High-definition map double-yellow-line post-processing method and device, and electronic equipment

Abstract

The application relates to a high-precision map double-yellow-line post-processing method and device and electronic equipment. The method comprises the following steps: obtaining a semantic segmentation result of high-precision map data, and extracting an initial contour set from the semantic segmentation result; the initial contour set comprises candidate contours corresponding to double yellow lines; the shortest Euclidean distances between all candidate contours in the initial contour set are calculated, and according to the distance calculation result and a preset adhesion threshold, a pair of adhesion contours is determined from the candidate contours; for each contour in the pair of adhesion contours, the corresponding two contours in the pair of adhesion contours are translated by a preset translation amount in the normal direction of the contour, and a separated contour set is obtained; and a double-yellow-line geometric element is generated according to the separated contour set. The scheme provided by the application can effectively solve the misrecognition and missed detection problems of double yellow lines caused by pixel adhesion in complex scenes such as bad weather, improve the double yellow line recognition accuracy and data integrity, and further improve the production accuracy of the high-precision map.

Description

Technical Field

[0001] This application relates to the field of high-precision map technology, and in particular to a method, apparatus and electronic device for post-processing double yellow lines in high-precision maps. Background Technology

[0002] In the process of creating and updating high-precision maps, double yellow lines are an important road marking element. Their accurate identification and geometric reconstruction are crucial for the lane keeping, path planning and other functions of autonomous driving systems.

[0003] In related technologies, methods based on color segmentation or edge detection are typically used to extract double yellow line contours from road images used to construct high-precision maps. A second yellow line is then generated directly from these extracted contours using a fixed displacement method, thus completing the generation of the double yellow line geometric elements. However, in adverse weather conditions (such as rain or fog), environmental interference factors such as shadows and reflections can cause the two yellow lines to be misidentified as a single contour due to pixel adhesion, leading to missed detections of the double yellow lines. This results in missing geometric elements in the generated double yellow lines, affecting the accuracy of high-precision map production. Summary of the Invention

[0004] To address or partially address the problems existing in related technologies, this application provides a high-precision map double yellow line post-processing method, apparatus, and electronic device, which can effectively solve the problem of misidentification and missed detection of double yellow lines caused by pixel adhesion in complex scenarios such as severe weather, improve the double yellow line recognition accuracy and data integrity, and thus improve the production accuracy of high-precision maps.

[0005] The first aspect of this application provides a post-processing method for double yellow lines on high-precision maps, including: Obtain the semantic segmentation results of high-precision map data, and extract an initial contour set from the semantic segmentation results; wherein, the initial contour set includes candidate contours corresponding to double yellow lines; Calculate the shortest Euclidean distance between all candidate contours in the initial contour set, and determine the adhesive contour pairs from the candidate contours based on the distance calculation results and a preset adhesion threshold. For each contour in the adhesive contour pair, according to the normal direction of its contour, the corresponding two contours in the adhesive contour pair are translated in opposite directions along their respective normal directions by a preset translation amount to obtain a set of separated contours. Based on the separated contour set, generate double yellow line geometric elements.

[0006] In some implementations, the normal direction is determined in the following way: The target point is determined from the contour points of the contour, and the M adjacent points before and after the target point are selected to form a local point set; where M is an integer greater than 1. By fitting a straight line or curve to the local point set, the tangent direction at the target point is obtained; Rotate the tangent direction by 90 degrees to obtain the normal direction of the contour.

[0007] In some implementations, generating the double yellow line geometry based on the separated contour set includes: Obtain road curvature information, and perform curvature adaptive translation processing on the separated contour set based on the road curvature information to generate double yellow line geometric elements.

[0008] In some implementations, the step of acquiring road curvature information and performing curvature-adaptive translation processing on the separated contour set based on the road curvature information to generate double yellow line geometric features includes: Obtain road curvature information, and calculate the compensation translation amount based on the road curvature information and the preset dynamic translation compensation rules; The translation direction is determined based on the lane direction field. The separated contour set is translated according to the translation direction and the compensation translation amount to generate double yellow line geometric elements. The translation direction is the normal direction of the direction vector at the corresponding point in the lane direction field.

