Lane-level center line acquisition method and device, equipment and medium

Abstract

The invention provides a lane-level center line acquisition method and device, equipment and a medium, and relates to the technical field of computers, in particular to the fields of intelligent traffic, automatic driving technologies and map navigation. According to the specific implementation scheme, a target high-precision track matched with a lane-level road chain is obtained from high-precision track data; segmenting the lane-level road chain according to a preset target distance to obtain a plurality of segmented sections; obtaining intersection points of the target high-precision track and each segmented section, and clustering the intersection points on each segmented section to obtain a clustering point set; according to attribute information of different clustering points in the clustering point set, selecting a target lane center point of each segmented section from the clustering point set; fusing the target lane center points of each segmented cross section to obtain a target center line of a lane-level road chain; and determining the center lines of other lanes on the lane-level link according to the target center line and the position relationship between different lanes on the lane-level link and the lane information.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, in particular to the fields of intelligent transportation, autonomous driving technology, and map navigation, and specifically to a lane-level centerline acquisition method, apparatus, device, and medium. Background Art

[0002] Map data can generally be divided into standard maps (SD Map), lane-level maps (LD Map) and high-precision maps (HD Map), and the accuracy requirements range from low to high.

[0003] Because real-world maps are constantly changing, and some lane-level maps have accuracy issues due to limitations in their production methods, finding a solution to improve lane-level map accuracy that balances accuracy, timeliness, and cost is a pressing issue. Summary of the Invention

[0004] The present disclosure provides a lane-level centerline acquisition method, apparatus, device, and medium.

[0005] According to one aspect of the present disclosure, a lane-level centerline acquisition method is provided, comprising:

[0006] Obtain the target high-precision trajectory that matches the lane-level road link from the high-precision trajectory data;

[0007] Segmenting the lane-level road link according to a preset target distance to obtain a plurality of segmented sections;

[0008] Obtaining the intersection points of the target high-precision trajectory and each segmented cross section, and clustering the intersection points on each segmented cross section to obtain a cluster point set;

[0009] selecting a target lane center point of each segmented cross section from the cluster point set according to attribute information of different cluster points in the cluster point set;

[0010] The target lane center points of each segmented section are merged to obtain the target center line of the lane-level link;

[0011] The center lines of other lanes on the lane-level link are determined based on the positional relationship between the target center line and different lanes on the lane-level link and lane information.

[0012] According to another aspect of the present disclosure, a lane-level centerline acquisition device is provided, comprising:

[0013] The acquisition module is used to obtain the target high-precision trajectory that matches the lane-level road link from the high-precision trajectory data;

[0014] A segmentation module, configured to segment the lane-level road link according to a preset target distance to obtain a plurality of segmented sections;

[0015] A clustering module is used to obtain the intersection points of the target high-precision trajectory and each segmented section, and cluster the intersection points on each segmented section to obtain a cluster point set;

[0016] A selection module, configured to select a target lane center point of each segmented cross section from the cluster point set according to attribute information of different cluster points in the cluster point set;

[0017] a fusion module, configured to fuse the target lane center points of each segmented section to obtain a target center line of the lane-level link;

[0018] A determination module is used to determine the center lines of other lanes on the lane-level road link based on the positional relationship between the target center line and different lanes on the lane-level road link and lane information.

[0019] According to another aspect of the present disclosure, there is provided an electronic device, comprising:

[0020] at least one processor; and

[0021] a memory communicatively connected to the at least one processor; wherein,

[0022] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the lane-level centerline acquisition method described in any embodiment of the present disclosure.

[0023] According to another aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable a computer to execute the lane-level centerline acquisition method described in any embodiment of the present disclosure.

[0024] According to another aspect of the present disclosure, a computer program product is provided, including a computer program, which, when executed by a processor, implements the lane-level centerline acquisition method according to any embodiment of the present disclosure.

[0025] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The accompanying drawings are used to better understand the present invention and do not constitute a limitation of the present invention.

