Lane line correlation degree determination method, device, equipment and storage medium

Abstract

The present invention relates to the field of autonomous driving technology, and discloses a lane line correlation degree determination method, device, equipment and storage medium, which are used to improve the scoring accuracy of the correlated lane lines between every two frames of lane line detection image data when there is an offset distance between lane lines in different frames and they are relatively misaligned. The lane line correlation degree determination method comprises: acquiring multiple frames of lane line detection image data, each frame of lane line detection image data includes multiple detected lane lines; performing lane line discrete point correlation processing on the multiple frames of lane line detection image data to obtain multiple lane line correlation degree scores between every two frames of lane line detection image data; and determining the correlated lane lines between every two frames of lane line detection image data according to the multiple lane line correlation degree scores.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and in particular to a method, device, equipment and storage medium for determining the degree of association of lane lines. Background Art

[0002] After obtaining the lane line output by the perception detection model, it is necessary to associate multiple lane line detection results in different frames to build a local semantic map. This local high-precision semantic map not only contains the image detected by the vehicle in the current frame, but also includes the image detection results at the time or distance before the current moment. In order to fuse the lane line detection results in different frames at different times, it is necessary to associate and fit the lane lines in different frames into a complete lane line.

[0003] The roads that existing autonomous vehicles travel on generally consist of three lane lines and two lanes. Assuming that the visual distance of the camera of the autonomous vehicle is 100 meters, when the autonomous vehicle travels at high speed from one side of the road to the other side of the road, the perception system continuously detects and outputs the lane line detection results of different frames. When fusing the same three lane lines in different frames, it is necessary to calculate the correlation score between different lane line detection results. The higher the correlation between two objects, the higher the score.

[0004] In traditional object detection and tracking scoring systems, the intersection over union (IoU) calculation method is usually used, and the object expression form of the application scenario is mainly the detection frame. The overlap and the detection frame are well matched because the detection frame can cover the entire detected object. However, when the overlap and the detection frame are applied to the lane line detection scenario, there is a problem of poor accuracy in the lane line association degree scoring. Summary of the invention

[0005] The present invention provides a method, device, equipment and storage medium for determining the degree of lane line association, which are used to improve the scoring accuracy of the associated lane lines between every two frames of lane line detection image data when there is an offset distance between lane lines in different frames and they are relatively misaligned.

[0006] To achieve the above-mentioned purpose, the first aspect of the present invention provides a method for determining the degree of lane line association, comprising: acquiring multiple frames of lane line detection image data, each frame of lane line detection image data including multiple detected lane lines; performing lane line discrete point association processing on the multiple frames of lane line detection image data to obtain multiple lane line association degree scores between every two frames of lane line detection image data; and determining the associated lane lines between every two frames of lane line detection image data based on the multiple lane line association degree scores.

[0007] In a feasible implementation manner, the obtaining of multiple frames of lane line detection image data includes: receiving road environment image data sent by a preset sensor; performing target detection on the road environment image data through a preset perception detection model to obtain multiple frames of lane line detection image data.

[0008] In a feasible implementation manner, the lane line discrete point association processing is performed on the multiple frames of lane line detection image data to obtain multiple lane line association degree scores between every two frames of lane line detection image data, including: measuring the length of each detected lane line corresponding to each frame of lane line detection image data to obtain a lane line length value corresponding to each detected lane line; performing discrete processing on each detected lane line corresponding to each frame of lane line detection image data to obtain multiple discrete points corresponding to each detected lane line; and determining multiple lane line association degree scores between every two frames of lane line detection image data based on the lane line length value corresponding to each detected lane line and the multiple discrete points corresponding to each detected lane line.

[0009] In a feasible implementation manner, the method of determining multiple lane line association degree scores between every two frames of lane line detection image data based on the lane line length value corresponding to each detected lane line and the multiple discrete points corresponding to each detected lane line includes: adding the lane line length values ​​corresponding to any two detected lane lines between every two frames of lane line detection image data to obtain the sum of the lengths of multiple lane lines to be associated between every two frames of lane line detection image data; calculating the sum of the distances of multiple lane lines to be associated between every two frames of lane line detection image data based on the multiple discrete points corresponding to each detected lane line; determining multiple lane line association degree scores between every two frames of lane line detection image data based on the sum of the lengths of the multiple lane lines to be associated and the sum of the distances of the multiple lane lines to be associated, and there is a one-to-one correspondence between the sum of the lengths of the lane lines to be associated, the sum of the distances of the lane lines to be associated, and the association degree scores of the lane lines.

[0010] In a feasible implementation manner, the determining of multiple lane line association degree scores between every two frames of lane line detection image data based on the sum of the lengths of the multiple lane lines to be associated and the sum of the distances of the multiple lane lines to be associated includes: judging whether the sum of the distances of the lane lines to be associated is less than the sum of the lengths of the lane lines to be associated; if the sum of the distances of the lane lines to be associated is less than the sum of the lengths of the lane lines to be associated, calculating a target ratio between the sum of the distances of the lane lines to be associated and the sum of the lengths of the lane lines to be associated, calculating a target difference between a preset target value and the target ratio, and setting the lane line association degree score between every two frames of lane line detection image data to the target difference; if the sum of the distances of the lane lines to be associated is greater than or equal to the sum of the lengths of the lane lines to be associated, setting the lane line association degree score between every two frames of lane line detection image data to a preset score.

