is higher. In combination with complete road marking line information, the missing parts in the marking line extraction results are filled. Compared with other road marking line extraction methods, this patent can achieve complete extraction of road marking lines in road environments with missing data.

[0125] See also Figure 3 , Figure 3 It is a structural diagram of a road marking line extraction device provided in an embodiment of the present application. The device can be built into an electronic device. The device mainly includes: an acquisition module 310, a convolution module 320, a point extraction module 330 and a line extraction module 340.

[0126] The acquisition module 310 is used to acquire all scan lines of the target road, where all scan lines include road surface marking line contour points.

[0127] The convolution module 320 is used to perform adaptive convolution on the echo intensity of the scanning point in each scanning line to obtain the intensity convolution result of the scanning point in each scanning line.

[0128] The point extraction module 330 is used to extract the road surface marking line contour points in each scanning line based on the intensity convolution result of the scanning points in each scanning line.

[0129] The line extraction module 340 is used to extract the road surface marking lines of the target road based on the road surface marking line contour points in each of the scan lines.

[0130] In one embodiment of the present application, the point extraction module 330 includes:

[0131] A first determination submodule, used to determine a convolution result threshold of each scan line;

[0132] An extraction submodule, used for extracting scanning points whose convolution results are greater than corresponding convolution result thresholds in each scanning line;

[0133] The second determination submodule determines the road surface marking line contour point in each scanning line based on the scanning point in each scanning line.

[0134] In one embodiment of the present application, the second determining submodule is specifically used to determine the scanning point in each of the scanning lines as the road surface marking line contour point.

[0135] In one embodiment of the present application, the second determining submodule includes:

[0136] An acquisition unit, used for acquiring two adjacent points of each scanning point in a corresponding scanning line;

[0137] A calculation unit, used for calculating the intensity change value between each scanning point and the corresponding two adjacent points;

[0138] An extraction unit, used for extracting scanning points whose intensity change value is greater than a preset intensity threshold value in each scanning line;

[0139] The determination unit is used to determine the scanning point in each scanning line as the road surface marking line contour point.

[0140] In one embodiment of the present application, the first determining submodule includes:

[0141] A mean-variance calculation unit, used for the mean and variance of the convolution results of all scanning points in each scanning line;

[0142] A threshold calculation unit, used for calculating a convolution result threshold of each scan line based on the mean and the variance;

[0143] Among them, T1=μ+a×σ, T1 represents the convolution result threshold, μ represents the mean of the convolution results of all scanning points in each scanning line, σ represents the variance of the convolution results of all scanning points in each scanning line, and a represents a constant.

[0144] In one embodiment of the present application, the convolution module 320 includes:

[0145] The acquisition submodule is used to obtain the scanning point p in each scanning line. i , p i (i∈[1,n]);

[0146] The search submodule is used to scan the point p i is the center of the circle and the distance r is the radius to search for the scanning point p i The scan points in the neighborhood {p i1 , p i2 , ..., p it};

[0147] The generation submodule is used to generate the scanning point p using the scanning point within the distance r and the Gaussian first-order derivative as a template i The corresponding convolution kernel The calculation formula of the convolution kernel is as follows:

[0148] x j =jt / 2(j∈[1,t])

[0149]

[0150] Among them, j is the scanning point p i The jth point in the neighborhood, k′ j is the convolution kernel value corresponding to the j-th point;

[0151] Normalization submodule is used to normalize the convolution kernel Perform normalization to obtain the final convolution kernel;

[0152] The convolution submodule is used to transform the scan point p i Convolve with the corresponding final convolution kernel to obtain the convolution result of each scanning point in the scan line.

[0153] In one embodiment of the present application, the line extraction module 340 includes:

[0154] A fitting submodule is used to fit all the road surface marking line contour points to obtain the preliminary road surface marking line of the target road;

[0155] A missing portion determination submodule, used to determine the missing portion of the road surface marking line of the target road based on the preliminary road surface marking line of the target road;

[0156] The filling submodule is used to fill in the missing part of the road surface marking line of the target road by using a Kalman filter algorithm to obtain the road surface marking line of the target road.

[0157] See also Figure 4 , Figure 4 A hardware structure diagram of an electronic device is shown.

[0158] The electronic device described in this embodiment includes:

[0159] The memory 41, the processor 42 and the computer program stored in the memory 41 and executable on the processor, when the processor executes the program, the aforementioned Figure 1 The road surface marking line extraction method described in the illustrated embodiment.

[0160] Furthermore, the electronic device also includes:

[0161] At least one input device 43 ; at least one output device 44 .

[0162] The memory 41 , processor 42 , input device 43 and output device 44 are connected via a bus 45 .

[0163] The input device 43 may be a camera, a touch panel, a physical button or a mouse, etc. The output device 44 may be a display screen.

[0164] The memory 41 may be a high-speed random access memory (RAM) memory or a non-volatile memory, such as a disk memory. The memory 41 is used to store a set of executable program codes, and the processor 42 is coupled to the memory 41 .