[0027] Figure 1 is a schematic diagram of a lane-level centerline acquisition method according to an embodiment of the present disclosure;

[0028] Figure 2 is a schematic diagram of obtaining a high-precision trajectory in a lane-level centerline acquisition method according to an embodiment of the present disclosure;

[0029] Figure 3 is a schematic diagram of obtaining a target centerline in a lane-level centerline obtaining method according to an embodiment of the present disclosure;

[0030] Figure 4 is a schematic diagram of a lane-level centerline acquisition device according to an embodiment of the present disclosure;

[0031] Figure 5 3 is a block diagram of an electronic device used to implement the lane-level centerline acquisition method of an embodiment of the present disclosure. DETAILED DESCRIPTION

[0032] The following description of exemplary embodiments of the present disclosure is made in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding. These details should be considered as merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0033] Figure 1 This is a schematic diagram of a lane-level centerline acquisition method according to an embodiment of the present disclosure. This embodiment is applicable to the case where lane-level centerlines are acquired based on high-precision trajectory data to improve the accuracy of lane-level map data. It relates to the field of computer technology, and in particular to the fields of intelligent transportation, autonomous driving, and map navigation technology. The method can be executed by a lane-level centerline acquisition device, which is implemented in software and / or hardware, and is preferably configured in an electronic device, such as a computer device or a server. Figure 1 As shown, the method specifically includes the following:

[0034] S101. Obtain a target high-precision trajectory that matches a lane-level road link from high-precision trajectory data.

[0035] S102: Segment the lane-level road link according to a preset target distance to obtain multiple segmented sections.

[0036] S103 , obtaining the intersection points between the target high-precision trajectory and each segmented cross section, and clustering the intersection points on each segmented cross section to obtain a cluster point set.

[0037] S104 . Selecting a target lane center point of each segmented section from the cluster point set based on attribute information of different cluster points in the cluster point set.

[0038] S105: Fuse the target lane center points of each segmented section to obtain the target center line of the lane-level link.

[0039] S106. Determine the center lines of other lanes on the lane-level link based on the positional relationship between the target center line and different lanes on the lane-level link and the lane information.

[0040] Specifically, high-precision trajectory data refers to trajectory data in high-precision maps. High-precision trajectories within high-precision trajectory data offer advantages such as high timeliness, excellent positioning accuracy, and comprehensive coverage. Obtaining lane centerlines based on high-precision trajectories can effectively improve the accuracy of lane-level map data.

[0041] Lane-level links are generated based on the lane-level basemap of the lane-level map. Each lane-level link represents a road segment with specific attributes and coordinates, such as start and end coordinates and direction. By connecting consecutive links, a complete lane-level road path can be represented. Lane-level links are generated by structuring and discretizing the road information in the lane-level basemap, facilitating subsequent matching and analysis with high-precision trajectory data.

[0042] During specific implementation, for road scenarios with hard isolation, the roads on both sides of the hard isolation can be treated as lane-level links, that is, there is one lane-level link for each driving direction. In this case, the lane-level links in both directions can be treated as one-way roads. For road scenarios without hard isolation, the entire road can be treated as a lane-level link, or it can be divided according to the driving direction, and one or more lanes with the same driving direction on the road can be treated as a lane-level link. For example, the current road has four lanes and no hard isolation, of which there are two lanes for driving from south to north and two lanes for driving from north to south. In this case, the four lanes can be treated as a lane-level link as a whole, or the two lanes for driving from south to north can be treated as a lane-level link, and the two lanes for driving from north to south can be treated as a lane-level link. The disclosed embodiments do not impose any restrictions on this.

[0043] Obtaining a target high-precision trajectory that matches a lane-level link from high-precision trajectory data means that if a certain high-precision trajectory passes through a certain lane-level link in the high-precision trajectory data, then the high-precision trajectory is the target high-precision trajectory that matches the lane-level link. Each lane-level link has a corresponding target high-precision trajectory, and the number of target high-precision trajectories is at least one. In order to more accurately capture the shape of the trajectory, the lane-level link is segmented according to a preset target distance in the embodiment of the present disclosure to obtain multiple segmented sections. The target distance can be configured according to the road scene. The road scene is related to dimensions such as the straight distance of the road, whether it includes intersections, and the number of curves. For example, in a straight section that does not include an intersection, a larger target distance can be configured to segment the corresponding link, while in a section with many curves or near an intersection, a smaller target distance can be configured to segment the corresponding link. Dynamically adjusting the target distance based on different road scenes can capture the trajectory shape more accurately and finely. For example, the target distance can be configured to 10 meters.