[0011] In a feasible implementation manner, determining the associated lane lines between every two frames of lane line detection image data based on the multiple lane line association degree scores includes: when the association degree score of each lane line is equal to a preset score, determining that the detected lane lines corresponding to the lane line association degree scores between every two frames of lane line detection image data are not associated; when the association degree score of each lane line is greater than the preset score, setting the detected lane lines corresponding to the lane line association degree scores as the associated lane lines between every two frames of lane line detection image data.

[0012] In a feasible implementation manner, after determining the associated lane lines between every two frames of lane line detection image data based on the multiple lane line association degree scores, the lane line association degree determination method also includes: performing multi-frame fusion on the associated lane lines between every two frames of lane line detection image data to obtain multi-frame fused lane line image data; and generating a local semantic map based on the multi-frame fused lane line image data.

[0013] The second aspect of the present invention provides a lane line association degree determination device, including: an acquisition module, used to acquire multiple frames of lane line detection image data, each frame of lane line detection image data includes multiple detected lane lines; an association module, used to perform lane line discrete point association processing on the multiple frames of lane line detection image data to obtain multiple lane line association degree scores between every two frames of lane line detection image data; a determination module, used to determine the associated lane lines between every two frames of lane line detection image data based on the multiple lane line association degree scores.

[0014] In a feasible implementation, the acquisition module is specifically used to: receive road environment image data sent by a preset sensor; perform target detection on the road environment image data through a preset perception detection model to obtain multiple frames of lane line detection image data.

[0015] In a feasible implementation, the associated module also includes: a measuring unit, used to measure the length of each detected lane line corresponding to each frame of lane line detection image data to obtain a lane line length value corresponding to each detected lane line; a discrete unit, used to discretize each detected lane line corresponding to each frame of lane line detection image data to obtain multiple discrete points corresponding to each detected lane line; a determination unit, used to determine multiple lane line association degree scores between every two frames of lane line detection image data based on the lane line length value corresponding to each detected lane line and the multiple discrete points corresponding to each detected lane line.

[0016] In a feasible implementation manner, the determination unit also includes: an adding subunit, used to add the lane line length values ​​corresponding to any two detected lane lines between every two frames of lane line detection image data to obtain the sum of the lengths of multiple lane lines to be associated between every two frames of lane line detection image data; a calculating subunit, used to calculate the sum of the distances of multiple lane lines to be associated between every two frames of lane line detection image data based on multiple discrete points corresponding to each detected lane line; a determining subunit, used to determine multiple lane line association degree scores between every two frames of lane line detection image data based on the sum of the lengths of the multiple lane lines to be associated and the sum of the distances of the multiple lane lines to be associated, and there is a one-to-one correspondence between the sum of the lengths of the lane lines to be associated, the sum of the distances of the lane lines to be associated, and the association degree scores of the lane lines.

[0017] In a feasible implementation manner, the determination subunit is specifically used to: determine whether the sum of the distances of the lane lines to be associated is less than the sum of the lengths of the lane lines to be associated; if the sum of the distances of the lane lines to be associated is less than the sum of the lengths of the lane lines to be associated, calculate the target ratio between the sum of the distances of the lane lines to be associated and the sum of the lengths of the lane lines to be associated, calculate the target difference between the preset target value and the target ratio, and set the association degree score of each lane line between every two frames of lane line detection image data to the target difference; if the sum of the distances of the lane lines to be associated is greater than or equal to the sum of the lengths of the lane lines to be associated, set the association degree score of each lane line between every two frames of lane line detection image data to the preset score.

[0018] In a feasible implementation manner, the determination module is specifically used to: when the correlation degree score of each lane line is equal to a preset score, determine that the detected lane lines corresponding to the correlation degree scores of each lane line between every two frames of lane line detection image data are not correlated; when the correlation degree score of each lane line is greater than the preset score, set the detected lane lines corresponding to the correlation degree scores of each lane line as the correlated lane lines between every two frames of lane line detection image data.

[0019] In a feasible implementation manner, the lane line association degree determination device also includes: a fusion module, used to perform multi-frame fusion of the associated lane lines between every two frames of lane line detection image data to obtain multi-frame fused lane line image data; and a generation module, used to generate a local semantic map based on the multi-frame fused lane line image data.

[0020] The third aspect of the present invention provides a lane line association degree determination device, comprising: a memory and at least one processor, the memory storing instructions, the memory and the at least one processor being interconnected via lines; the at least one processor calling the instructions in the memory so that the lane line association degree determination device executes the above-mentioned lane line association degree determination method.

[0021] A fourth aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute the above-mentioned method for determining the degree of lane line association.

