Lane deviation correction method, device, electronic device and storage medium

Abstract

The present disclosure provides a lane line correction method, device, electronic device and storage medium, belonging to the field of unmanned driving technology. The method includes: based on the position information of multiple historical nodes on the historical lane line of the actual line attribute and the position information of multiple reference vector nodes on the reference vector element matching the historical lane line, determining the correction vectors corresponding to the multiple historical nodes, the correction vector is used to indicate the correction direction and correction amplitude of the historical node; based on the correction vectors corresponding to the multiple historical nodes, adjusting the position information of the multiple historical nodes to achieve the correction of the historical lane line. After the present disclosure determines the correction vector corresponding to the historical node, it adjusts the position information of the historical node according to the correction direction and correction amplitude of the historical node indicated by the correction vector, thereby achieving automatic correction of the historical lane line. Since there is no need for cartographers to draw manually, the lane line correction efficiency is high.

Description

Technical Field

[0001] The present disclosure relates to the field of unmanned driving technology, and in particular to a lane deviation correction method, device, electronic equipment and storage medium. Background Art

[0002] Environmental perception and high-precision maps are important technologies to support autonomous driving. Unlike environmental perception, which focuses on real-time performance, high-precision maps can be used stably for a long time after they are built. However, when roads are rebuilt, traffic is reorganized, or the DOM (Digital Orthophoto Map) base map of the driving scene is iterated, the high-precision map needs to be updated to meet the needs of autonomous driving. The lane lines with actual attributes in the high-precision map are elemental representations of the ground lane lines in the actual scene. They have standardized serial numbers, and the element attributes are consistent with the ground lane lines. The lane lines with virtual attributes in the high-precision map are auxiliary lane lines artificially added in the unmarked area. As a link between the high-precision map and the real environment, the lane lines are important map data for reference by autonomous driving. How to correct the lane lines is the key to updating the high-precision map. Summary of the invention

[0003] The embodiments of the present disclosure provide a lane deviation correction method, device, electronic device and storage medium, which can improve the lane deviation correction efficiency. The technical solution is as follows:

[0004] In a first aspect, a lane deviation correction method is provided, the method comprising:

[0005] Determine, based on the position information of multiple historical nodes on the historical lane line of the actual line attribute and the position information of multiple reference vector nodes on the reference vector element, the correction vector corresponding to the multiple historical nodes, the reference vector element is a vectorized representation of the ground lane line matching the historical lane line, and the correction vector is used to indicate the correction direction and correction amplitude of the historical node;

[0006] Based on the correction vectors corresponding to the multiple historical nodes, the position information of the multiple historical nodes is adjusted to achieve correction of the historical lane line.

[0007] In another embodiment of the present disclosure, the method further includes:

[0008] When the historical lane line is a double line in the actual line attribute, another historical lane line forming a double line with the historical lane line is obtained;

[0009] Obtaining a vector element matching the other historical lane line;

[0010] When the vector element and the reference vector element coincide with each other, no deviation correction is performed on the historical lane line.

[0011] In another embodiment of the present disclosure, the determining of the correction vectors corresponding to the plurality of historical nodes based on the position information of the plurality of historical nodes on the historical lane line of the actual line attribute and the position information of the plurality of reference vector nodes on the reference vector element includes:

[0012] Performing node interpolation on the plurality of historical nodes to obtain a plurality of historical interpolation nodes;

[0013] Segmenting the historical lane line with the plurality of historical interpolation nodes as endpoints to obtain a plurality of historical lane line segments, wherein the historical lane line segments include at least one of the historical nodes;

[0014] Determine endpoint correction vectors corresponding to the two endpoints according to the position information of the two endpoints of each of the historical lane line segments and the position information of the reference vector nodes closest to the two endpoints;

[0015] Determining the total arc length of the historical line segment according to the position information of the two endpoints;

[0016] Determine the arc length from the historical node to any endpoint according to the position information of any endpoint of the two endpoints and the position information of any historical node in the same historical lane line segment;

[0017] Based on the endpoint correction vector, the total arc length, and the arc length from the historical node to any endpoint, a correction vector corresponding to the historical node is determined.

[0018] In another embodiment of the present disclosure, the method further includes:

[0019] In the case where the historical lane line is a solid and dotted line in the actual line attribute, for the endpoint of the historical lane line, when the perpendicular distance from the endpoint to the line connecting the two reference vector nodes closest to the reference vector element does not meet the distance condition, the two reference vector nodes corresponding to the endpoint are determined to be unreliable vector nodes;

[0020] In the case that the two reference vector nodes are unreliable vector nodes, the position information of the endpoint is not adjusted.

[0021] In another embodiment of the present disclosure, after not adjusting the location information of the endpoint, the method further includes:

[0022] Obtain other historical nodes adjacent to the endpoint on the historical lane line;

[0023] Obtain the vector node closest to the other historical nodes;

[0024] When the vector node closest to the other historical nodes belongs to the reference vector element, and the position information of the other historical nodes has been adjusted, the position information of the endpoint is adjusted based on the two reference vector nodes.

[0025] In another embodiment of the present disclosure, after the position information of the plurality of historical nodes is adjusted based on the correction vectors corresponding to the plurality of historical nodes to correct the historical lane line, the method further includes:

[0026] Based on the predecessor-successor relationship between the historical lane lines after correction and the lane lines, the target virtual lane line to be corrected is determined;

[0027] The corrected historical lane line is used as a position reference of the target virtual lane line, and the target virtual lane line is corrected according to the corrected historical lane line correction method.

[0028] In another embodiment of the present disclosure, after the position information of the plurality of historical nodes is adjusted based on the correction vectors corresponding to the plurality of historical nodes to correct the historical lane line, the method further includes:

[0029] When there is an unengaged endpoint in any lane line after deviation correction, based on the predecessor and successor relationship between lane lines, the adjacent endpoint of the endpoint is obtained from the predecessor lane line or the successor lane line of the lane line to obtain an endpoint pair to be engaged including the endpoint and the adjacent endpoint;

[0030] The position information of the adjacent endpoints is replaced with the position information of the endpoints to achieve the occlusion of the pair of endpoints to be occluded.

[0031] In a second aspect, a lane deviation correction device is provided, the device comprising:

[0032] A determination module, configured to determine, based on position information of multiple historical nodes on a historical lane line with actual line attributes and position information of multiple reference vector nodes on a reference vector element, a correction vector corresponding to the multiple historical nodes, wherein the reference vector element is a vectorized representation of a ground lane line matching the historical lane line, and the correction vector is used to indicate a correction direction and a correction amplitude of the historical node;

[0033] The adjustment module is used to adjust the position information of the multiple historical nodes based on the correction vectors corresponding to the multiple historical nodes to achieve correction of the historical lane lines.