[0165] Furthermore, the present disclosure also provides a computer-readable storage medium, which may be disposed in the electronic device in the above embodiments. Figure 4 The electronic device in the embodiment shown. The computer readable storage medium stores a computer program, which, when executed by the processor, implements the aforementioned Figure 1 The road marking line extraction method described in the illustrated embodiment. Furthermore, the computer storage medium may also be a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0166] It should be noted that the functional modules in the various embodiments of the present disclosure may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module. The above-mentioned integrated modules may be implemented in the form of hardware or in the form of software functional modules.

[0167] If the integrated module is implemented in the form of a software function module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product.

[0168] It should be noted that, for the convenience of description, the aforementioned method embodiments are all described as a series of action combinations, but those skilled in the art should be aware that the present invention is not limited by the described action sequence, because according to the present invention, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present invention.

[0169] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0170] The above is a description of a road marking line extraction method, device, electronic device and readable storage medium provided by the present invention. For technicians in this field, according to the ideas of the embodiments of the present invention, there will be changes in the specific implementation methods and application scopes. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A method for extracting road marking lines, characterized in that: include: Acquire all scan lines of the target road, wherein all scan lines include road surface marking line contour points; Adaptively convolve the echo intensity of the scanning point in each scanning line to obtain the intensity convolution result of the scanning point in each scanning line; Extracting the road marking line contour points in each of the scan lines based on the intensity convolution results of the scan points in each of the scan lines; Extracting the road surface marking lines of the target road based on the road surface marking line contour points in each of the scan lines; The step of extracting the road marking line contour points in each scanning line based on the intensity convolution result of the scanning points in each scanning line comprises: Determining a convolution result threshold for each of the scan lines; Extracting scanning points in each of the scanning lines whose convolution results are greater than the corresponding convolution result threshold; Based on the scan points in each of the scan lines, determining a road surface marking line contour point in each of the scan lines; The determining of the road surface marking line contour points in each of the scan lines based on the scan points in each of the scan lines comprises: Obtaining two adjacent points of each of the scanning points in the corresponding scanning line; Calculating the intensity change value between each of the scanning points and the corresponding two adjacent points; Extracting scanning points in each of the scanning lines whose intensity change values ​​are greater than a preset intensity threshold; Determine the scanning point in each of the scanning lines as the road surface marking line contour point; The step of adaptively convolving the echo intensity of the scanning point in each scanning line to obtain the intensity convolution result of the scanning point in each scanning line includes: For each of the scan lines, obtain the scan points in the scan line ; The scanning point is the center of the circle, and the distance r is the radius to search for the scanning point Scan points in the neighborhood ; The scanning point within the distance r and the Gaussian first-order derivative are used as templates to generate the scanning point The corresponding convolution kernel , the calculation formula of the convolution kernel is as follows: in, The scanning point Neighborhood Points, For the The convolution kernel value corresponding to the point; For the convolution kernel Perform normalization to obtain the final convolution kernel; The scanning point Convolve with the corresponding final convolution kernel to obtain the convolution result of each scanning point in the scanning line.

2. The road marking line extraction method according to claim 1, characterized in that: Determining the convolution result threshold of each of the scan lines comprises: Calculating the mean and variance of the convolution results of all scanning points in each of the scanning lines; Calculating a convolution result threshold of each of the scan lines based on the mean and the variance; in, , represents the convolution result threshold, represents the mean value of the convolution results of all scanning points in each of the scanning lines, represents the variance of the convolution results of all the scanning points in each of the scanning lines, and a represents a constant.

3. The road marking line extraction method according to claim 1, characterized in that: The step of extracting the road surface marking lines of the target road based on the road surface marking line contour points in each of the scan lines comprises: Fitting all the road surface marking line contour points to obtain preliminary road surface marking lines of the target road; Based on the preliminary road surface marking lines of the target road, determining the missing portion of the road surface marking lines of the target road; A Kalman filter algorithm is used to fill in the missing part of the road surface marking line of the target road to obtain the road surface marking line of the target road.

4. A road marking line extraction device, used to implement the road marking line extraction method according to any one of claims 1 to 3, characterized in that: include: An acquisition module, used for acquiring all scan lines of a target road, wherein all scan lines include road surface marking line contour points; A convolution module, used for performing adaptive convolution on the echo intensity of the scanning point in each scanning line to obtain the intensity convolution result of the scanning point in each scanning line; A point extraction module, used for extracting the road surface marking line contour points in each scanning line based on the intensity convolution result of the scanning points in each scanning line; The line extraction module is used to extract the road surface marking lines of the target road based on the road surface marking line contour points in each of the scan lines.

5. An electronic device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, each step of the road marking line extraction method described in any one of claims 1 to 3 is implemented.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, each step of the road surface marking line extraction method according to any one of claims 1 to 3 is implemented.

Images