[0022] In the technical solution provided by the present invention, multiple frames of lane line detection image data are obtained, each frame of lane line detection image data includes multiple detected lane lines; lane line discrete point association processing is performed on the multiple frames of lane line detection image data to obtain multiple lane line association degree scores between every two frames of lane line detection image data; and the associated lane lines between every two frames of lane line detection image data are determined based on the multiple lane line association degree scores. In an embodiment of the present invention, by performing lane line discrete point association processing on multiple frames of lane line detection image data to obtain multiple lane line association degree scores between every two frames of lane line detection image data, and determining the associated lane lines between every two frames of lane line detection image data based on the multiple lane line association degree scores, when there is an offset distance between lane lines in different frames and they are relatively misaligned, the scoring accuracy of the associated lane lines between every two frames of lane line detection image data is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 A schematic diagram of an embodiment of a method for determining the degree of association of lane lines in an embodiment of the present invention;

[0024] Figure 2 Schematic diagram of calculation of associated lane lines between every two frames of lane line detection image data in an embodiment of the present invention;

[0025] Figure 3 A schematic diagram of another embodiment of a method for determining the degree of association of lane lines in an embodiment of the present invention;

[0026] Figure 4 A schematic diagram of an embodiment of a lane line correlation degree determination device in an embodiment of the present invention;

[0027] Figure 5 A schematic diagram of another embodiment of a lane line correlation degree determination device in an embodiment of the present invention;

[0028] Figure 6 Schematic diagram of an embodiment of a lane line correlation degree determination device in an embodiment of the present invention. DETAILED DESCRIPTION

[0029] Embodiments of the present invention provide a method, apparatus, device and storage medium for determining the degree of lane line association, which are used to improve the accuracy of scoring the associated lane lines between every two frames of lane line detection image data when there is an offset distance between lane lines in different frames and they are relatively misaligned.

[0030] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0031] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 , an embodiment of the method for determining the degree of association of lane lines in an embodiment of the present invention includes:

[0032] 101. Acquire multiple frames of lane line detection image data, each frame of lane line detection image data including multiple detected lane lines.

[0033] Among them, the multiple frames of lane line detection image data are used to indicate the color or black and white road image data containing lane lines detected by the unmanned vehicle at different viewing distances. There is an association relationship between each frame of lane line detection image data, each frame of lane line detection image data includes multiple detected lane lines, and also includes a lane composed of multiple detected lane lines, and each detected lane line includes a lane line represented by a solid line and a lane line represented by a dotted line. In some embodiments, the server receives road environment image data sent by a preset sensor, wherein the preset sensor can be a camera, and the preset sensor can also be a laser radar, which is not specifically limited here, and the road environment image data is used to indicate the driving environment image information collected during the driving process of the unmanned vehicle; the server performs target detection on the road environment image data through a preset perception detection model to obtain multiple frames of lane line detection image data, that is, the server performs lane line detection on the road environment image data through a preset perception detection model to obtain multiple frames of lane line detection image data, so as to ensure that each frame of lane line detection image data includes multiple detected lane lines. Among them, the preset perception detection model is a pre-trained perception detection model, and the preset perception detection model can belong to a neural network model or a deep learning model, which is not specifically limited here. Furthermore, the server stores the multiple frames of lane line detection image data in a preset database.

[0034] It is understandable that the execution subject of the present invention may be a lane line correlation degree determination device, or may be a terminal or a server, which is not limited here. The embodiment of the present invention is described by taking a server as the execution subject as an example.

[0035] 102. Perform lane line discrete point association processing on multiple frames of lane line detection image data to obtain multiple lane line association degree scores between every two frames of lane line detection image data.

[0036] That is, the server performs overlap calculations on each frame of lane line detection image data between every two frames of lane line detection image data to obtain multiple lane line correlation scores between every two frames of lane line detection image data. It should be noted that there is an offset distance between each detected lane line between every two frames of lane line detection image data, and they are relatively misaligned. The present invention is applicable to detected lane lines of the solid line type, and is also applicable to detected lane lines of the dotted line type. This embodiment only takes the detected lane lines of the solid line type as an example. Figure 2 As shown, it includes detected lane line a, detected lane line b, detected lane line c and detected lane line d, where a and c belong to the same frame (that is, the same frame) lane line detection image data, and b and d belong to the same frame lane line detection image data. There is an offset distance between a and b and d, and they are relatively misaligned, and there is an offset distance between c and b and d, and they are relatively misaligned, that is, the server needs to determine the lane line association degree score between a and b and d, and the lane line association degree score between c and b and d.