[0034] In another embodiment of the present disclosure, the device further comprises:

[0035] An acquisition module, used for acquiring another historical lane line that forms a double line with the historical lane line when the historical lane line is a double line in the actual line attribute;

[0036] The acquisition module is further used to acquire a vector element matching the another historical lane line;

[0037] The adjustment module is further configured to not correct the historical lane line when the vector element coincides with the reference vector element.

[0038] In another embodiment of the present disclosure, the determination module is used to perform node interpolation on the multiple historical nodes to obtain multiple historical interpolation nodes; segment the historical lane line with the multiple historical interpolation nodes as endpoints to obtain multiple historical lane line segments, and the historical lane line segments include at least one historical node; determine the endpoint correction vector corresponding to the two endpoints according to the position information of the two endpoints of each historical lane line segment and the position information of the reference vector node closest to the two endpoints; determine the total arc length of the historical line segment according to the position information of the two endpoints; determine the arc length from the historical node to any endpoint according to the position information of any one of the two endpoints and the position information of any historical node in the same historical lane line segment; determine the correction vector corresponding to the historical node based on the endpoint correction vector, the total arc length and the arc length from the historical node to any endpoint.

[0039] In another embodiment of the present disclosure, the determination module is further configured to, when the historical lane line is a solid dashed line in the actual line attribute, for an endpoint of the historical lane line, when a perpendicular distance from the endpoint to a line connecting two reference vector nodes closest to the reference vector element does not meet a distance condition, determine that the two reference vector nodes corresponding to the endpoint are unreliable vector nodes;

[0040] The adjustment module is further configured to not adjust the position information of the endpoint when the two reference vector nodes are unreliable vector nodes.

[0041] In another embodiment of the present disclosure, the device further comprises:

[0042] An acquisition module, used for acquiring other historical nodes adjacent to the endpoint on the historical lane line;

[0043] The acquisition module is further used to acquire the vector node closest to the other historical nodes;

[0044] The adjustment module is further configured to adjust the position information of the endpoint based on the two reference vector nodes when the vector node closest to the other historical nodes belongs to the reference vector element and the position information of the other historical nodes has been adjusted.

[0045] In another embodiment of the present disclosure, the apparatus comprises:

[0046] The determination module is further used to determine the target virtual lane line to be corrected based on the historical lane line after correction and the predecessor-successor relationship between the lane lines;

[0047] The adjustment module is also used to use the corrected historical lane line as a position reference of the target virtual lane line, and correct the target virtual lane line according to the method for correcting the corrected historical lane line.

[0048] In another embodiment of the present disclosure, the device further comprises:

[0049] An acquisition module is used for, when there is an unengaged endpoint in any lane line after deviation correction, acquiring an adjacent endpoint of the endpoint from a predecessor lane line or a successor lane line of the lane line based on a predecessor-successor relationship between lane lines, to obtain an endpoint pair to be engaged including the endpoint and the adjacent endpoint;

[0050] A replacement module is used to replace the position information of the adjacent endpoints with the position information of the endpoints to achieve the occlusion of the pair of endpoints to be occluded.

[0051] In a third aspect, an electronic device is provided, comprising a processor and a memory, wherein the memory stores at least one program code, and the at least one program code is loaded and executed by the processor to implement the lane correction method as described in the first aspect.

[0052] In a fourth aspect, a computer-readable storage medium is provided, wherein at least one program code is stored in the storage medium, and the at least one program code is loaded and executed by a processor to implement the lane correction method as described in the first aspect.

[0053] In a fifth aspect, a computer program product is provided, which includes a computer program code, and the computer program code is stored in a computer-readable storage medium. A processor of a server reads the computer program code from the computer-readable storage medium, and the processor executes the computer program code, so that the server executes the lane correction method as described in the first aspect.

[0054] The technical solution provided by the embodiments of the present disclosure has the following beneficial effects:

[0055] There is no need for cartographers to compare and check one by one. Based on the position information of historical nodes on the historical lane line and the position information of reference vector nodes on the reference vector elements that match the historical lane line, the correction vector corresponding to the historical node is determined. The correction vector represents the correction direction and correction amplitude of the historical node. Based on the correction vector, the position information of the historical node is adjusted so that the shape of the corrected historical lane line can fit with the historical lane line, and the corrected historical lane line is as close as possible to the reference vector elements, which not only improves the correction efficiency of the historical lane line, but also has a high accuracy of the correction result. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0057] Figure 1 is a flow chart of a lane deviation correction method provided by an embodiment of the present disclosure;

[0058] Figure 2 It is a schematic diagram of a lane line correction process in the prior art;

[0059] Figure 3 is a schematic diagram of a lane deviation correction processing result in the prior art;

[0060] Figure 4 is a schematic diagram of a historical lane densification process provided by an embodiment of the present disclosure;

[0061] Figure 5 is a schematic diagram of a smoothing correction strategy provided by an embodiment of the present disclosure;

[0062] Figure 6 is a schematic diagram of a virtual line correction provided by an embodiment of the present disclosure;

[0063] Figure 7 is a schematic diagram of a lane deviation correction process provided by an embodiment of the present disclosure;

[0064] Figure 8 is a structural schematic diagram of a lane deviation correction device provided by an embodiment of the present disclosure;

[0065] Fig. 9 A structural block diagram of an electronic device provided by an exemplary embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0066] In order to make the objectives, technical solutions and advantages of the present disclosure more clear, the embodiments of the present disclosure will be further described in detail below with reference to the accompanying drawings.

[0067] It can be understood that the terms "each", "multiple", and "any" used in the embodiments of the present disclosure include two or more, each refers to each of the corresponding multiple, and any refers to any one of the corresponding multiple. For example, the multiple words include 10 words, and each word refers to each word of the 10 words, and any word refers to any one of the 10 words.

[0068] It should be noted that all actions of acquiring signals, information or data in the present disclosure are carried out in compliance with the relevant data protection laws and policies of the country where the device is located and with the authorization given by the owner of the corresponding device.

[0069] The embodiment of the present disclosure provides a lane deviation correction method. Taking the electronic device executing the embodiment of the present disclosure as an example, the electronic device can be a terminal with strong computing power, such as a smart phone, a laptop computer, a vehicle-mounted terminal, etc.; the electronic device can also be a server, for example, it can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, etc. The embodiment of the present disclosure does not specifically limit the type of electronic device. Figure 1 , the method flow provided by the embodiment of the present disclosure includes:

[0070] 101. The electronic device pre-processes multiple historical lane lines and multiple vector elements.