[0009] In some implementations, the method further includes: The geometric elements of the double yellow lines are subjected to topological verification. If the topological verification result does not meet the preset conditions, the geometric elements of the double yellow lines are automatically geometrically corrected, and the reconstructed double yellow line data is output.

[0010] In some implementations, the topology verification includes verifying the parallelism between the double yellow line geometry and the road boundary.

[0011] In some implementations, the method further includes: The geometric elements of the double yellow lines are subjected to fracture interval identification. When a fracture interval is determined to exist, the interval data corresponding to the fracture interval is curve fitted according to a preset reconstruction rule to generate a reconstruction curve that connects the fracture interval.

[0012] A second aspect of this application provides a high-precision map double yellow line post-processing device, comprising: An initial contour acquisition module is used to acquire the semantic segmentation results of high-precision map data and extract an initial contour set from the semantic segmentation results; wherein, the initial contour set includes candidate contours corresponding to double yellow lines; The adhesion contour determination module is used to calculate the shortest Euclidean distance between all candidate contours in the initial contour set, and determine the adhesion contour pairs from the candidate contours based on the distance calculation results and a preset adhesion threshold. The contour separation module is used to translate the corresponding two contours in the adhesive contour pair in opposite directions along their respective normal directions by a preset translation amount according to the normal direction of their contour points, so as to obtain a set of separated contours. The feature generation module is used to generate double yellow line geometric features based on the separated contour set.

[0013] A third aspect of this application provides an electronic device, comprising: Processor; and A memory that stores executable code, which, when executed by the processor, causes the processor to perform the method described above.

[0014] A fourth aspect of this application provides a computer-readable storage medium having executable code stored thereon, which, when executed by a processor of an electronic device, causes the processor to perform the method described above.

[0015] The technical solution provided in this application may include the following beneficial effects: The technical solution of this application calculates the shortest Euclidean distance between all candidate contours in an initial contour set and compares it with an adhesion threshold. Based on the normal direction of each contour, it performs translational separation of the adhered contour pairs in opposite directions. Finally, it generates double yellow line geometric elements based on this precisely separated contour set. This technical solution can automatically and accurately identify double yellow line adhered contour pairs from adhered images generated in complex environments and precisely separate these pairs. It can effectively solve the problem of misidentification and missed detection of double yellow lines caused by pixel adhesion in complex scenarios such as severe weather, effectively improving the accuracy of double yellow line recognition and data integrity, thereby improving the accuracy of high-precision map production.

[0016] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0017] The above and other objects, features and advantages of this application will become more apparent from the more detailed description of exemplary embodiments thereof in conjunction with the accompanying drawings, wherein the same reference numerals generally represent the same components in the exemplary embodiments thereof.

[0018] Figure 1 This is a schematic flowchart illustrating the high-precision map double yellow line post-processing method shown in the embodiments of this application; Figure 2 This is another schematic flowchart illustrating the high-precision map double yellow line post-processing method shown in the embodiments of this application; Figure 3 This is a schematic diagram of the high-precision map double yellow line post-processing device shown in the embodiments of this application; Figure 4 This is another schematic diagram of the high-precision map double yellow line post-processing device shown in the embodiments of this application; Figure 5 This is a schematic diagram of the structure of an electronic device shown in an embodiment of this application. Detailed Implementation

[0019] Embodiments of this application will now be described in more detail with reference to the accompanying drawings. While embodiments of this application are shown in the drawings, it should be understood that this application may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to make this application more thorough and complete, and to fully convey the scope of this application to those skilled in the art.

[0020] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0021] It should be understood that although the terms "first," "second," "third," etc., may be used in this application to describe various information, this information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0022] In related technologies, methods based on color segmentation or edge detection are typically used to extract the outline of the double yellow lines. A second yellow line is then generated directly from the extracted outline using a fixed displacement method, thus completing the generation of the double yellow line geometric elements. However, in adverse weather conditions (such as rain or fog), due to environmental interference factors like shadows and reflections, the two yellow lines are easily misidentified as a single outline in the image due to pixel adhesion. This leads to missed detections of the double yellow lines, resulting in missing geometric elements in the generated double yellow lines and affecting the accuracy of high-precision map production.