[0037] In some embodiments, the server extracts the lane line interest regions corresponding to the multiple frames of lane line detection image data respectively; the server performs Hough line detection on the lane line interest regions corresponding to the lane line detection image data of each frame to extract multiple detected lane lines corresponding to the lane line detection image data of each frame, and each detected lane line has a corresponding lane line length value; the server discretizes each detected lane line corresponding to each frame of lane line detection image data into corresponding multiple discrete points, for example, Figure 2 As shown, the detected lane line a, the detected lane line b, the detected lane line c and the detected lane line d are all discretized into multiple discrete points; the server calculates the lane line distance between each detected lane line between every two frames of lane line detection image data based on the multiple discrete points corresponding to each detected lane line; the server performs lane line correlation degree scoring processing based on the lane line length value corresponding to each detected lane line and the lane line distance between each detected lane line, and obtains multiple lane line correlation degree scores between every two frames of lane line detection image data. Among them, each lane line correlation degree score can be less than the preset score, or greater than the preset score, which is not limited here. The preset score can be 0 or other values, which are not limited here. Figure 2As shown, for example, the server determines that the lane line association degree score between the detected lane line a and the detected lane line b is greater than 0 and less than or equal to 1, the server determines that the lane line association degree score between the detected lane line c and the detected lane line d is greater than 0 and less than or equal to 1, the server determines that the lane line association degree score between the detected lane line a and the detected lane line d is less than 0, and the server determines that the lane line association degree score between the detected lane line c and the detected lane line b is less than 0.

[0038] 103. Determine the associated lane line between every two frames of lane line detection image data according to the multiple lane line association degree scores.

[0039] It should be noted that the lane line association degree score is used to determine the association degree between any two detected lane lines corresponding to the lane line association degree score between each two frames of lane line detection image data. That is, if the lane line association degree score is greater than the preset score, the server determines that any two detected lane lines corresponding to the lane line association degree score are associated, and if the lane line association degree score is less than the preset score, the server determines that any two detected lane lines corresponding to the lane line association degree score are not associated.

[0040] In some embodiments, the server determines whether the lane line association degree score is greater than a preset score; if the lane line association degree score is greater than the preset score, the server sets any two detected lane lines corresponding to the lane line association degree score as the associated lane lines between every two frames of lane line detection image data; if the lane line association degree score is less than or equal to the preset score, the server determines that any two detected lane lines corresponding to the lane line association degree score are not associated. Figure 2 As shown, for example, if the preset score is 0, the lane line association score between the detected lane line a and the detected lane line b is 0.22, the lane line association score between the detected lane line c and the detected lane line d is 0.34, the lane line association score between the detected lane line a and the detected lane line d is -1, and the lane line association score between the detected lane line c and the detected lane line b is -1, then the server determines that the detected lane line a and the detected lane line b are the associated lane lines between the two frames of lane line detection image data, and the detected lane line c and the detected lane line d are the associated lane lines between the two frames of lane line detection image data. Furthermore, the server can also perform operations such as lane line fusion and lane line fitting processing on the detected lane line a and the detected lane line b, and the detected lane line c and the detected lane line d, respectively.

[0041] In an embodiment of the present invention, lane line discrete point association processing is performed on multiple frames of lane line detection image data to obtain multiple lane line association degree scores between every two frames of lane line detection image data, and the associated lane lines between every two frames of lane line detection image data are determined based on the multiple lane line association degree scores. When there is an offset distance between the lane lines in different frames and they are relatively misaligned, the scoring accuracy of the associated lane lines between every two frames of lane line detection image data is improved.

[0042] See also Figure 3 Another embodiment of the method for determining the degree of association of lane lines in the embodiment of the present invention includes:

[0043] 301. Acquire multiple frames of lane line detection image data, each frame of lane line detection image data including multiple detected lane lines.

[0044] The specific execution process of step 301 is similar to the specific execution process of step 101, and will not be repeated here.

[0045] 302. Measure the length of each detected lane line corresponding to each frame of lane line detection image data to obtain a lane line length value corresponding to each detected lane line.

[0046] Among them, each frame of lane line detection image data includes multiple detected lane lines, lanes composed of multiple detected lane lines, and other road sign information (for example, straight-ahead signs and speed limit signs, etc.), which are not specifically limited here. In some embodiments, the server performs corner feature extraction and lane line extraction on each frame of lane line detection image data to obtain each detected lane line corresponding to each frame of lane line detection image data; the server calculates the pixel accuracy corresponding to each frame of lane line detection image data based on each detected lane line corresponding to each frame of lane line detection image data; the server determines the lane line length value corresponding to each detected lane line based on the pixel accuracy corresponding to each frame of lane line detection image data and the preset distortion parameters corresponding to the preset sensor. For example, Figure 2 As shown, the detected lane lines are a, b, c and d respectively, and the lane line length corresponding to the detected lane line a is L a , the lane line length value corresponding to the detected lane line b is L b , the lane line length value corresponding to the detected lane line c is L c And the lane line length value corresponding to the detected lane line d is L d .

[0047] 303. Discrete each detected lane line corresponding to each frame of lane line detection image data to obtain a plurality of discrete points corresponding to each detected lane line.

[0048] That is, the server discretizes each detected lane line corresponding to each frame of lane line detection image data into multiple discrete points corresponding to each detected lane line according to the preset discretization rule. The preset discretization rule is used to indicate that each detected lane line corresponding to each frame of lane line detection image data is sampled at a preset distance. The preset distance can be 1 meter or 1.5 meters, which is not limited here. For example, Figure 2 As shown in FIG. 1 , the detected lane lines are a, b, c, and d. The server samples one point every one meter on the detected lane line a, the detected lane line b, the detected lane line c, and the detected lane line d. Accordingly, the multiple discrete points corresponding to the detected lane line a are A. 1 , A 2 , A 3 , A 4 and A 5 , the multiple discrete points corresponding to the detected lane line b are B 1 , B 2 , B 3 , B 4 and B 5 , the multiple discrete points corresponding to the detected lane line c are C 1 , C 2 , C 3 , C 4 and C 5 , the multiple discrete points corresponding to the detected lane line b are D 1 , D 2 , D 3 , D 4 and D 5 .