[0071] Among them, the historical lane line is the lane line displayed on the high-precision map before the road reconstruction, traffic reorganization or DOM base map iteration of the driving scene occurs. The historical lane line includes multiple historical nodes. When selecting the multiple historical nodes, multiple points randomly selected from the historical lane line can be used as multiple historical nodes, or multiple points can be selected from the historical lane line at a certain distance (for example, 10 meters, 20 meters, etc.) as multiple historical nodes, and multiple points with identification functions can be selected from the historical lane line as multiple historical nodes. The embodiment of the present disclosure does not specifically limit the selection method of multiple historical nodes. Lane lines include lane lines with actual line (non-virtual) attributes and lane lines with virtual line (virtual) attributes. Lane lines with actual attributes are elemental representations of ground lane lines in actual scenes, including actual lines, real and virtual lines, etc. Lane lines with virtual line attributes are virtual lane lines artificially constructed in unmarked areas to facilitate driving guidance of unmanned vehicles. Lane lines with virtual line attributes are not marked on the ground of the actual scene, but are only displayed on the high-precision map of the unmanned driving scene.

[0072] Among them, the vector element is the vectorized representation of the ground lane line in the DOM. The vector element includes multiple vector nodes. When selecting the vector nodes, multiple points randomly selected from the vector element can be used as multiple vector nodes, or multiple points can be selected from the vector element at a certain distance (for example, 20 meters, 30 meters, etc.) as multiple vector nodes. Multiple points with identification functions can also be selected from the vector element as multiple vector nodes. The embodiment of the present disclosure does not make specific restrictions on the selection method of multiple vector nodes. Furthermore, after the vector element extraction is completed, the vector element will be subjected to certain data compression processing, and some vector elements may be filtered out during the processing.

[0073] In actual applications, although both vector elements and historical lane lines correspond to the ground lane lines in the actual scene, vector elements lack standardized serial numbers compared to historical lane lines, and are not aligned with the endpoints of historical lane lines. Therefore, when correcting the historical lane lines, it is impossible to directly replace the historical nodes with vector nodes. However, although it is impossible to correct the historical lane lines directly by replacing the nodes, since the vector elements provide a more accurate actual lane line position, they can also provide a reference for correcting the historical lane lines of the actual line attributes.

[0074] In order to improve the efficiency and accuracy of correcting the historical lane lines, the method provided in the embodiment of the present disclosure will also pre-process multiple vector elements and multiple historical lane lines before correcting the historical lane lines. The pre-processing stage mainly generates some necessary data required for correcting the historical lane lines, so as to correct the historical lane lines based on these necessary data. The necessary data includes the KD tree of the vector elements, the predecessor and successor relationship of the lane lines, the KD tree of the historical lane lines, etc.

[0075] Among them, the predecessor-successor relationship is mainly the adjacency relationship between lane lines. Simply put, for a lane line, determining the predecessor-successor relationship of the lane line is mainly to find two or more lane lines adjacent to the starting and ending endpoints of the lane line. Considering that lane lines can be divided into two-way and one-way, the predecessor-successor relationship mainly includes four types: starting endpoint predecessor and ending endpoint successor, starting endpoint successor and ending endpoint predecessor.

[0076] A KD tree is a tree structure that stores instance points in a K-dimensional space for quick retrieval. For vector elements, the KD tree of the vector elements stores the identification of the vector elements, the order of each vector node in the vector elements, the location information of each vector node, etc., and the location information can be the location coordinates of the vector nodes. For historical lane lines, a KD tree can be constructed for each historical lane line, and a KD tree can also be constructed for historical lane lines with double-line attributes. Taking into account the time of the preprocessing process and the search requirements in actual scenarios, the embodiment of the present disclosure mainly constructs a KD tree for historical lane lines with double-line attributes. The historical lane lines with double-line attributes refer to two lane lines with a distance less than a distance threshold. The distance threshold can be 30 cm, 50 cm, etc. The embodiment of the present disclosure does not specifically limit the distance threshold. In real life, historical lane lines with double-line attributes can be double yellow lines, double white lines, etc. Considering the complexity of the road conditions in actual scenarios, the distances between historical lane lines with double-line attributes in different sections may be different, and the distances between historical lane lines with double-line attributes at different locations in the same section are also different. The KD tree constructed for the historical lane lines with double-line attributes in the disclosed embodiment stores the identification of each historical lane line constituting the double line, the order of each historical node in each historical lane line, the location information of each historical node, etc.

[0077] For the KD tree corresponding to the vector element, the embodiment of the present disclosure may adopt the following method when constructing:

[0078] In the first step, the electronic device performs node interpolation on multiple original vector nodes based on multiple original vector nodes of each segment of vector elements to obtain multiple interpolation vector nodes, and combines the multiple original vector nodes and the multiple interpolation vector nodes to form multiple vector nodes of the segment of vector elements.

[0079] The electronic device performs node interpolation operations on each segment of vector elements at every preset distance, and the obtained interpolated vector nodes and the original vector nodes are both used as vector nodes. The preset distance may be 30 cm, 40 cm, etc., and the embodiment of the present disclosure does not specifically limit the preset distance.

[0080] In the second step, the electronic device constructs a KD tree corresponding to the vector elements based on the vector nodes of all segment vector elements.

[0081] When constructing a KD tree corresponding to a vector element, the electronic device records the identifier of each segment of the vector element and the position information of the first and last endpoints of the segment of the vector element.

[0082] Since the matching of historical lane lines and vector elements in the disclosed embodiment is node matching, and whether the nodes are matched is mainly determined based on the distance, and since the KD tree of the vector elements stores the relationship between multiple vector nodes and vector elements and the position information of multiple vector nodes, the KD tree of the vector elements is used in the subsequent steps to search for matching vector elements for the historical lane lines, which can greatly improve the search efficiency.

[0083] Since historical lane lines with double-line attributes have strong correlation, they often need to be combined for correction processing during the correction process. Therefore, in order to improve the subsequent correction accuracy of historical lane lines with double-line attributes, the electronic device needs to build a KD tree for historical lane lines with double-line attributes. For the KD tree corresponding to the historical lane lines with double-line attributes, the following method can be used in the embodiment of the present disclosure when constructing it:

[0084] In the first step, the electronic device determines the historical lane lines with double-line attributes based on the distance between the historical lane lines.

[0085] In the second step, the electronic device constructs a KD tree corresponding to the historical lane lines with double-line attributes based on the historical nodes in the historical lane lines with double-line attributes.