[0023] To address the aforementioned issues, this application provides a post-processing method for double yellow lines in high-precision maps. This method effectively solves the problems of misidentification and missed detection of double yellow lines caused by pixel adhesion in complex scenarios such as severe weather, improves the recognition accuracy and data integrity of double yellow lines, and thus enhances the production accuracy of high-precision maps.

[0024] The technical solutions of the embodiments of this application are described in detail below with reference to the accompanying drawings.

[0025] Figure 1 This is a schematic flowchart illustrating the high-precision map double yellow line post-processing method shown in the embodiments of this application.

[0026] See Figure 1 The high-precision map double yellow line post-processing method of this application includes: S110, Obtain the semantic segmentation results of the high-precision map data, and extract the initial contour set from the semantic segmentation results; wherein, the initial contour set includes candidate contours corresponding to the double yellow lines.

[0027] In this step, the initial contours corresponding to the double yellow lines are extracted from the semantic segmentation results of the road images used to construct the high-precision map.

[0028] The semantic segmentation result can be obtained by performing pixel-level classification on road image data used to construct high-precision maps using a deep learning model. Essentially, the semantic segmentation result corresponds to extracting semantic category label data from the high-precision map image source. For example, a convolutional neural network can be used to process high-precision map data to obtain the segmentation mask for road marking areas.

[0029] In this process, an edge detection algorithm can be used to identify the segmentation mask corresponding to the double yellow lines, thereby identifying closed boundary lines from the semantic segmentation results and forming an initial contour set containing candidate contours of the double yellow lines. This initial contour set can contain candidate regions of double yellow lines that have been misidentified as single contours.

[0030] S120: Calculate the shortest Euclidean distance between all candidate contours in the initial contour set, and determine the adhesive contour pairs from the candidate contours based on the distance calculation results and the preset adhesion threshold.

[0031] In this step, the shortest Euclidean distance between all candidate contours in the initial contour set is calculated. Based on the calculation results and a preset adhesion threshold for determining whether adhesion exists, the adhesion contour pairs are determined from all candidate contours.

[0032] It should be understood that an adhesion profile pair refers to two yellow lines that are adhered together but are presented as a single profile on the bottom of the plate.

[0033] One approach is to use a spatial indexing structure to calculate the minimum spatial distance between two candidate contours. For example, a KD-tree can be used to calculate the shortest Euclidean distance between two candidate contours.

[0034] S130, for each contour in the adhesive contour pair, according to the normal direction of its contour points, the corresponding two contours in the adhesive contour pair are translated in opposite directions along their respective normal directions by a preset translation amount to obtain the separated contour set.

[0035] In this step, after determining the adhesive contour pair, based on the normal directions of the two contours in the adhesive contour pair, the two contours in the adhesive contour pair are translated in opposite directions by a preset translation amount to obtain the separation contour set.

[0036] It should be understood that the separated contour set includes contours separated from the adhered contour pairs as well as candidate contours that are not adhered to each other in the initial contour set.

[0037] In some implementations, the normal direction can be determined by the following steps: S131, determine the target point from the contour points of the contour, and select M adjacent points before and after the target point to form a local point set; where M is an integer greater than 1.

[0038] The target point can refer to a reference positioning point on the contour used to calculate the normal direction. The target point can be determined using uniform sampling or keypoint detection. The local point set can be formed using a sliding window mechanism.

[0039] S132, perform straight line or curve fitting on the local point set to obtain the tangent direction at the target point.

[0040] The process of fitting a straight line or curve can use the least squares method to fit a straight line or spline curve, thereby performing geometric smoothing on the local point set.

[0041] S133, rotate the tangent direction by 90 degrees to obtain the normal direction of the contour.

[0042] Rotating the tangent direction by 90 degrees can be a geometric transformation of the tangent direction to obtain the normal direction in the two-dimensional plane that is perpendicular to the tangent direction.

[0043] By determining the normal direction through the above steps, the normal direction can be made dependent on the smoothed tangent, effectively ensuring the perpendicularity and stability of the normal direction. This provides precise guidance for the separation and translation process of the adhered contour pairs, achieving reliable separation of the double yellow line contours.

[0044] S140 generates double yellow line geometric elements based on the separated contour set.

[0045] In this step, all contours are collected based on the obtained separated contour sets to construct the complete geometric elements of the double yellow line road markings.

