Lane line reconstruction method and device of autonomous vehicle, vehicle and medium

Abstract

This invention discloses a lane line reconstruction method, device, vehicle, and medium for autonomous vehicles, relating to the field of intelligent driving vehicle technology. The method includes: collecting environmental perception data corresponding to the target vehicle based on intelligent driving sensors installed on the target vehicle; generating a lane line attribute dataset corresponding to the driving road based on the environmental perception data; determining the lane line type corresponding to each lane line of the driving road based on the lane line attribute dataset; and reconstructing the target lane lines based on the lane line types and the lane line attribute dataset. This technical solution can improve the accuracy and completeness of lane line data, avoid path deviation problems caused by lane line data defects, and ensure the stability of the vehicle's driving trajectory.

Description

Technical Field

[0001] This invention relates to the field of intelligent driving vehicle technology, and in particular to a lane line reconstruction method, device, vehicle, and medium for autonomous driving vehicles. Background Technology

[0002] Currently, there is a significant technological gap in the field of intelligent driving, specifically addressing lane line perception quality optimization. As intelligent driving algorithms evolve towards L3 and higher levels of autonomous driving, and as the hardware performance of onboard intelligent driving sensors (such as LiDAR, high-definition cameras, and millimeter-wave radar) upgrades, the driving scenarios that intelligent driving systems need to cover have expanded from conventional highways / urban arterial roads to extreme and long-tail scenarios such as complex intersections, construction zones, and low-visibility conditions in rainy or foggy weather.

[0003] However, existing technologies often suffer from coordinate deviations, discontinuities, and missing semantic information in the original lane line data output by onboard sensors under complex non-standard road environments (such as lane line wear, obstruction, and overlap of temporary and original lane markings). Since lane lines are a core input parameter for the environmental perception and path planning modules in intelligent driving systems, their quality defects directly lead to decreased path planning accuracy, delayed decision response, and even deviations from the expected driving trajectory, seriously affecting the functional safety and driving stability of the intelligent driving system. Summary of the Invention

[0004] This invention provides a lane line reconstruction method, device, vehicle, and medium for autonomous vehicles. By determining the lane line attribute dataset based on environmental perception data and classifying lane line types, lane line reconstruction is performed based on the classified lane line types and lane line attribute datasets. This improves the accuracy and completeness of lane line data, avoids path deviation problems caused by lane line data defects, and ensures the stability of the vehicle's driving trajectory.

[0005] According to one aspect of the present invention, a lane line reconstruction method for an autonomous vehicle is provided, comprising:

[0006] Based on intelligent driving sensors installed on the target vehicle, environmental perception data corresponding to the target vehicle is collected;

[0007] Generate a lane line attribute dataset corresponding to the driving road based on environmental perception data;

[0008] Determine the lane line type corresponding to each lane line of the driving road based on the lane line attribute dataset;

[0009] The target lane line is obtained by reconstructing the lane line based on the lane line type and lane line attribute dataset.

[0010] According to another aspect of the present invention, a lane line reconstruction apparatus for an autonomous vehicle is provided, comprising:

[0011] The data acquisition module is used to collect environmental perception data corresponding to the target vehicle based on the intelligent driving sensors installed on the target vehicle;

[0012] The dataset generation module is used to generate a dataset of lane line attributes corresponding to the driving road based on environmental perception data;

[0013] The lane line type classification module is used to determine the lane line type corresponding to each lane line of the driving road based on the lane line attribute dataset.

[0014] The lane line reconstruction module is used to reconstruct the target lane lines on the driving road based on the lane line type and lane line attribute dataset.

[0015] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0016] At least one processor;

[0017] and memory that is communicatively connected to at least one processor;

[0018] The memory stores a computer program that can be executed by at least one processor, which enables the at least one processor to execute the lane line reconstruction method for an autonomous vehicle according to any embodiment of the present invention.

[0019] According to another aspect of the present invention, a computer-readable storage medium is provided, which stores computer instructions for causing a processor to execute and implement the lane line reconstruction method for an autonomous vehicle according to any embodiment of the present invention.

[0020] The technical solution of this invention involves collecting environmental perception data corresponding to the target vehicle using intelligent driving sensors installed on the target vehicle; generating a lane line attribute dataset corresponding to the driving road based on the environmental perception data; determining the lane line type corresponding to each lane line of the driving road based on the lane line attribute dataset; and reconstructing the lane lines of the driving road based on the lane line types and the lane line attribute dataset to obtain the target lane lines. Based on this technical solution, by determining the lane line attribute dataset based on the environmental perception data and classifying lane line types, and then reconstructing lane lines based on the classified lane line types and the lane line attribute dataset, the accuracy and completeness of lane line data are improved, path deviation problems caused by lane line data defects are avoided, and the stability of the vehicle's driving trajectory is ensured.