[0086] For the predecessor and successor relationship of lane lines, the electronic device can obtain the predecessor and successor relationship of each road section from the relevant files of the historical high-precision map, and then integrate the predecessor and successor relationship of each road section into the historical lane line. This method of determining the predecessor and successor relationship of the lane line is the inheritance formula. The electronic device will also calculate the distance between the endpoints of each historical lane line based on the position information of the endpoints on each historical lane line, so as to determine the predecessor lane line and the successor lane line of the historical lane line based on the distance. This method of determining the predecessor and successor relationship of the lane line is the geometric formula. For example, the distance threshold is set to 20 centimeters. For any historical lane line, if the distance between the starting endpoint of the historical lane line and the ending endpoint of another historical lane line is less than 20 centimeters, the other historical lane line is determined to be the predecessor lane line of the historical lane line; if the distance between the ending endpoint of the historical lane line and the starting endpoint of another historical lane line is less than 20 centimeters, the other historical lane line is determined to be the successor lane line of the historical lane line. In order to obtain a complete and robust predecessor and successor relationship of lane lines, the embodiment of the present disclosure can use the above inheritance formula and geometric formula to mutually verify and complement the predecessor and successor relationship of the determined lane lines. The predecessor-follower relationship of the lane lines determined by the embodiment of the present disclosure can be applied in subsequent actual line deviation correction, virtual line deviation correction and post-processing.

[0087] 102. The electronic device determines a reference vector element that matches the historical lane line based on the position information of multiple historical nodes on the historical lane line of the actual line attribute and the position information of multiple vector nodes on multiple vector elements.

[0088] Since both the vector elements and the historical lane lines with actual line attributes correspond to the ground lane lines in the actual scene, the electronic device can correct the historical lane lines with actual line attributes based on the matching vector elements. Before that, the electronic device needs to determine the reference vector elements that match the historical lane lines with actual line attributes. The specific determination process is as follows:

[0089] 1021. The electronic device calculates the distance between each historical node and each vector node based on the position information of each historical node and the position information of multiple vector nodes.

[0090] The distance between the history node and the vector node may be the Euclidean distance or the like.

[0091] 1022. When the two vector nodes with the smallest distance from any historical node belong to the same vector element, and the distance between the historical node and the two vector nodes is less than the first distance threshold, the electronic device determines the vector element as a reference vector element that matches the historical lane line.

[0092] The first distance threshold may be determined according to the correction accuracy, and the first distance threshold may be 50 cm, 60 cm, etc. In a possible implementation, the electronic device may calculate the distance between the historical node and each vector node in the multiple vector elements, and sort the multiple vector nodes in order of distance (from small to large or from large to small), and then obtain the two vector nodes with the smallest distance from the historical node. If the two vector nodes belong to the same vector element and the distance between the historical node and the two vector nodes is less than the first distance threshold, the vector element is determined as a reference vector element matching the historical lane line.

[0093] In another possible implementation, the electronic device obtains two vector nodes with the smallest distance from the historical node based on a pre-built KD tree of vector elements. If the historical node and the two vector nodes belong to the same vector element and the distance between the historical node and the two vector nodes is less than a first distance threshold, the vector element is determined as a reference vector element that matches the historical lane line.

[0094] In another embodiment of the present disclosure, taking into account the particularity of historical lane lines with double-line attributes, usually one group of historical lane lines with double-line attributes will have another group of historical lane lines with double-line attributes that are close to it and have a relatively fixed spacing, and these two groups of historical lane lines need to be combined for deviation correction. However, in actual applications, the vector element extraction algorithm of the base map is prone to problems such as missed detection, false detection, and concatenation on the double-line lane lines. In order to improve the accuracy of the deviation correction results of the historical lane lines with double-line attributes, when the historical lane lines are double lines in the actual line, the electronic device will also verify whether the historical lane lines with double-line attributes are corrected. Specifically, the following method can be used:

[0095] In the first step, the electronic device obtains another historical lane line that forms a double line with the historical lane line.

[0096] In this step, the electronic device may search for another historical lane line that forms a double line with the historical lane line by calculating the distance between the historical lane lines based on the KD tree pre-constructed for the historical lane line with double-line attributes.

[0097] In a second step, the electronic device obtains a vector element that matches another historical lane line.

[0098] The way in which the electronic device obtains the vector element matching another historical lane line is the same as the way in which the above-mentioned electronic device obtains the vector element matching the historical lane line. For details, please refer to the above-mentioned step 202, which will not be repeated here.

[0099] In the third step, when the vector element matching another historical lane line coincides with the reference vector element, the electronic device does not correct the historical lane line.

[0100] This step compares the vector elements of the two sets of historical lane line matches to determine whether there are errors in the vector elements extracted from the base map. If there are no errors in the vector elements extracted from the base map, the two sets of historical lane lines will be corrected based on their respective matching vector elements; if there are errors in the vector elements extracted from the base map, no correction will be made to the historical lane lines. When judging whether there are errors in the vector elements extracted from the base map, it can be determined whether the vector elements of the two sets of historical lane line matches overlap. Since the endpoints of the historical lane lines with double-line attributes are not strictly aligned, when obtaining the matching vector elements of the two sets of historical lane lines, the matching vector elements of their predecessors and successors can be used to expand the reference vector element set.

[0101] The disclosed embodiment verifies the historical lane lines with double-line attributes before correction, so as to avoid using erroneous vector elements to correct the historical lane lines, which may cause the historical lane lines to become more and more erroneous during correction and affect the driving safety of the unmanned driving equipment.

[0102] 103. The electronic device determines the correction vectors corresponding to the multiple historical nodes based on the position information of the multiple historical nodes on the historical lane line and the position information of the multiple reference vector nodes on the reference vector element.

[0103] Before using the method provided in the embodiment of the present disclosure to correct the historical lane line, the related technology mainly corrects the lane line perpendicular to the matched vector element. Figure 2 The main correction process of the related technology shown in the figure is:

[0104] In the first step, search for two vector nodes with the smallest distance to the historical node on the historical lane line as two reference vector nodes.

[0105] The second step is to connect each historical node to the corresponding two reference vector nodes as a perpendicular line, and use the foot of the perpendicular as the correction position of each historical node. When all historical nodes have completed the correction, a new line segment combination will be formed, that is, the overall correction of the historical lane line is completed. When this correction method is used for correction, for each historical node of the historical lane line, after obtaining the two reference vector nodes closest to the node, it is determined whether the two reference vector nodes belong to the same vector element. If the two reference vector nodes belong to the same vector element, the historical node is corrected. If the two reference vector nodes do not belong to the same vector element, the historical node is not corrected.

[0106] However, the curvature of the historical lane lines and vector elements are not strictly consistent. Figure 3 When using the strategy to correct the deviation, it may appear Figure 3 The phenomenon that the correction result is not aligned with the vector element, and this part of the misalignment may not meet the correction requirements, resulting in invalid correction. In order to improve the accuracy of the correction results of the historical lane lines, the node correction strategy proposed in the embodiment of the present disclosure determines the correction vector corresponding to each historical node. The correction vector can characterize the vector of the correction direction and correction amplitude of the historical node, so as to effectively maintain the shape of the historical lane line while trying to approach the vector element, that is, to rely on the shape fitting of the historical lane line to the actual lane marking, and at the same time make full use of the position information provided by the reference vector element. When the electronic device determines the correction vector corresponding to each historical node based on the position information of each historical node and the position information of multiple reference vector nodes on the reference vector element, the following method can be used:

[0107] 1031. The electronic device performs node interpolation on multiple historical nodes to obtain multiple historical interpolation nodes.