[0046] In this embodiment, the high-precision map double yellow line post-processing method of this application calculates the shortest Euclidean distance between all candidate contours in the initial contour set and compares it with the adhesion threshold. Based on the normal direction of each contour, the adhered contour pairs are separated by translation in opposite directions. Finally, the double yellow line geometric elements are generated based on this precisely separated contour set. The technical solution of this application can automatically and accurately identify double yellow line adhered contour pairs from adhered images generated in complex environments and accurately separate the adhered contour pairs. This effectively solves the problem of misidentification and missed detection of double yellow lines caused by pixel adhesion in complex scenarios such as severe weather, effectively improving the double yellow line recognition accuracy and data integrity, thereby improving the production accuracy of high-precision maps.

[0047] Figure 2 This is another schematic diagram of the post-processing method for double yellow lines on high-precision maps shown in the embodiments of this application.

[0048] See Figure 2 The high-precision map double yellow line post-processing method of this application includes: S210, Obtain the semantic segmentation results of the high-precision map data, and extract the initial contour set from the semantic segmentation results; wherein, the initial contour set includes candidate contours corresponding to the double yellow lines.

[0049] In this step, the initial contours corresponding to the double yellow lines are extracted from the semantic segmentation results of the road images used to construct the high-precision map.

[0050] S220: Calculate the shortest Euclidean distance between all candidate contours in the initial contour set, and determine the adhesive contour pairs from the candidate contours based on the distance calculation results and the preset adhesion threshold.

[0051] In this step, the shortest Euclidean distance between all candidate contours in the initial contour set is calculated. Based on the calculation results and a preset adhesion threshold for determining whether adhesion exists, the adhesion contour pairs are determined from all candidate contours.

[0052] S230, for each contour in the adhesive contour pair, according to the normal direction of its contour points, the corresponding two contours in the adhesive contour pair are translated in opposite directions along their respective normal directions by a preset translation amount to obtain the separated contour set.

[0053] In this step, after determining the adhesive contour pair, based on the normal directions of the two contours in the adhesive contour pair, the two contours in the adhesive contour pair are translated in opposite directions by a preset translation amount to obtain the separation contour set.

[0054] S240: Obtain road curvature information, and perform curvature adaptive translation processing on the separated contour set based on the road curvature information to generate double yellow line geometric elements.

[0055] In this step, the road curvature information corresponding to the initial contour set is obtained, the corresponding compensation translation amount is calculated based on the road curvature information, and the curvature adaptive translation processing of the separated contour set is performed using the compensation translation amount to generate the double yellow line geometric elements.

[0056] Road curvature information refers to quantitative parameters that describe the degree of geometric curvature of a road. Specifically, road curvature information can be characterized using the radius of curvature or curvature value of the curve fitted to the road's centerline.

[0057] It should be understood that road curvature information can be obtained from high-precision map data.

[0058] By performing curvature-adaptive translation processing on the separated contour set based on road curvature information, the translation amount of the generated double yellow line geometric elements can be adaptively adjusted according to the degree of road curvature. This avoids systematic deviations or conflicts between the double yellow line geometric elements and the actual road marking positions in curved areas due to fixed translation amounts.

[0059] In some implementations, obtaining road curvature information and performing curvature-adaptive translation processing on the separated contour set based on the road curvature information to generate double yellow line geometric features may include the following steps: S241, Obtain road curvature information, and calculate the compensation translation amount based on the road curvature information and the preset dynamic translation compensation rules.

[0060] Among them, the preset dynamic translation compensation rule can characterize the mapping relationship between road curvature and translation compensation value.

[0061] The preset dynamic translation compensation rules may include calculating the translation compensation amount for each contour in the separated contour set using the following formula: dy=dynamic_base_dy×(1+α×curvature); Where curvature is the road curvature, dynamic_base_dy is the pre-set dynamic baseline spacing that changes with position, α is the curvature compensation coefficient, and the translation compensation amount for each contour point is calculated.

[0062] The translation compensation amount for each contour is calculated using the above formula. This compensation translation amount is then applied to the translation operation of the separated contour set, thereby ensuring that the generated double yellow line geometric elements can accurately match the spatial distribution of the actual road markings in both straight and curved areas.