[0049] It should be noted that the number of discrete points corresponding to each detected lane line a, b, c and d may be the same or different, and is not specifically limited here.

[0050] 304. Determine multiple lane line association degree scores between every two frames of lane line detection image data according to the lane line length value corresponding to each detected lane line and the multiple discrete points corresponding to each detected lane line.

[0051] It should be noted that the road on which the autonomous driving vehicle is traveling includes at least three lane lines and two lanes, and there are at least three pairs of lane lines associated between every two frames of lane line detection image data, and the number of lane line association degree scores between every two frames of lane line detection image data is at least 3. In some embodiments, the server adds the lane line length values ​​corresponding to any two detected lane lines between every two frames of lane line detection image data to obtain the sum of the lengths of multiple lane lines to be associated between every two frames of lane line detection image data, such as Figure 2As shown, for example, the detected lane lines are a, b, c and d, and the lane line length values ​​corresponding to the detected lane lines a, b, c and d are L a , L b , L c and L d , where the detected lane line a is the lane line to be associated with the detected lane line b and the detected lane line d, respectively, and the detected lane line c is the lane line to be associated with the detected lane line b and the detected lane line d, respectively. Accordingly, the sum of the lengths L of the multiple lane lines to be associated between each two frames of lane line detection image data is L a +L b , L a +L d , L c +L b and L c +L d ; The server calculates the sum of the distances of multiple lane lines to be associated between every two frames of lane line detection image data based on the multiple discrete points corresponding to each detected lane line; The server determines the multiple lane line association degree scores between every two frames of lane line detection image data based on the sum of the lengths of the multiple lane lines to be associated and the sum of the distances of the multiple lane lines to be associated, and there is a one-to-one correspondence between the sum of the lengths of the lane lines to be associated, the sum of the distances of the lane lines to be associated, and the association degree scores of the lane lines.

[0052] Furthermore, when the server executes the step of calculating the sum of the distances of multiple lane lines to be associated between every two frames of lane line detection image data based on the multiple discrete points corresponding to each detected lane line, in some embodiments, the server obtains the number of discrete points corresponding to each detected lane line, and respectively counts the distances between the multiple discrete points corresponding to each detected lane line between every two frames of lane line detection image data to obtain the multiple discrete point distance values ​​corresponding to each lane line to be associated; the server adds up the multiple discrete point distance values ​​corresponding to each lane line to be associated to obtain the sum of the discrete point distance values ​​corresponding to each lane line to be associated; the server determines the sum of the distances of multiple lane lines to be associated between every two frames of lane line detection image data based on the discrete point distance values ​​corresponding to each lane line to be associated and the number of discrete points corresponding to each detected lane line. For example, Figure 2 As shown, the multiple discrete points corresponding to the detected lane line a are A 1 , A 2 , A 3 , A 4 and A 5 , the multiple discrete points corresponding to the detected lane line b are B 1 , B 2 , B 3 , B 4 and B 5 , the number of discrete points N corresponding to the detected lane line aa is 5, the number of discrete points N corresponding to the detected lane line b b is 5, the server will detect the discrete point A on lane line a 1 As the starting point, the server searches for discrete points A on the detected lane line b. 1 The nearest discrete point B 1 Then the server searches for the discrete point B on the detected lane line a. 1 The nearest discrete point A 3 , and the discrete point B 1 With discrete point A 3 The distance between discrete points A 1 Distance to detected lane line b Similarly, the server will discrete point B 1 With discrete point A 3 The distance between discrete points A 2 Distance to detected lane line b The server will discrete point B 1 With discrete point A 3 The distance between discrete points A 3 Distance to detected lane line b The server will discrete point B 2 With discrete point A 4 The distance between discrete points A 4 Distance to detected lane line b The server will discrete point B 3 With discrete point A 5 The distance between discrete points A 5 Distance to detected lane line b in, and are the distance values ​​of multiple discrete points corresponding to the lane line to be associated. The server traverses all discrete points in the detected lane line a to obtain the distance between the detected lane line a and the detected lane line b: The server traverses all discrete points in the detected lane line b to obtain the distance between the detected lane line b and the detected lane line a. The total distance between lane line detection image data of each two frames to be associated is D = D a +D b The calculation method of the sum of the distances of other lane lines to be associated between every two frames of lane line detection image data is similar, and the details will not be repeated here.