[0108] Since the node positions of the vector elements and the nodes in the historical lane lines are not consistent, the vector elements and the historical lane lines are quite different. In order to achieve the approximation of the vector elements, the historical nodes of the historical lane lines need to be aligned with the vector nodes of the vector elements. In order to achieve the alignment of the historical nodes of the historical lane lines with the vector nodes of the vector elements, so that there are historical nodes on the historical lane lines that are close enough to any reference vector node, it is necessary to perform node interpolation on the historical lane lines. When performing node interpolation on the historical lane lines, interpolation can be performed according to a certain distance. For details, please refer to the above interpolation method for vector elements, which will not be repeated here.

[0109] Furthermore, there may be redundant historical interpolation nodes among multiple historical interpolation nodes. In order to avoid invalid correction based on redundant historical interpolation nodes, the disclosed embodiment can screen multiple historical interpolation nodes. Specifically, the electronic device searches for a historical interpolation node that is closest to each reference vector node among multiple historical interpolation nodes of the historical lane line, and retains the closest historical interpolation node as a valid historical interpolation node, and deletes other historical interpolation nodes. The valid historical interpolation node will participate in the correction process together with the original historical node. By performing node interpolation on historical nodes, the nodes on the historical lane line are densified. The densified historical lane line can be seen in Figure 4 .

[0110] 1032. The electronic device segments the historical lane line using multiple historical interpolation nodes as endpoints to obtain multiple historical lane line segments.

[0111] The historical lane line segment may include at least one historical node. Of course, the historical lane line segment may not include a historical node, in which case only the two endpoints of the historical lane line segment need to be corrected.

[0112] 1033. The electronic device determines the endpoint correction vectors corresponding to the two endpoints based on the position information of the two endpoints of each historical lane line segment and the position information of the reference vector node closest to the two endpoints.

[0113] For the starting endpoint of the historical lane line segment, the electronic device determines the reference vector node on the reference vector that is closest to the starting endpoint based on the position information of the starting endpoint, and uses a directed line segment to connect the starting endpoint and the reference vector node of the starting endpoint with the starting endpoint, to obtain the correction vector corresponding to the starting endpoint. The direction of the correction vector corresponding to the starting endpoint is that the starting endpoint points to the reference vector node, and the position information of the correction vector can be determined based on the position information of the reference vector node and the position information of the starting endpoint. For example, subtract the coordinates of the reference vector node from the coordinates of the same dimension of the starting endpoint. The correction vector corresponding to the starting endpoint can be found in Figure 5 The vector A′A in .

[0114] For the end endpoint of the historical lane line segment, the electronic device determines the reference vector node on the reference vector that is closest to the end endpoint based on the position information of the end endpoint, and uses a directed line segment to connect the end endpoint and the reference vector node of the end endpoint with the end endpoint, to obtain the correction vector corresponding to the end endpoint. The direction of the correction vector corresponding to the end endpoint is that the end endpoint points to the corresponding reference vector node, and the position information of the correction vector can be determined based on the position information of the reference vector node and the position information of the end endpoint. For example, subtract the coordinates of the reference vector node from the coordinates of the end endpoint in the same dimension. The correction vector corresponding to the end endpoint can be found in Figure 5 The vector B′B in .

[0115] 1034. The electronic device determines the total arc length of the historical lane line segment based on the position information of the two endpoints.

[0116] Based on the position information of the starting endpoint and the ending endpoint of the historical lane line segment, the electronic device calculates the arc length between each node, and takes the sum of the arc lengths between each node as the total arc length of the historical lane line segment.

[0117] 1035. The electronic device determines the arc length from the historical node to any of the two endpoints based on the position information of the two endpoints and the position information of the historical node.

[0118] The electronic device calculates the distance from the historical node to the starting endpoint based on the sum of arc lengths of each node segment between the historical node and the starting endpoint in the historical lane line segment. Similarly, the electronic device calculates the distance from the historical node to the ending endpoint based on the sum of arc lengths of each node segment between the historical node and the ending endpoint in the historical lane line segment.

[0119] 1036. The electronic device determines the correction vector corresponding to the historical node based on the endpoint correction vector, the total arc length and the distance from the historical node to any endpoint.

[0120] Specifically, the electronic device can apply the following formula to determine the correction vector of the historical node based on the correction vector of the starting endpoint, the correction vector of the ending endpoint, the total arc length of the historical lane line segment, and the arc length from the historical node to the starting endpoint:

[0121]

[0122] Among them, e i represents the correction vector of the historical node, l i Indicates the arc length from the historical node to the starting endpoint, l total represents the total arc length of the historical lane segment, e A′→Arepresents the correction vector of the starting endpoint, e B′→B represents the correction vector of the end endpoint, A represents the start endpoint, A′ represents the target correction position corresponding to the start endpoint, B represents the end endpoint, and B′ represents the target correction position corresponding to the end endpoint.

[0123] The above takes the determination of the correction vector of any historical node in any historical lane line segment as an example. The process of determining the correction vectors of other historical nodes in the historical lane line segment, and the process of determining the correction vectors corresponding to the historical nodes in other historical lane line segments are the same as the process of determining the correction vector corresponding to the historical node. Please refer to the above process for details, and they will not be described one by one here.

[0124] In another embodiment of the present disclosure, in order to improve the accuracy of the correction result, it is necessary to verify the reliability of the reference vector node on which the correction of each historical node depends. Specifically, for each of the multiple historical nodes, the electronic device obtains the two reference vector nodes closest to the historical node on the reference vector element, and draws a perpendicular line to the line between the two reference vector nodes, and then obtains the perpendicular distance from the historical node to the line between the two reference vector nodes. When the perpendicular distance is less than the second distance threshold, the electronic device determines that the two reference vector nodes are reliable vector nodes, and then, when the two reference vector nodes are reliable vector nodes, performs an operation based on the position information of each historical node and the position information of multiple reference vector nodes on the reference vector element to determine the correction vector corresponding to each historical node. Among them, the second distance threshold can be determined according to the correction accuracy, and the second distance threshold can be 20 cm, 30 cm, and so on.