[0063] S242, determine the translation direction based on the lane direction field, and perform translation processing on the separated contour set according to the translation direction and the compensation translation amount to generate double yellow line geometric elements; wherein, the translation direction is the normal direction of the direction vector at the corresponding point in the lane direction field.

[0064] The lane orientation field can refer to a vector field representing the local orientation of the road. This field can be a tangent vector field generated based on the road centerline or an orientation field extracted from image features. It can be understood that defining the translation direction as the normal direction of the orientation vector at the corresponding point in the lane orientation field effectively ensures that the translation direction is perpendicular to the road centerline.

[0065] By performing translation processing on the separated contour set in the above manner, in areas with large or abrupt road curvature, the translation amount can be dynamically adjusted according to the road curvature and the translation direction can be precisely controlled. This effectively avoids the positional offset of the geometric elements of the double yellow lines, ensuring that the double yellow lines remain parallel to the road boundary, thereby improving the geometric reconstruction accuracy of high-precision maps in complex road scenarios.

[0066] S250 performs topology verification on the geometric features of the double yellow lines. If the topology verification result does not meet the preset conditions, it automatically corrects the geometric features of the double yellow lines and outputs the reconstructed double yellow line data.

[0067] In this step, a topology check is performed on all double yellow line geometric elements before outputting them. If the topology check result of a double yellow line geometric element does not meet the preset conditions, the double yellow line geometric element is automatically geometrically corrected, and the reconstructed double yellow line data is output.

[0068] It should be understood that when the topology verification results of the double yellow line geometric elements meet the preset conditions, there is no need to perform automatic geometric correction.

[0069] Topology verification refers to the process of systematically verifying the geometric elements of double yellow lines based on road topology rules. Preset conditions can refer to pre-defined thresholds for topology consistency judgment. For example, a pre-defined parallelism deviation tolerance, such as greater than 95%.

[0070] Topology verification may include verifying the parallelism between the geometric features of the double yellow lines and the road boundary.

[0071] Topology verification can be achieved using the following steps: S251 calculates the parallelism between the geometric elements of the double yellow lines and the road boundary. When the parallelism does not meet the preset parallelism threshold, it automatically corrects the geometric elements of the double yellow lines and outputs the reconstructed double yellow line data.

[0072] Parallelism refers to a quantitative indicator of the directional consistency between the geometric elements of the double yellow lines and the road boundary, used to assess whether the double yellow lines conform to the road boundary orientation. Automatic geometric correction refers to the process of dynamically adjusting the geometric shape of the double yellow lines based on parallelism deviations. For example, parametric curve fitting can be used to realign the geometric elements of double yellow lines whose parallelism does not meet a preset parallelism threshold with the road boundary, thereby eliminating positional deviations.

[0073] It is understandable that the parallelism is calculated based on the geometric features of the double yellow lines and the road boundary using the above method. This parallelism serves as a criterion for determining positional accuracy. When the parallelism falls below a preset parallelism threshold, a correction mechanism is automatically triggered, and the double yellow lines are geometrically adjusted according to the deviation. This closed-loop feedback process effectively ensures that the geometric features of the double yellow lines maintain a reasonable parallel relationship with the road boundary, thereby maintaining the continuity and accuracy of the double yellow line geometric features even in curved road scenarios.

[0074] Furthermore, the parallelism between the geometric features of the double yellow lines and the road boundary can be calculated in the following ways: S2511, the geometric elements of the double yellow lines and the adjacent road boundaries are discretized into sampling point sets respectively.

[0075] One approach is to use equal-interval sampling to convert the geometric elements of the double yellow lines and the continuous geometric curves of the adjacent road boundaries into a discrete point sequence.

[0076] S2512, based on the sampling point set, calculate the shortest distance line from each sampling point on the double yellow line to the road boundary, and obtain the angle between the line and the tangent direction at the location of the geometric element of the double yellow line.

[0077] S2513, calculate the average angle deviation based on the included angle of all sampling points, and determine the parallelism between the geometric elements of the double yellow lines and the road boundary based on the average angle deviation.

[0078] The average angle deviation can be calculated by using the arithmetic mean or a weighted average based on the sampling point weights to calculate the angle between all sampling points.