[0053] Furthermore, when the server executes the step of determining multiple lane line association degree scores between every two frames of lane line detection image data based on the sum of the lengths of multiple lane lines to be associated and the sum of the distances of multiple lane lines to be associated, in some embodiments, the server determines whether the sum of the distances of the lane lines to be associated is less than the sum of the lengths of the lane lines to be associated; if the sum of the distances of the lane lines to be associated is less than the sum of the lengths of the lane lines to be associated, the server calculates the target ratio between the sum of the distances of the lane lines to be associated and the sum of the lengths of the lane lines to be associated, the server calculates the target difference between the preset target value and the target ratio, and sets the lane line association degree score between every two frames of lane line detection image data to the target difference; if the sum of the distances of the lane lines to be associated is greater than or equal to the sum of the lengths of the lane lines to be associated, the server sets the lane line association degree score between every two frames of lane line detection image data to the preset score. For example, Figure 2 As shown, the number of discrete points corresponding to the detected lane line c and the detected lane line d are both 5, the lane line length value corresponding to the detected lane line c is 4 meters, and the lane line length value corresponding to the detected lane line d is 4 meters. Then the total length L of the lane lines to be associated corresponding to the detected lane line c and the detected lane line d is 8 meters. When the total distance D of the lane lines to be associated corresponding to the detected lane line c and the detected lane line d is 0.10 meters, the lane line association degree score is 0.988. When the total distance D of the lane lines to be associated corresponding to the detected lane line c and the detected lane line d is 3.00 meters, the lane line association degree score is 0.625. When the total distance D of the lane lines to be associated corresponding to the detected lane line c and the detected lane line d is 10.00 meters, the lane line association degree score is 0.

[0054] 305. Determine the associated lane line between every two frames of lane line detection image data according to the multiple lane line association degree scores.

[0055] That is, the server selects the associated lane lines between every two frames of lane line detection image data from the multiple frames of lane line detection image data according to the lane line association degree score. In some embodiments, when the lane line association degree score is equal to the preset score, the server determines that the detected lane lines corresponding to the lane line association degree score between every two frames of lane line detection image data are not associated, such as Figure 2 As shown in FIG. 1 , the detected lane line a is not associated with the detected lane line d, and the detected lane line b is not associated with the detected lane line c. When the association degree score of each lane line is greater than the preset score, the server sets the detected lane line corresponding to the association degree score of each lane line as the associated lane line between every two frames of lane line detection image data, such as Figure 2As shown, the detected lane line a and the detected lane line b are associated lane lines, and the detected lane line c and the detected lane line d are associated lane lines. The preset score can be set according to actual business needs, for example, the preset score can be 0.

[0056] Furthermore, the server performs multi-frame fusion on the associated lane lines between every two frames of lane line detection image data to obtain multi-frame fused lane line image data; the server generates a local semantic map based on the multi-frame fused lane line image data. The server then improves the accuracy of constructing the semantic map through the associated lane lines between every two frames of lane line detection image data.

[0057] In an embodiment of the present invention, lane line discrete point association processing is performed on multiple frames of lane line detection image data to obtain multiple lane line association degree scores between every two frames of lane line detection image data, and the associated lane lines between every two frames of lane line detection image data are determined based on the multiple lane line association degree scores. When there is an offset distance between the lane lines in different frames and they are relatively misaligned, the scoring accuracy of the associated lane lines between every two frames of lane line detection image data is improved.

[0058] The above describes the lane line correlation degree determination method in the embodiment of the present invention. The following describes the lane line correlation degree determination device in the embodiment of the present invention. Figure 4 In one embodiment of the present invention, a lane line correlation degree determination device includes:

[0059] An acquisition module 401 is used to acquire multiple frames of lane line detection image data, each frame of lane line detection image data includes multiple detected lane lines;

[0060] An association module 402 is used to perform lane line discrete point association processing on the multiple frames of lane line detection image data to obtain multiple lane line association degree scores between every two frames of lane line detection image data;

[0061] The determination module 403 is used to determine the associated lane lines between every two frames of lane line detection image data according to the multiple lane line association degree scores.

[0062] In an embodiment of the present invention, lane line discrete point association processing is performed on multiple frames of lane line detection image data to obtain multiple lane line association degree scores between every two frames of lane line detection image data, and the associated lane lines between every two frames of lane line detection image data are determined based on the multiple lane line association degree scores. When there is an offset distance between the lane lines in different frames and they are relatively misaligned, the scoring accuracy of the associated lane lines between every two frames of lane line detection image data is improved.

[0063] See also Figure 5Another embodiment of the lane line correlation degree determination device in the embodiment of the present invention includes:

[0064] An acquisition module 401 is used to acquire multiple frames of lane line detection image data, each frame of lane line detection image data includes multiple detected lane lines;

[0065] An association module 402 is used to perform lane line discrete point association processing on the multiple frames of lane line detection image data to obtain multiple lane line association degree scores between every two frames of lane line detection image data;

[0066] The determination module 403 is used to determine the associated lane lines between every two frames of lane line detection image data according to the multiple lane line association degree scores.