[0044] Next, the intersection of the target high-precision trajectory and each segmented section is obtained, and the intersection points on each segmented section are clustered to obtain a set of clustered points. Among them, the intersection point represents the point where the object that generates the target high-precision trajectory enters each segmented section in its driving direction. For example, a vehicle is driving on a road and generates a target high-precision trajectory that matches the lane-level road link corresponding to the road. The point where the vehicle enters a segmented section of the lane-level road link in its driving direction is the intersection of the target high-precision trajectory and the segmented section. The intersection point is also the point where the target high-precision trajectory intersects with the segmented section for the first time. Since the number of target high-precision trajectories that match the lane-level road links is at least one, there is also at least one intersection point on each segmented section. It can be understood that the center point of the lane is included in these intersection points. Therefore, the intersection points on each segmented section continue to be clustered to obtain a set of clustered points.

[0045] In one embodiment, a first clustering algorithm is employed, using the number of lanes in each segmented section as the target number of clustering points. The intersection points on each segmented section are clustered to obtain a set of clustered points. For example, the first clustering algorithm may be a kmeans clustering algorithm. The location of the segment center of the segmented section within the lane-level base map can be determined, and the number of lanes on the road corresponding to that location within the base map can be obtained. This number of lanes is then used as the target number of clustered points. For example, if the number of lanes is three, then the clustered point set for the segmented section obtained through clustering will include three clustered points.

[0046] Furthermore, before performing clustering using the first clustering algorithm, a second clustering algorithm can be used to perform a second clustering on the intersection points on each segmented cross section to remove outliers from the intersection points. Then, the first clustering is performed on the intersection points after removing the outliers to improve clustering accuracy. The second clustering algorithm can be, for example, the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) clustering algorithm.

[0047] To further improve the accuracy of centerline acquisition, the disclosed embodiment selects the target lane center point for each segmented section from the cluster point set based on the attribute information of the different cluster points in the cluster point set. The target lane center points for each segmented section are then fused to obtain the target centerline for the lane-level link. The centerlines of the remaining lanes on the lane-level link are then determined based on the positional relationships between the target centerline and the lanes on the lane-level link, as well as the lane information.

[0048] In one embodiment, the attribute information of different cluster points in a cluster point set may include the lane scenario of the lane in which the cluster point is located, the location of the cluster point on the road, and the location of the cluster point on the road under different lane scenarios. Lane scenarios include at least parallel road scenarios and non-parallel road scenarios. A parallel road scenario refers to two adjacent lane-level links, one a main road and the other a secondary road. In this case, for the lane-level link of the secondary road, its rightmost cluster point can be selected as the cluster point with the highest confidence and used as the target lane center point for the segmented section. For the lane-level link of the main road, its leftmost cluster point can be selected as the target lane center point for the segmented section of the main road. For non-parallel road scenarios other than parallel road scenarios, the leftmost cluster point can also be selected as the target lane center point. The target lane center points of each segmented section are then fused to obtain the target centerline of the lane-level link. For example, when the leftmost cluster point is selected as the target lane center point, the target centerline obtained after fusion is the lane centerline of the leftmost lane of the lane-level link. The centerlines of other lanes on the lane-level link can be derived based on the positional relationships between the target centerline and the different lanes, as well as lane information. For example, this can be derived based on information such as the number of lanes, their positional relationships, and lane widths on the lane-level link contained in the map data.

[0049] In another embodiment, the attribute information of different cluster points in the cluster point set may further include the confidence of the cluster point. The confidence can be determined based on artificial intelligence technology, that is, a pre-trained model is used to score each cluster point, and the higher the score, the higher the confidence. Specifically, a classification model, such as xgboost (distributed gradient boosting library), can be trained. Using this classification model, the confidence of each cluster point is predicted based on the number of trajectory points of each cluster point, the number of points of adjacent cluster points, the distance from the adjacent cluster points to the current cluster point, and other information, so as to determine the lane and lane center point of the cluster point with the highest confidence based on its relative position.

[0050] Selecting cluster points for fusion based on lane scenarios or location information, or selecting cluster points with the highest confidence for fusion, can improve the accuracy of the obtained centerline and reduce the occurrence of centerline deviation, cornering, curvature errors, or unevenness.