[0079] By calculating the parallelism between the geometric elements of the double yellow lines and the road boundary in the above manner, the local parallelism of areas containing curved roads can be accurately assessed, thereby avoiding assessment distortion caused by changes in road curvature and reducing misjudgments or omissions in geometric correction during topology verification.

[0080] S260 identifies the fracture intervals of the geometric elements of the double yellow lines. When a fracture interval is determined, curve fitting is performed on the interval data corresponding to the fracture interval according to the preset reconstruction rules to generate a reconstruction curve that connects the fracture intervals.

[0081] In this step, before outputting the geometric features of the double yellow lines, the break intervals of all the geometric features of the double yellow lines are identified. When a break interval is determined, the interval data corresponding to the break interval is curve-fitted according to the preset reconstruction rules to generate a reconstruction curve, so as to connect the break intervals by generating the reconstruction curve.

[0082] The process of identifying fracture zones can detect discontinuous regions by analyzing the topological continuity of the double yellow line geometric elements, thereby enabling the identification of fracture zones. For example, methods based on point sequence distance threshold detection or geometric connectivity analysis can be used to detect fracture locations in all double yellow line geometric elements.

[0083] The preset reconstruction rules can be predefined curve fitting rules. For example, the least squares method can be used to fit a polynomial curve to generate a reconstructed curve that naturally connects the broken intervals.

[0084] Among these methods, curve fitting of the interval data corresponding to the fracture interval according to the preset reconstruction rules can include: Effective contour points of a preset length on both sides of the fracture interval are extracted as fitting data. Based on the fitting data, a reconstruction curve is generated through curve fitting.

[0085] It is understandable that by extracting effective contour points within a preset length range from both sides of the fractured section as fitting data, the geometric continuity information of the complete contours on both sides of the fracture can be utilized. Based on these fitting data, a curve fitting operation can be performed, and the generated reconstructed curve can ensure the directional consistency of the contours on both sides of the fracture, thereby forming a smooth transition curve at the fractured section.

[0086] In this application, steps S250 and S260 can be performed separately and sequentially, or only one step can be selected for performance. That is to say, in the double yellow line post-processing method of this application, after obtaining the geometric elements of the double yellow line, steps S250 and S260 can be performed sequentially, or only step S250 can be performed, or only step S260 can be performed.

[0087] In this embodiment, the double yellow line post-processing method of this application, by introducing road curvature information for adaptive translation processing, can effectively overcome the geometric distortion caused by using only fixed displacement in road sections with changing curvature, thereby effectively ensuring that the spacing of the generated double yellow line geometric elements is highly compatible with the road direction; furthermore, by performing topology verification and automatic correction on the generated elements, it can detect and correct the parallelism deviation between the double yellow line geometric elements and the road boundary in real time, significantly improving the geometric consistency in complex road topologies; furthermore, the double yellow line post-processing method of this application also sets up a break interval identification and fitting reconstruction mechanism, which can automatically repair the discontinuity of the marking caused by occlusion or wear, effectively ensuring the topological integrity and continuity of the output double yellow line geometric elements.

[0088] Corresponding to the aforementioned application function implementation method embodiments, this application also provides a high-precision map double yellow line post-processing device, electronic device, and corresponding embodiments.

[0089] Figure 3 This is a schematic diagram of the structure of the high-precision map double yellow line post-processing device shown in the embodiments of this application.

[0090] See Figure 3 The high-precision map double yellow line post-processing device 300 of this application includes: an initial contour acquisition module 310, an adhesive contour determination module 320, a contour separation module 330, and a feature generation module 340.

[0091] The initial contour acquisition module 310 is used to acquire the semantic segmentation results of high-precision map data and extract an initial contour set from the semantic segmentation results; wherein, the initial contour set includes candidate contours corresponding to the double yellow lines.

[0092] The adhesion contour determination module 320 is used to calculate the shortest Euclidean distance between all candidate contours in the initial contour set, and to determine the adhesion contour pairs from the candidate contours based on the distance calculation results and the preset adhesion threshold.

[0093] The contour separation module 330 is used to translate the corresponding two contours in the adhesive contour pair in opposite directions along their respective normal directions by a preset translation amount according to the normal direction of their contour points, so as to obtain a set of separated contours.