[0125] In another embodiment of the present disclosure, a type of endpoint correction error may occur in the real and dotted line correction. For example, when the adjacent endpoints of the connected historical lane lines are located in a wireless area, the distance between them and the nearest vector node will usually exceed the reliable distance threshold. At this time, the endpoint of the historical lane line cannot be corrected based on the reference vector element. In judgment for this situation, the electronic device can obtain the perpendicular distance between the endpoint and the line between the two closest reference vector nodes on the reference vector element for the endpoint of the historical lane line when the historical lane line is a real and dotted line in the actual line attribute. When the perpendicular distance is greater than or equal to the second distance threshold, the electronic device determines that the two reference vector nodes of the endpoint are unreliable vector nodes. In the case where the two reference vector nodes are unreliable vector nodes, the electronic device does not correct the endpoint. In the embodiment of the present disclosure, the electronic device ensures the accuracy of the correction result by performing reliability verification on the reference vector nodes on which the endpoint correction of the historical lane line depends.

[0126] However, when the electronic device does not correct the endpoint of the historical lane line, if the historical node adjacent to the endpoint has been effectively corrected at this time, the corrected historical lane line will have a certain angle here. In response to such problems, the disclosed embodiment can comprehensively compare the reference vector nodes between adjacent historical nodes, whether to correct the endpoint, and other information, to determine whether to correct the endpoint based on the unreliable vector node, thereby ensuring the smoothness of the lane line after correction. Specifically, the electronic device obtains other historical nodes adjacent to the endpoint on the historical lane line, and obtains the vector node closest to the other historical nodes. When the vector node closest to the other historical nodes belongs to the reference vector element, and the position information of the other historical nodes has been adjusted, the electronic device adjusts the position information of the endpoint based on the two reference vector nodes. In the disclosed embodiment, the electronic device uses an unreliable vector node to correct the endpoint when the historical node adjacent to the endpoint in the historical lane line is effectively corrected, so that the corrected historical lane line can remain smooth as a whole.

[0127] 104. The electronic device adjusts the position information of the multiple historical nodes based on the correction vectors corresponding to the multiple historical nodes to achieve correction of the historical lane lines.

[0128] Based on the correction vector corresponding to the historical node, the electronic device can correct the historical node by adjusting the position information of the historical node. Specifically, the electronic device adds the position coordinates of the historical node with the coordinates of the same dimension in the correction vector to obtain the position coordinates of the correction node corresponding to the historical node, and corrects the historical node by assigning the coordinates of the correction node to the historical node. When each historical node on the historical lane line completes the above correction, the electronic device connects each corrected historical node to complete the correction of the historical lane line. Since the correction is based on the smoothness of the historical lane line, the historical lane line corrected by this method will ensure the smoothness of the original historical lane line as much as possible, while effectively approximating the position of the vector element.

[0129] Furthermore, after completing the correction of the historical lane lines of each actual line attribute, the electronic device will mark the corrected historical lane lines and record the correction status of the starting endpoints and ending endpoints of the corrected historical lane lines, thereby correcting and post-processing the virtual lane lines based on the recorded information.

[0130] 105. The electronic device corrects the target virtual lane line based on the corrected historical lane line.

[0131] Since the historical lane lines with vector elements and actual line attributes correspond to the ground lane lines in the actual scene, while the historical lane lines with virtual line attributes do not correspond to the ground lane lines in the actual scene, only the historical lane lines with actual line attributes can be corrected based on vector elements, and the historical lane lines with virtual line attributes cannot be corrected directly. However, the corrected historical lane lines can be used for correction. Specifically, the following methods can be used:

[0132] 1051. The electronic device determines the target virtual lane line to be corrected based on the historical lane line after correction and the predecessor-successor relationship between the lane lines.

[0133] Figure 6 The process of determining the target virtual lane line to be corrected is shown in FIG. Figure 6 The electronic device obtains the corrected historical lane line, and uses the starting endpoint or the ending endpoint (endpoint A) of the corrected historical lane line and the predecessor and successor relationship between lane lines to obtain the target virtual lane line connected to endpoint A, and determines endpoint A' (the endpoint of the target virtual lane line) connected to endpoint A. Then, the electronic device obtains the other endpoint B' of the target virtual lane line, and uses the predecessor and successor relationship between endpoint B' and the lane line to obtain endpoint B of the corrected historical lane line connected to endpoint B' (the attribute of the corrected historical lane line can be an actual line or a virtual line), thereby obtaining the target correction endpoints A and B of the target virtual lane line, and the endpoints A' and B' to be corrected.

[0134] 1052. The electronic device uses the corrected historical lane line as a position reference for the target virtual lane line, and corrects the target virtual lane line according to the method for correcting the corrected historical lane line.

[0135] 106. The electronic device performs bite processing on the endpoints of the corrected lane line.

[0136] Due to issues such as mapping technology and algorithm accuracy, the quality of vector elements extracted from the DOM base map is poor in some areas. Usually, no correction will be performed in such areas to avoid incorrect correction. However, this processing method will cause the historical lane lines connected at the regional boundaries of such areas to be disconnected, resulting in mapping errors. In order to solve this problem, the embodiment of the present disclosure can perform bite processing on the endpoints of the corrected lane lines based on the predecessor and successor relationship of the lane lines.

[0137] Specifically, when there are unengaged endpoints in any lane line after deviation correction, the electronic device obtains the adjacent endpoint of the endpoint from the predecessor lane line or the successor lane line of the lane line based on the predecessor-successor relationship between the lane lines, obtains the pair of endpoints to be engaged including the endpoint and the adjacent endpoint, and then replaces the position information of the adjacent endpoint with the position information of the endpoint to achieve engagement of the pair of endpoints to be engaged. The disclosed embodiment ensures the continuity of the lane line after deviation correction by engaging the endpoints of the lane line after deviation correction to avoid interruption.

[0138] Figure 7 The four stages of the lane correction method provided by the embodiment of the present disclosure are shown, namely the pre-processing stage, the actual line correction stage, the virtual line correction stage and the processing stage. The pre-processing stage mainly generates some necessary data, including the KD tree of the vector element nodes, the KD tree based on the historical lane lines and the predecessor and successor relationship between the lane lines, for use in subsequent steps. The actual line correction stage mainly matches the historical lane lines of the actual line attributes with the vector elements, and corrects the node positions on the historical lane lines of the actual line attributes based on the position information provided by the vector elements. The virtual line correction stage mainly uses the correction results of the historical lane lines of the actual line attributes and the predecessor and successor relationship of the lane lines to correct the position of the virtual lane lines. The post-processing stage is mainly for the disconnected lane lines, using the correction of the historical lane lines of the actual line attributes, the correction of the historical lane lines of the virtual line attributes and the predecessor and successor relationship of the lane lines, to perform endpoint bite correction on the connected lane lines.