[0051] The technical solution of the disclosed embodiment first obtains a target high-precision trajectory that matches the lane-level road link from the high-precision trajectory data, then segments the lane-level road link according to a preset target distance to obtain multiple segmented sections, then obtains the intersection points of the target high-precision trajectory and each segmented section, and clusters the intersection points on each segmented section to obtain a cluster point set. It can be seen that the disclosed embodiment implements joint clustering of intersection points on multiple lanes based on the number of lanes, then selects the cluster point with the highest confidence in the cluster point set as the target lane center point, and fuses the target lane center points of each segmented section to obtain the target centerline of the lane-level road link. Finally, the centerlines of other lanes on the lane-level road link are determined based on the positional relationship between the target centerline and the different lanes on the lane-level road link and the lane information. This ensures, on the one hand, the accuracy of the target centerline, and on the other hand, the consistency of the centerline morphology between lanes, thereby improving the accuracy of the lane-level centerline obtained based on the high-precision trajectory. This allows the lane-level map to be updated and corrected based on the obtained high-precision lane-level centerline, thereby improving the precision and accuracy of the lane-level map.

[0052] Figure 2 This is a schematic diagram of obtaining a high-precision track in the lane-level centerline acquisition method according to an embodiment of the present disclosure. This embodiment is further optimized based on the above embodiment. Figure 2 As shown, the method specifically includes the following:

[0053] S201. Obtain lane-level road links according to the lane-level base map.

[0054] S202: Retrieve high-precision trajectory data according to the coordinates of the lane-level road links to obtain a target high-precision trajectory that matches the lane-level road links.

[0055] Since the number of high-precision trajectories is too large, it is impossible to load all the trajectories for processing. In order to improve processing efficiency, the lane-level road links are first obtained according to the lane-level base map, and then the coordinates of the lane-level road links are used to retrieve the massive amount of high-precision trajectory data. Specifically, the position information of each high-precision trajectory can be compared with the coordinate range of the lane-level road link to determine whether the high-precision trajectory passes through a certain lane-level road link. If the point on the high-precision trajectory falls within the coordinate range of a certain lane-level road link, then this high-precision trajectory is the covering high-precision trajectory of the lane-level road link. In this way, the high-precision trajectories related to the lane-level road links can be quickly screened out, which greatly improves the efficiency and pertinence of data processing.

[0056] Exemplarily, any lane-level road link may refer to the road between two adjacent intersections. For roads without hard isolation, the center line of the lane-level road link corresponding to the entire road may be selected, and the coordinate range of the lane-level road link may be determined based on the area formed by extending a certain width to the left and right of the center line. For roads with hard isolation, the lane-level road links corresponding to roads with different driving directions on both sides of the hard isolation may be processed separately in the above manner to determine their respective coordinate ranges. Of course, for roads without hard isolation, each lane-level road link with different driving directions may also be processed separately. This disclosed embodiment does not impose any limitation on this, and may be configured as needed.

[0057] Accordingly, in the disclosed embodiment, for roads without hard barriers, the target high-precision trajectory can be acquired, segmented, and the intersection points can be obtained for each lane-level link of the road in each driving direction. The intersection points are then clustered, and the cluster point with the highest confidence is selected as the target lane center point. After fusing them, the target center line of each lane-level link in each driving direction can be obtained, that is, two target center lines can be obtained. To further ensure the consistency of each lane center line in the scenario and the accuracy of the lane center line, in one embodiment, these two target center lines can also be used as trajectories and segmented clustering can be performed again. In other words, the intersection points of each target center line and its corresponding lane-level link after segmentation are determined, and then the intersection points on each segmented section are clustered separately, in this case, only one center point is clustered. These obtained center points are then connected to obtain the road center line of the entire road without hard barriers. Finally, based on this road center line, the lane center lines of each driving direction on both sides are determined separately. Furthermore, for roads with hard barriers, each direction of travel can be treated as a lane-level link, configured as needed. In one embodiment, obtaining target high-precision trajectories matching lane-level links from high-precision trajectory data further includes filtering the target high-precision trajectories based on their trajectory information, retaining those that match lane-level links.

[0058] Specifically, high-precision trajectories can also contain abnormal trajectories, such as incorrectly bound trajectories or other cluttered trajectories caused by trajectory positioning issues. The retrieved target high-precision trajectories can be filtered based on dimensions such as trajectory angle, trajectory distance, and the relationship between different trajectories. This removes erroneous or cluttered trajectories and retains those that match lane-level links, ensuring the accuracy of subsequent centerline extraction.

[0059] Figure 3 FIG. 1 is a schematic diagram of obtaining a target centerline in a lane-level centerline obtaining method according to an embodiment of the present disclosure. This embodiment is further optimized based on the above embodiment. Figure 3 As shown, the method specifically includes the following:

[0060] S301: Fuse the target lane center points of each segmented section to obtain an initial center line.