[0094] The normal direction is determined as follows: the target point is determined from the contour points of the contour, and M adjacent points before and after the target point are selected to form a local point set; where M is an integer greater than 1; the local point set is fitted with a straight line or curve to obtain the tangent direction at the target point; the tangent direction is rotated by 90 degrees to obtain the normal direction of the contour.

[0095] Feature generation module 340 is used to generate double yellow line geometric features based on the separated contour set.

[0096] The element generation module 340 can also acquire road curvature information and perform curvature adaptive translation processing on the separated contour set based on the road curvature information to generate double yellow line geometric elements.

[0097] The element generation module 340 can also acquire road curvature information, calculate the compensation translation amount based on the road curvature information and the preset dynamic translation compensation rules; determine the translation direction based on the lane direction field, and perform translation processing on the separated contour set based on the translation direction and the compensation translation amount to generate double yellow line geometric elements; wherein, the translation direction is the normal direction of the direction vector at the corresponding point in the lane direction field.

[0098] Figure 4 This is another schematic diagram of the high-precision map double yellow line post-processing device shown in the embodiments of this application.

[0099] See Figure 3 The high-precision map double yellow line post-processing device 300 of this application includes: an initial contour acquisition module 310, an adhesion contour determination module 320, a contour separation module 330, a feature generation module 330, a topology verification module 340, and a fracture interval identification module 350.

[0100] The topology verification module 340 is used to perform topology verification on the geometric features of the double yellow lines. When the topology verification result does not meet the preset conditions, the geometric features of the double yellow lines are automatically geometrically corrected and the reconstructed double yellow line data is output.

[0101] The fracture interval identification module 350 is used to identify fracture intervals of the geometric elements of the double yellow lines. When a fracture interval is determined to exist, the module performs curve fitting on the interval data corresponding to the fracture interval according to the preset reconstruction rules to generate a reconstruction curve that connects the fracture intervals.

[0102] In this embodiment, the high-precision map double yellow line post-processing device of this application calculates the shortest Euclidean distance between all candidate contours in the initial contour set and compares it with the adhesion threshold. Based on the normal direction of each contour, it performs translational separation of the adhered contour pairs in opposite directions, and finally generates double yellow line geometric elements based on the precisely separated contour set. The technical solution of this application can automatically and accurately identify double yellow line adhered contour pairs from adhered images generated in complex environments and accurately separate the adhered contour pairs. This effectively solves the problem of misidentification and missed detection of double yellow lines caused by pixel adhesion in complex scenarios such as severe weather, effectively improving the double yellow line recognition accuracy and data integrity, thereby improving the production accuracy of high-precision maps.

[0103] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated further here.

[0104] Figure 5 This is a schematic diagram of the structure of an electronic device shown in an embodiment of this application.

[0105] See Figure 5 The electronic device 1000 includes a memory 1010 and a processor 1020.

[0106] The processor 1020 can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0107] Memory 1010 may include various types of storage units, such as system memory, read-only memory (ROM), and permanent storage devices. ROM may store static data or instructions required by processor 1020 or other modules of the computer. Permanent storage devices may be read-write storage devices. Permanent storage devices may be non-volatile storage devices that retain stored instructions and data even when the computer is powered off. In some embodiments, permanent storage devices use mass storage devices (e.g., magnetic or optical disks, flash memory) as permanent storage devices. In other embodiments, permanent storage devices may be removable storage devices (e.g., floppy disks, optical drives). System memory may be a read-write storage device or a volatile read-write storage device, such as dynamic random access memory. System memory may store some or all of the instructions and data required by the processor during operation. Furthermore, memory 1010 may include any combination of computer-readable storage media, including various types of semiconductor memory chips (e.g., DRAM, SRAM, SDRAM, flash memory, programmable read-only memory), and disks and / or optical disks may also be used. In some embodiments, the memory 1010 may include a removable storage device that is readable and / or writable, such as a laser disc (CD), a read-only digital multifunction optical disc (e.g., DVD-ROM, dual-layer DVD-ROM), a read-only Blu-ray disc, an ultra-high density optical disc, a flash memory card (e.g., SD card, mini SD card, Micro-SD card, etc.), a magnetic floppy disk, etc. Computer-readable storage media do not contain carrier waves or transient electronic signals transmitted wirelessly or via wired connections.

[0108] The memory 1010 stores executable code, which, when processed by the processor 1020, can cause the processor 1020 to execute part or all of the methods described above.