[0139] The method provided by the disclosed embodiment does not require the draftsman to compare and check one by one. Based on the position information of the historical nodes on the historical lane line and the position information of the vector nodes on the vector elements, the reference vector elements matching the historical lane line can be determined, thereby improving the search efficiency of the reference vector elements. Based on the position information of the historical nodes on the historical lane line and the position information of the reference vector nodes on the reference vector elements matching the historical lane line, the correction vector corresponding to the historical node is determined. The correction vector characterizes the correction direction and correction amplitude of the historical node. Based on the correction vector, the position information of the historical node is adjusted so that the shape of the corrected historical lane line can fit the historical lane line, and the corrected historical lane line is as close to the reference vector element as possible, which not only improves the correction efficiency of the historical lane line, but also has a high accuracy of the correction result.

[0140] See also Figure 8 , the embodiment of the present disclosure provides a lane deviation correction device, the device comprising:

[0141] A determination module 801 is used to determine the correction vectors corresponding to the multiple historical nodes based on the position information of the multiple historical nodes on the historical lane line of the actual line attribute and the position information of the multiple reference vector nodes on the reference vector element, where the reference vector element is a vectorized representation of the ground lane line matching the historical lane line, and the correction vector is used to indicate the correction direction and correction amplitude of the historical node;

[0142] The adjustment module 802 is used to adjust the position information of multiple historical nodes based on the correction vectors corresponding to the multiple historical nodes to achieve correction of the historical lane lines.

[0143] In another embodiment of the present disclosure, the device further comprises:

[0144] An acquisition module, used for acquiring another historical lane line that forms a double line with the historical lane line when the historical lane line is a double line in the actual line attribute;

[0145] The acquisition module is also used to obtain a vector element matching another historical lane line;

[0146] The adjustment module is also used to not correct the historical lane line when the vector element and the reference vector element coincide with each other.

[0147] In another embodiment of the present disclosure, a determination module is used to perform node interpolation on multiple historical nodes to obtain multiple historical interpolation nodes; segment the historical lane line with multiple historical interpolation nodes as endpoints to obtain multiple historical lane line segments, and the historical lane line segments include at least one historical node; determine the endpoint correction vectors corresponding to the two endpoints based on the position information of the two endpoints of each historical lane line segment and the position information of the reference vector node closest to the two endpoints; determine the total arc length of the historical line segment based on the position information of the two endpoints; determine the arc length from the historical node to any endpoint based on the position information of any of the two endpoints and the position information of any historical node in the same historical lane line segment; determine the correction vector corresponding to the historical node based on the endpoint correction vector, the total arc length and the arc length from the historical node to any endpoint.

[0148] In another embodiment of the present disclosure, the determination module is further used to, when the historical lane line is a solid dashed line in the actual line attribute, for the endpoint of the historical lane line, when the perpendicular distance between the endpoint and the line connecting the two closest reference vector nodes on the reference vector element does not meet the distance condition, determine that the two reference vector nodes corresponding to the endpoint are unreliable vector nodes;

[0149] The adjustment module is further used for not adjusting the position information of the endpoint when the two reference vector nodes are unreliable vector nodes.

[0150] In another embodiment of the present disclosure, the device further comprises:

[0151] An acquisition module is used to acquire other historical nodes adjacent to the endpoint on the historical lane line;

[0152] The acquisition module is also used to obtain the vector node closest to other historical nodes;

[0153] The adjustment module is also used to adjust the position information of the endpoint based on two reference vector nodes when the vector node closest to other historical nodes belongs to the reference vector element and the position information of other historical nodes has been adjusted.

[0154] In another embodiment of the present disclosure, the apparatus comprises:

[0155] The determination module is further used to determine the target virtual lane line to be corrected based on the historical lane line after correction and the predecessor and successor relationship between the lane lines;

[0156] The adjustment module is also used to use the corrected historical lane line as a position reference for the target virtual lane line, and to correct the target virtual lane line according to the method for correcting the corrected historical lane line.

[0157] In another embodiment of the present disclosure, the device further comprises:

[0158] An acquisition module is used to acquire the adjacent endpoint of the endpoint from the predecessor lane line or the successor lane line of the lane line when there is an unengaged endpoint in any lane line after deviation correction, based on the predecessor and successor relationship between the lane lines, to obtain an endpoint pair to be engaged including the endpoint and the adjacent endpoint;

[0159] The replacement module is used to replace the position information of the adjacent endpoints with the position information of the endpoints to achieve the occlusion of the endpoint pairs to be occluded.

[0160] In summary, the device provided by the embodiment of the present disclosure does not require the draftsman to compare and check one by one. Based on the position information of the historical node on the historical lane line and the position information of the reference vector node on the reference vector element matching the historical lane line, the correction vector corresponding to the historical node is determined, and the correction vector characterizes the correction direction and correction amplitude of the historical node. Based on the correction vector, the position information of the historical node is adjusted so that the shape of the corrected historical lane line can be fitted with the historical lane line, and the corrected historical lane line is as close as possible to the reference vector element, which not only improves the correction efficiency of the historical lane line, but also has a high accuracy of the correction result.

[0161] Fig. 9 The structure block diagram of an electronic device 900 provided by an exemplary embodiment of the present disclosure is shown. Generally, the electronic device 900 includes: a processor 901 and a memory 902 .

[0162] The processor 901 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 901 may be implemented in at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). The processor 901 may also include a main processor and a coprocessor. The main processor is a processor for processing data in the awake state, also known as a CPU (Central Processing Unit); the coprocessor is a low-power processor for processing data in the standby state. In some embodiments, the processor 901 may be integrated with a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 901 may also include an AI (Artificial Intelligence) processor, which is used to process computing operations related to machine learning.

[0163] The memory 902 may include one or more computer-readable storage media, which may be non-transitory. The memory 902 may also include a high-speed random access memory, and a non-volatile memory, such as one or more disk storage devices, flash memory storage devices. In some embodiments, the non-transitory computer-readable storage medium in the memory 902 is used to store at least one instruction, which is used to be executed by the processor 901 to implement the lane line deviation correction method provided in the method embodiment of the present disclosure.

[0164] In some embodiments, the electronic device 900 may further optionally include: a peripheral device interface 903 and at least one peripheral device. The processor 901, the memory 902 and the peripheral device interface 903 may be connected via a bus or a signal line. Each peripheral device may be connected to the peripheral device interface 903 via a bus, a signal line or a circuit board. Specifically, the peripheral device includes: a power supply 904.

[0165] The peripheral device interface 903 may be used to connect at least one peripheral device related to I / O (Input / Output) to the processor 901 and the memory 902. In some embodiments, the processor 901, the memory 902, and the peripheral device interface 903 are integrated on the same chip or circuit board; in some other embodiments, any one or two of the processor 901, the memory 902, and the peripheral device interface 903 may be implemented on a separate chip or circuit board, which is not limited in this embodiment.