[0061] S302: Perform a first smoothing process on the initial center line according to the road shape of the lane-level road link to obtain a first center line.

[0062] S303: Perform a second smoothing process on the first center line based on a smoothing algorithm to obtain a target center line of the lane-level link.

[0063] The target lane center points of each segmented section are fused together, for example, by connecting the center points to obtain the initial centerline. Considering that the initial centerline may have corners and other non-smoothness, centerline smoothing can be performed to improve accuracy.

[0064] First, based on the road morphology of the lane-level road link, the initial centerline is smoothed for the first time to obtain the first centerline. The road morphology of the lane-level road link refers to the original road morphology of the lane-level road link in the lane-level base map. The centerline of any road should be consistent with the morphology of the road to which it belongs. Therefore, smoothing the initial centerline based on the original road morphology can improve the accuracy of the initial centerline. Then, based on smoothing algorithms such as Douglas thinning and Bessel logic, the first centerline can be smoothed for the second time to obtain the target centerline of the lane-level road link. The second smoothing process can further eliminate the remaining non-smooth trajectory problem of the first centerline, so that the target centerline obtained after the two smoothing processes has higher accuracy.

[0065] The technical solution of the embodiment of the present disclosure first performs trajectory retrieval to retrieve the covering high-precision trajectory corresponding to the lane-level road link from the high-precision trajectory data. Then, trajectory filtering is performed to filter out the erroneous or messy trajectories in the covering high-precision trajectory, and retain the trajectory that matches the lane-level road link. Then, trajectory center point clustering is performed, and the lane-level road link is first fine-grained segmented to obtain segmented sections. Then, the intersection points of the high-precision trajectory and the segmented sections are clustered according to the number of lanes, and the cluster point with the highest confidence is selected as the lane center point according to the road scene. After the lane center points are fused, the lane centerline can be obtained, and the centerlines of the remaining lanes on the lane-level road link are derived based on the lane relationship. Therefore, on the one hand, the embodiment of the present disclosure can achieve more fine-grained trajectory segmentation, so that the trajectory shape can be better captured, providing a basis for obtaining high-precision centerlines; on the other hand, it can also ensure the consistency of the centerline shape between different lanes on the lane-level road link, further improving the accuracy of the lane centerline. Finally, the acquired lane-level centerline data is pushed to the operating platform. After a series of quality inspections and screening, it is updated to the master database of lane-level map data, improving the accuracy of the lane-level map.

[0066] Figure 4 This is a schematic diagram of a lane-level centerline acquisition device according to an embodiment of the present disclosure. This embodiment can be applied to the case where lane-level centerlines are acquired based on high-precision trajectory data to improve the accuracy of lane-level map data. It relates to the field of computer technology, and in particular to the field of intelligent transportation and autonomous driving technology. The device can implement the lane-level centerline acquisition method described in any embodiment of the present disclosure. Figure 4 As shown, the device 400 specifically includes:

[0067] An acquisition module 401 is configured to acquire a target high-precision trajectory matching a lane-level link from high-precision trajectory data;

[0068] A segmentation module 402 is configured to segment the lane-level road link according to a preset target distance to obtain a plurality of segmented sections;

[0069] The clustering module 403 is used to obtain the intersection points between the target high-precision trajectory and each segmented cross section, and cluster the intersection points on each segmented cross section to obtain a cluster point set;

[0070] A selection module 404 is configured to select a target lane center point of each segmented cross section from the cluster point set based on attribute information of different cluster points in the cluster point set;

[0071] A fusion module 405 is configured to fuse the target lane center points of each segmented section to obtain a target center line of the lane-level link;

[0072] The determination module 406 is configured to determine the center lines of other lanes on the lane-level link based on the positional relationship between the target center line and different lanes on the lane-level link and the lane information.

[0073] Optionally, the acquisition module 401 includes:

[0074] A generating unit, configured to obtain the lane-level road links according to a lane-level base map;

[0075] A retrieval unit is used to retrieve the high-precision trajectory data according to the coordinates of the lane-level road link to obtain a target high-precision trajectory that matches the lane-level road link.

[0076] Optionally, the acquisition module 401 further includes:

[0077] A filtering unit is configured to filter the target high-precision trajectory according to trajectory information of the target high-precision trajectory, wherein the trajectory information includes at least one of a trajectory angle, a trajectory distance, and an association relationship between different trajectories.