[0067] Optionally, the acquisition module 401 may also be specifically used for:

[0068] Measuring the length of each detected lane line corresponding to each frame of lane line detection image data to obtain a lane line length value corresponding to each detected lane line;

[0069] Discretely process each detected lane line corresponding to each frame of lane line detection image data to obtain multiple discrete points corresponding to each detected lane line;

[0070] A plurality of lane line association degree scores between every two frames of lane line detection image data are determined according to the lane line length value corresponding to each detected lane line and a plurality of discrete points corresponding to each detected lane line.

[0071] Optionally, the association module 402 may further include:

[0072] The measuring unit 4021 is used to measure the length of each detected lane line corresponding to each frame of lane line detection image data to obtain a lane line length value corresponding to each detected lane line;

[0073] The discrete unit 4022 is used to perform discrete processing on each detected lane line corresponding to each frame of lane line detection image data to obtain a plurality of discrete points corresponding to each detected lane line;

[0074] The determination unit 4023 is used to determine multiple lane line association degree scores between every two frames of lane line detection image data based on the lane line length value corresponding to each detected lane line and the multiple discrete points corresponding to each detected lane line.

[0075] Optionally, the determining unit 4023 may also be specifically configured to:

[0076] An adding subunit 40231 is used to add the lane line length values ​​corresponding to any two detected lane lines between every two frames of lane line detection image data to obtain the sum of the lengths of multiple lane lines to be associated between every two frames of lane line detection image data;

[0077] A calculation subunit 40232 is used to calculate the sum of the distances of multiple lane lines to be associated between every two frames of lane line detection image data according to multiple discrete points corresponding to each detected lane line;

[0078] Determine subunit 40233, for determining multiple lane line association degree scores between every two frames of lane line detection image data based on the sum of the lengths of the multiple lane lines to be associated and the sum of the distances of the multiple lane lines to be associated, and there is a one-to-one correspondence between the sum of the lengths of the lane lines to be associated, the sum of the distances of the lane lines to be associated and the association degree scores of the lane lines.

[0079] Optionally, the determining subunit 40233 may further include:

[0080] Determine whether the sum of the distances of the lane lines to be associated is less than the sum of the lengths of the lane lines to be associated; if the sum of the distances of the lane lines to be associated is less than the sum of the lengths of the lane lines to be associated, calculate a target ratio between the sum of the distances of the lane lines to be associated and the sum of the lengths of the lane lines to be associated, calculate a target difference between a preset target value and the target ratio, and set the score of the degree of association of each lane line between every two frames of lane line detection image data as the target difference;

[0081] If the sum of the distances of the lane lines to be associated is greater than or equal to the sum of the lengths of the lane lines to be associated, the score of the degree of association of each lane line between every two frames of lane line detection image data is set to a preset score.

[0082] Optionally, the determining module 403 may also be specifically configured to:

[0083] When the lane line association degree score is equal to the preset score, it is determined that the detected lane lines corresponding to the lane line association degree scores between every two frames of lane line detection image data are not associated;

[0084] When the association degree score of each lane line is greater than a preset score, the detected lane line corresponding to the association degree score of each lane line is set as the associated lane line between every two frames of lane line detection image data.

[0085] Optionally, the lane line association degree determination device further includes:

[0086] A fusion module 404 is used to perform multi-frame fusion on the associated lane lines between every two frames of lane line detection image data to obtain multi-frame fused lane line image data;

[0087] The generating module 405 is used to generate a local semantic map according to the multi-frame fused lane line image data.

[0088] In an embodiment of the present invention, lane line discrete point association processing is performed on multiple frames of lane line detection image data to obtain multiple lane line association degree scores between every two frames of lane line detection image data, and the associated lane lines between every two frames of lane line detection image data are determined based on the multiple lane line association degree scores. When there is an offset distance between the lane lines in different frames and they are relatively misaligned, the scoring accuracy of the associated lane lines between every two frames of lane line detection image data is improved.

[0089] above Figure 4 and Figure 5 The lane line correlation degree determination device in the embodiment of the present invention is described in detail from the perspective of modularization, and the lane line correlation degree determination device in the embodiment of the present invention is described in detail from the perspective of hardware processing.

[0090] Figure 6 is a schematic diagram of the structure of a lane line association degree determination device provided by an embodiment of the present invention. The lane line association degree determination device 600 may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) 610 (for example, one or more processors) and a memory 620, and one or more storage media 630 (for example, one or more mass storage devices) storing application programs 633 or data 632. Among them, the memory 620 and the storage medium 630 can be short-term storage or permanent storage. The program stored in the storage medium 630 may include one or more modules (not shown in the figure), and each module may include a series of computer program operations in the lane line association degree determination device 600. Furthermore, the processor 610 may be configured to communicate with the storage medium 630 to execute a series of computer program operations in the storage medium 630 on the lane line association degree determination device 600.

[0091] The lane line correlation degree determination device 600 may also include one or more power supplies 640, one or more wired or wireless network interfaces 650, one or more input and output interfaces 660, and / or one or more operating systems 631, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, etc. It will be appreciated by those skilled in the art that Figure 6 The structure of the lane line correlation degree determination device shown does not constitute a limitation on the lane line correlation degree determination device, and may include more or fewer components than shown in the figure, or a combination of certain components, or a different arrangement of components.