[0166] The power supply 904 is used to power various components in the electronic device 900. The power supply 904 can be an alternating current, a direct current, a disposable battery, or a rechargeable battery. When the power supply 904 includes a rechargeable battery, the rechargeable battery can be a wired rechargeable battery or a wireless rechargeable battery. A wired rechargeable battery is a battery charged through a wired line, and a wireless rechargeable battery is a battery charged through a wireless coil. The rechargeable battery can also be used to support fast charging technology.

[0167] Those skilled in the art will understand that Fig. 9 The structure shown in the figure does not constitute a limitation on the electronic device 900, and may include more or less components than those shown in the figure, or combine some components, or adopt a different component arrangement.

[0168] In an exemplary embodiment, a computer-readable storage medium including instructions is also provided, such as a memory including instructions, and the above instructions can be executed by a processor of the electronic device 900 to complete the above lane line correction method. Optionally, the storage medium can be a non-temporary computer-readable storage medium, for example, the non-temporary computer-readable storage medium can be a CD-ROM (Compact Disc Read-Only Memory), ROM (Read-Only Memory), RAM (Random Access Memory), magnetic tape, floppy disk, optical data storage device, etc.

[0169] The embodiment of the present disclosure provides a computer-readable storage medium, in which at least one program code is stored, and the at least one program code is loaded and executed by a processor to implement the lane deviation correction method. The computer-readable storage medium may be non-transient.

[0170] An embodiment of the present disclosure provides a computer program product, which includes a computer program code. The computer program code is stored in a computer-readable storage medium. A processor of a server reads the computer program code from the computer-readable storage medium, and the processor executes the computer program code, so that the server executes the above-mentioned lane line correction method.

[0171] A person skilled in the art will understand that all or part of the steps to implement the above embodiments may be accomplished by hardware or by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a disk or an optical disk, etc.

[0172] The above description is only an optional embodiment of the present disclosure and is not intended to limit the present disclosure. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A lane deviation correction method, characterized in that: The method comprises: Perform node interpolation on multiple historical nodes on the historical lane line of the actual line attribute to obtain multiple historical interpolation nodes; Segmenting the historical lane line with the plurality of historical interpolation nodes as endpoints to obtain a plurality of historical lane line segments, wherein the historical lane line segments include at least one of the historical nodes; Determine endpoint correction vectors corresponding to the two endpoints according to the position information of the two endpoints of each of the historical lane line segments and the position information of the reference vector nodes on the reference vector element closest to the two endpoints; Determine the total arc length of the historical lane line segment according to the position information of the two endpoints; Determine the arc length from the historical node to any endpoint according to the position information of any endpoint of the two endpoints and the position information of any historical node in the same historical lane line segment; Based on the endpoint correction vector, the total arc length, and the arc length from the historical node to any endpoint, determine the correction vector corresponding to the historical node, the reference vector element is a vectorized representation of the ground lane line matching the historical lane line, and the correction vector is used to indicate the correction direction and correction amplitude of the historical node; Based on the correction vectors corresponding to the multiple historical nodes, the position information of the multiple historical nodes is adjusted to achieve correction of the historical lane line.

2. The method according to claim 1, characterized in that The method further comprises: When the historical lane line is a double line in the actual line attribute, another historical lane line forming a double line with the historical lane line is obtained; Obtaining a vector element matching the other historical lane line; When the vector element and the reference vector element coincide with each other, no deviation correction is performed on the historical lane line.

3. The method according to claim 1, characterized in that: The method further comprises: In the case where the historical lane line is a solid and dotted line in the actual line attribute, for the endpoint of the historical lane line, when the perpendicular distance from the endpoint to the line connecting the two reference vector nodes closest to the reference vector element does not meet the distance condition, the two reference vector nodes corresponding to the endpoint are determined to be unreliable vector nodes; In the case that the two reference vector nodes are unreliable vector nodes, the position information of the endpoint is not adjusted.

4. The method according to claim 3, characterized in that After not adjusting the location information of the endpoint, the method further includes: Obtain other historical nodes adjacent to the endpoint on the historical lane line; Obtain the vector node closest to the other historical nodes; When the vector node closest to the other historical nodes belongs to the reference vector element, and the position information of the other historical nodes has been adjusted, the position information of the endpoint is adjusted based on the two reference vector nodes.

5. The method according to claim 1, characterized in that After the position information of the plurality of historical nodes is adjusted based on the deviation correction vectors corresponding to the plurality of historical nodes to correct the deviation of the historical lane line, the method further includes: Based on the predecessor-successor relationship between the historical lane lines after correction and the lane lines, the target virtual lane line to be corrected is determined; The corrected historical lane line is used as a position reference for the target virtual lane line, and the corrected target virtual lane line is corrected according to the corrected historical lane line correction method.

6. The method according to any one of claims 1 to 5, characterized in that After the position information of the plurality of historical nodes is adjusted based on the deviation correction vectors corresponding to the plurality of historical nodes to correct the deviation of the historical lane line, the method further includes: When there is an unengaged endpoint in any lane line after deviation correction, based on the predecessor and successor relationship between lane lines, the adjacent endpoint of the endpoint is obtained from the predecessor lane line or the successor lane line of the lane line to obtain an endpoint pair to be engaged including the endpoint and the adjacent endpoint; The position information of the adjacent endpoints is replaced with the position information of the endpoints to achieve the occlusion of the pair of endpoints to be occluded.

7. A lane deviation correction device, characterized in that: The device comprises: A determination module, used to perform node interpolation on multiple historical nodes on a historical lane line of actual line attributes to obtain multiple historical interpolation nodes; segment the historical lane line with multiple historical interpolation nodes as endpoints to obtain multiple historical lane line segments, wherein the historical lane line segments include at least one historical node; determine the endpoint correction vector corresponding to the two endpoints according to the position information of the two endpoints of each historical lane line segment and the position information of the reference vector node on the reference vector element closest to the two endpoints; determine the total arc length of the historical lane line segment according to the position information of the two endpoints; determine the arc length from the historical node to any endpoint according to the position information of any endpoint of the two endpoints and the position information of any historical node in the same historical lane line segment; determine the correction vector corresponding to the historical node based on the endpoint correction vector, the total arc length and the arc length from the historical node to any endpoint, wherein the reference vector element is a vectorized representation of a ground lane line matching the historical lane line, and the correction vector is used to indicate the correction direction and correction amplitude of the historical node; The adjustment module is used to adjust the position information of the multiple historical nodes based on the correction vectors corresponding to the multiple historical nodes to achieve correction of the historical lane lines.

8. An electronic device, characterized in that: The electronic device includes a processor and a memory, wherein at least one program code is stored in the memory, and the at least one program code is loaded and executed by the processor to implement the lane correction method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: At least one program code is stored in the storage medium, and the at least one program code is loaded and executed by the processor to implement the lane correction method as described in any one of claims 1 to 6.

Images