[0078] Optionally, the intersection point represents a point at which the object generating the target high-precision trajectory enters each segmented cross section in its driving direction.

[0079] Optionally, the clustering module 403 includes:

[0080] The first clustering unit is used to adopt a first clustering algorithm, take the number of lanes in each segmented section as the target number of clustering points, and perform first clustering on the intersection points on each segmented section to obtain a cluster point set.

[0081] Optionally, the clustering module 403 further includes:

[0082] The second clustering unit is used to use a second clustering algorithm to perform second clustering on the intersection points on each segmented cross section before the first clustering unit uses the first clustering algorithm to perform clustering, so as to remove outliers in the intersection points.

[0083] Optionally, the attribute information of different cluster points in the cluster point set includes at least one of the following: lane scene, location of cluster points and confidence of cluster points.

[0084] Optionally, the fusion module 405 includes:

[0085] A fusion unit, configured to fuse the target lane center points of each segmented section to obtain an initial center line;

[0086] a first smoothing unit, configured to perform a first smoothing process on the initial center line according to the road shape of the lane-level road link to obtain a first center line;

[0087] The second smoothing unit is configured to perform a second smoothing process on the first center line based on a smoothing algorithm to obtain a target center line of the lane-level link.

[0088] The above-mentioned product can execute the method provided by any embodiment of the present disclosure, and has the corresponding functional modules and beneficial effects of the execution method.

[0089] In the technical solutions disclosed herein, the collection, storage, use, processing, transmission, provision and disclosure of user personal information involved comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0090] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0091] Figure 5 A schematic block diagram of an example electronic device 500 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are provided as examples only and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0092] like Figure 5 As shown, the device 500 includes a computing unit 501, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 502 or a computer program loaded from a storage unit 508 into a random access memory (RAM) 503. Various programs and data required for the operation of the device 500 can also be stored in the RAM 503. The computing unit 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0093] Various components in device 500 are connected to I / O interface 505, including: an input unit 506, such as a keyboard, mouse, etc.; an output unit 507, such as various types of displays, speakers, etc.; a storage unit 508, such as a magnetic disk, optical disk, etc.; and a communication unit 509, such as a network card, modem, wireless communication transceiver, etc. The communication unit 509 allows device 500 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0094] The computing unit 501 can be any general-purpose and / or specialized processing component with processing and computing capabilities. Some examples of the computing unit 501 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various specialized artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 501 performs the various methods and processes described above, such as the lane-level centerline acquisition method. For example, in some embodiments, the lane-level centerline acquisition method can be implemented as a computer software program tangibly embodied in a machine-readable medium, such as the storage unit 508. In some embodiments, part or all of the computer program can be loaded and / or installed onto the device 500 via the ROM 502 and / or the communication unit 509. When the computer program is loaded into the RAM 503 and executed by the computing unit 501, one or more steps of the lane-level centerline acquisition method described above can be performed. Alternatively, in other embodiments, the computing unit 501 can be configured to perform the lane-level centerline acquisition method by any other suitable means (e.g., via firmware).

[0095] Various embodiments of the systems and techniques described above can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), system-on-chip systems (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include being implemented in one or more computer programs that are executable and / or interpreted on a programmable system that includes at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0096] The program code for implementing the method of the present disclosure can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device so that when the program code is executed by the processor or controller, the functions / operations specified in the flow chart and / or block diagram are implemented. The program code can be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0097] In the context of the present disclosure, a machine-readable medium can be a tangible medium that can contain or store a program for use by or in conjunction with an instruction execution system, device or equipment. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium can include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0098] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0099] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with embodiments of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0100] A computer system may include a client and a server. The client and server are typically remote from each other and typically interact via a communication network. This client-server relationship arises through computer programs running on the respective computers, creating a client-server relationship. The server may be a cloud server, also known as a cloud computing server or cloud host, a hosting product within the cloud computing service ecosystem that addresses the management difficulties and limited scalability of traditional physical hosting and VPS services. The server may also be a server in a distributed system or a server integrated with blockchain.

[0101] Artificial intelligence (AI) is the study of how computers can simulate certain human thought processes and intelligent behaviors (such as learning, reasoning, thinking, and planning). It encompasses both hardware and software technologies. AI hardware technologies generally include sensors, specialized AI chips, cloud computing, distributed storage, and big data processing. AI software technologies primarily encompass computer vision, speech recognition, natural language processing, machine learning / deep learning, big data processing, and knowledge graphs.