[0092] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions are executed on a computer, the computer executes the steps of the lane line association degree determination method.

[0093] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0094] If the integrated unit is implemented in the form of a software functional unit 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. The computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program codes.

[0095] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for determining the degree of association of lane lines, characterized in that: The lane line association degree determination method comprises: Acquire multiple frames of lane line detection image data, each frame of lane line detection image data includes multiple detected lane lines; Measuring the length of each detected lane line corresponding to each frame of lane line detection image data to obtain a lane line length value corresponding to each detected lane line; Discretely process each detected lane line corresponding to each frame of lane line detection image data to obtain multiple discrete points corresponding to each detected lane line; Determine multiple lane line association degree scores between every two frames of lane line detection image data according to the lane line length value corresponding to each detected lane line and multiple discrete points corresponding to each detected lane line; Determine the associated lane lines between every two frames of lane line detection image data according to the multiple lane line association degree scores.

2. The method for determining the lane line association degree according to claim 1, characterized in that: The acquiring of multiple frames of lane line detection image data includes: Receiving road environment image data sent by a preset sensor; The road environment image data is subjected to target detection through a preset perception detection model to obtain multiple frames of lane line detection image data.

3. The method for determining the lane line association degree according to claim 1, characterized in that: The determining of a plurality of lane line association degree scores between every two frames of lane line detection image data according to the lane line length value corresponding to each detected lane line and the plurality of discrete points corresponding to each detected lane line comprises: Adding the lane line length values ​​corresponding to any two detected lane lines between every two frames of lane line detection image data to obtain the sum of the lengths of multiple lane lines to be associated between every two frames of lane line detection image data; Calculate the sum of the distances of multiple lane lines to be associated between each two frames of lane line detection image data according to multiple discrete points corresponding to each detected lane line; Multiple lane line association degree scores between every two frames of lane line detection image data are determined based on the sum of the lengths of the multiple lane lines to be associated and the sum of the distances of the multiple lane lines to be associated, and there is a one-to-one correspondence between the sum of the lengths of the lane lines to be associated, the sum of the distances of the lane lines to be associated, and the association degree scores of the lane lines.

4. The method for determining the lane line association degree according to claim 3, characterized in that: The determining of the multiple lane line association degree scores between every two frames of lane line detection image data according to the sum of the lengths of the multiple lane lines to be associated and the sum of the distances of the multiple lane lines to be associated comprises: Determine whether the sum of the distances of the lane lines to be associated is less than the sum of the lengths of the lane lines to be associated; If the sum of the distances of the lane lines to be associated is less than the sum of the lengths of the lane lines to be associated, a target ratio between the sum of the distances of the lane lines to be associated and the sum of the lengths of the lane lines to be associated is calculated, a target difference between a preset target value and the target ratio is calculated, and a score of the degree of association of each lane line between every two frames of lane line detection image data is set as the target difference; If the sum of the distances of the lane lines to be associated is greater than or equal to the sum of the lengths of the lane lines to be associated, the score of the degree of association of each lane line between every two frames of lane line detection image data is set to a preset score.

5. The method for determining the lane line association degree according to claim 1, characterized in that: The determining the associated lane line between every two frames of lane line detection image data according to the multiple lane line association degree scores includes: When the lane line association degree score is equal to the preset score, it is determined that the detected lane lines corresponding to the lane line association degree scores between every two frames of lane line detection image data are not associated; When the association degree score of each lane line is greater than a preset score, the detected lane line corresponding to the association degree score of each lane line is set as the associated lane line between every two frames of lane line detection image data.

6. The method for determining the lane line association degree according to any one of claims 1 to 5, characterized in that: After determining the associated lane lines between every two frames of lane line detection image data according to the multiple lane line association degree scores, the lane line association degree determination method further includes: Perform multi-frame fusion on the associated lane lines between every two frames of lane line detection image data to obtain multi-frame fused lane line image data; A local semantic map is generated based on the multi-frame fused lane line image data.

7. A lane line correlation degree determination device, characterized in that: The lane line association degree determination device comprises: An acquisition module is used to acquire multiple frames of lane line detection image data, each frame of lane line detection image data includes multiple detected lane lines; An association module is used to measure the length of each detected lane line corresponding to each frame of lane line detection image data to obtain a lane line length value corresponding to each detected lane line; Discretely process each detected lane line corresponding to each frame of lane line detection image data to obtain multiple discrete points corresponding to each detected lane line; Determine multiple lane line association degree scores between every two frames of lane line detection image data according to the lane line length value corresponding to each detected lane line and multiple discrete points corresponding to each detected lane line; A determination module is used to determine the associated lane lines between every two frames of lane line detection image data according to the multiple lane line association degree scores.

8. A lane line correlation degree determination device, characterized in that: The lane line association degree determination device comprises: a memory and at least one processor, wherein a computer program is stored in the memory; The at least one processor calls the computer program in the memory so that the lane line association degree determination device executes the lane line association degree determination method as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for determining the degree of lane line association as described in any one of claims 1 to 6 is implemented.

Images