[0109] Furthermore, the method according to this application can also be implemented as a computer program or computer program product, which includes computer program code instructions for performing some or all of the steps in the method described above.

[0110] Alternatively, this application may be implemented as a computer-readable storage medium (or a non-transitory machine-readable storage medium or a machine-readable storage medium) storing executable code (or computer program or computer instruction code) thereon, which, when executed by a processor of an electronic device (or server, etc.), causes the processor to perform part or all of the steps of the methods described above according to this application.

[0111] The various embodiments of this application have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or improvement of the technology in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A post-processing method for double yellow lines on a high-precision map, characterized in that, include: Obtain the semantic segmentation results of high-precision map data, and extract an initial contour set from the semantic segmentation results; wherein, the initial contour set includes candidate contours corresponding to double yellow lines; Calculate the shortest Euclidean distance between all candidate contours in the initial contour set, and determine the adhesive contour pairs from the candidate contours based on the distance calculation results and a preset adhesion threshold. For each contour in the adhesive contour pair, according to the normal direction of its contour, the corresponding two contours in the adhesive contour pair are translated in opposite directions along their respective normal directions by a preset translation amount to obtain a set of separated contours. Based on the separated contour set, generate double yellow line geometric elements.

2. The method according to claim 1, characterized in that, The direction of the normal is determined in the following way: The target point is determined from the contour points of the contour, and the M adjacent points before and after the target point are selected to form a local point set; where M is an integer greater than 1. By fitting a straight line or curve to the local point set, the tangent direction at the target point is obtained; Rotate the tangent direction by 90 degrees to obtain the normal direction of the contour.

3. The method according to claim 1, characterized in that, The generation of double yellow line geometric features based on the separated contour set includes: Obtain road curvature information, and perform curvature adaptive translation processing on the separated contour set based on the road curvature information to generate double yellow line geometric elements.

4. The method according to claim 3, characterized in that, The process of acquiring road curvature information and performing curvature-adaptive translation processing on the separated contour set based on the road curvature information to generate double yellow line geometric features includes: Obtain road curvature information, and calculate the compensation translation amount based on the road curvature information and the preset dynamic translation compensation rules; The translation direction is determined based on the lane direction field. The separated contour set is translated according to the translation direction and the compensation translation amount to generate double yellow line geometric elements. The translation direction is the normal direction of the direction vector at the corresponding point in the lane direction field.

5. The method according to claim 1, characterized in that, The method also includes: The geometric elements of the double yellow lines are subjected to topological verification. If the topological verification result does not meet the preset conditions, the geometric elements of the double yellow lines are automatically geometrically corrected, and the reconstructed double yellow line data is output.

6. The method according to claim 5, characterized in that, The topology verification includes verifying the parallelism between the geometric elements of the double yellow lines and the road boundary.

7. The method according to claim 1, characterized in that, The method also includes: The geometric elements of the double yellow lines are subjected to fracture interval identification. When a fracture interval is determined to exist, the interval data corresponding to the fracture interval is curve fitted according to a preset reconstruction rule to generate a reconstruction curve that connects the fracture interval.

8. A high-precision map double yellow line post-processing device, characterized in that, include: An initial contour acquisition module is used to acquire the semantic segmentation results of high-precision map data and extract an initial contour set from the semantic segmentation results; wherein, the initial contour set includes candidate contours corresponding to double yellow lines; The adhesion contour determination module is used to calculate the shortest Euclidean distance between all candidate contours in the initial contour set, and determine the adhesion contour pairs from the candidate contours based on the distance calculation results and a preset adhesion threshold. The contour separation module is used to translate the corresponding two contours in the adhesive contour pair in opposite directions along their respective normal directions by a preset translation amount according to the normal direction of their contour points, so as to obtain a set of separated contours. The feature generation module is used to generate double yellow line geometric features based on the separated contour set.

9. An electronic device, characterized in that, include: processor; as well as A memory having executable code stored thereon, which, when executed by the processor, causes the processor to perform the method as described in any one of claims 1-7.

10. A computer-readable storage medium having executable code stored thereon, characterized in that: When the executable code is executed by the processor of the electronic device, the processor performs the method as described in any one of claims 1-7.

Images