[0102] Cloud computing refers to a technology system that provides network access to elastically scalable shared pools of physical or virtual resources. These resources can include servers, operating systems, networks, software, applications, and storage devices, and can be deployed and managed on-demand in a self-service manner. Cloud computing technology provides efficient and powerful data processing capabilities for the application of technologies such as artificial intelligence and blockchain, as well as for model training.

[0103] It should be understood that the various forms of the processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this disclosure can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions provided by this disclosure can be achieved. This is not a limitation herein.

[0104] The above specific embodiments do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure shall be included within the scope of protection of this disclosure.

Claims

1. A lane-level centerline acquisition method, comprising: Obtain the target high-precision trajectory that matches the lane-level road link from the high-precision trajectory data; Segmenting the lane-level road link according to a preset target distance to obtain a plurality of segmented sections; Obtaining the intersection points of the target high-precision trajectory and each segmented cross section, and clustering the intersection points on each segmented cross section to obtain a cluster point set; selecting a target lane center point of each segmented cross section from the cluster point set according to attribute information of different cluster points in the cluster point set; The target lane center points of each segmented section are merged to obtain the target center line of the lane-level link; The center lines of other lanes on the lane-level link are determined based on the positional relationship between the target center line and different lanes on the lane-level link and lane information.

2. The method according to claim 1, wherein The step of obtaining a target high-precision trajectory matching a lane-level link from the high-precision trajectory data includes: Acquire the lane-level road links according to the lane-level base map; The high-precision trajectory data is retrieved according to the coordinates of the lane-level road link to obtain a target high-precision trajectory that matches the lane-level road link.

3. The method according to claim 2, wherein: The step of obtaining a target high-precision trajectory matching a lane-level link from the high-precision trajectory data further includes: The target high-precision trajectory is filtered according to trajectory information of the target high-precision trajectory, wherein the trajectory information includes at least one of a trajectory angle, a trajectory distance, and an association relationship between different trajectories.

4. The method according to claim 1, wherein The intersection point represents a point at which the object generating the target high-precision trajectory enters each segmented cross-section in its driving direction.

5. The method according to claim 1, wherein The clustering of the intersection points on each segmented cross section to obtain a cluster point set includes: A first clustering algorithm is adopted, the number of lanes in each segmented section is used as the target number of clustering points, and the first clustering is performed on the intersection points on each segmented section to obtain a cluster point set.

6. The method according to claim 5, wherein: The clustering of the intersection points on each segmented cross section to obtain a cluster point set further includes: Before adopting the first clustering algorithm to perform clustering, a second clustering algorithm is adopted to perform second clustering on the intersection points on each segmented cross section, and outliers in the intersection points are removed.

7. The method according to claim 1, wherein The attribute information of different cluster points in the cluster point set includes at least one of the following: lane scene, location of the cluster point, and confidence of the cluster point.

8. The method according to claim 1, wherein The step of fusing the target lane center points of each segmented section to obtain the target center line of the lane-level link includes: The target lane center points of each segmented section are merged to obtain an initial center line; performing a first smoothing process on the initial center line according to the road shape of the lane-level road link to obtain a first center line; Based on a smoothing algorithm, a second smoothing process is performed on the first center line to obtain a target center line of the lane-level link.

9. A lane-level centerline acquisition device, comprising: The acquisition module is used to obtain the target high-precision trajectory matching the lane-level road link from the high-precision trajectory data; A segmentation module, configured to segment the lane-level road link according to a preset target distance to obtain a plurality of segmented sections; A clustering module is used to obtain the intersection points of the target high-precision trajectory and each segmented section, and cluster the intersection points on each segmented section to obtain a cluster point set; A selection module, configured to select a target lane center point of each segmented cross section from the cluster point set according to attribute information of different cluster points in the cluster point set; a fusion module, configured to fuse the target lane center points of each segmented section to obtain a target center line of the lane-level link; A determination module is used to determine the center lines of other lanes on the lane-level road link based on the positional relationship between the target center line and different lanes on the lane-level road link and lane information.

10. An electronic device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the lane-level centerline acquisition method according to any one of claims 1 to 8.

11. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to enable the computer to execute the lane-level centerline acquisition method according to any one of claims 1-8.

12. A computer program product comprising a computer program / instruction, which, when executed by a processor, implements the lane-level centerline acquisition method according to any one of claims 1-8.

Images