[0021] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a flowchart of a lane line reconstruction method for an autonomous vehicle provided in an embodiment of the present invention;

[0024] Figure 2 This is a flowchart of a lane line reconstruction method for an autonomous vehicle provided in an embodiment of the present invention;

[0025] Figure 3 This is a schematic diagram of the lane line distribution of a driving road provided in an embodiment of the present invention;

[0026] Figure 4 A schematic diagram of the structure of a lane line reconstruction device for an autonomous vehicle provided in an embodiment of the present invention;

[0027] Figure 5 This is a schematic diagram of the structure of the electronic device provided in an embodiment of the present invention. Detailed Implementation

[0028] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0029] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0030] Figure 1 This is a flowchart illustrating a lane line reconstruction method for an autonomous vehicle according to an embodiment of the present invention. This embodiment is applicable to situations where lane lines are reconstructed based on collected environmental perception data during the operation of an autonomous vehicle. This method can be executed by a lane line reconstruction device for the autonomous vehicle, which can be implemented in hardware and / or software and can be configured in an electronic device. Figure 1 As shown, the method specifically includes the following steps:

[0031] S110. Collect environmental perception data corresponding to the target vehicle based on the intelligent driving sensors installed on the target vehicle.

[0032] The target vehicle is a passenger car equipped with an intelligent driving system, which carries perception sensors and performs autonomous driving-related operations. The intelligent driving sensors can be multimodal perception arrays that acquire lane markings and surrounding environmental information about the road surface. Environmental perception data can be understood as raw road environment data collected by sensors.

[0033] Specifically, the system collects environmental perception data corresponding to the target vehicle based on intelligent driving sensors installed on the target vehicle. For example, a LiDAR and intelligent driving camera can be deployed in the front-view area on the roof of the target vehicle, and a millimeter-wave radar can be integrated on the inside of the front bumper to form a multimodal intelligent driving sensor array. Each sensor works together according to a preset field of view and detection range to collect road environment perception data in real time during the vehicle's driving process. The system focuses on capturing the geometric shape and semantic attributes of lane lines, while also collecting information related to the surrounding environment such as lane boundary obstacles and road wear conditions, ensuring stable data collection in extreme scenarios such as rain, fog, and strong light.

[0034] It should be noted that a multimodal perception array is formed by deploying LiDAR (point cloud density ≥128 lines, ranging accuracy ≤±2cm), intelligent driving camera (frame rate ≥30fps, supports HDR high dynamic range), and 77GHz millimeter-wave radar (detection distance 0-200m). The LiDAR and camera are installed in the front-view area on the roof (horizontal field of view 120°-150°), while the millimeter-wave radar is integrated into the inside of the front bumper. The sensors collect road environment data in real time, focusing on capturing the geometric shape (position, width, curvature), semantic attributes (solid / dashed lines, color), and surrounding environmental information (such as lane boundary obstacles, road wear status) of lane lines, ensuring stable data collection capabilities even in extreme scenarios such as rain, fog, strong light, and tunnels.

[0035] This invention achieves complementary acquisition of different types of perception data through a multimodal sensor array, making up for the shortcomings of a single sensor in complex scenarios; the sensors are deployed in fixed positions and work together to ensure the comprehensiveness and real-time nature of environmental perception data acquisition; the accurate capture of the core attributes of lane lines and the surrounding environment provides complete and reliable raw data for the subsequent analysis and reconstruction of lane line information, laying a solid foundation for environmental perception in high-level autonomous driving.

[0036] S120. Generate a lane line attribute dataset corresponding to the driving road based on environmental perception data.

[0037] The driving road can be any road scenario the target vehicle is currently traveling on, including regular main roads, complex intersections, and construction zones. The lane line attribute dataset can be a standardized set of core lane line parameters used to represent the geometric shape and semantic attributes of lane lines. Lane line attributes can include geometric and semantic parameters, serving as the basis for lane line state judgment and reconstruction.

[0038] Specifically, generating a lane line attribute dataset corresponding to the driving road based on environmental perception data can be achieved by using a deep learning perception algorithm with an improved LaneNet semantic segmentation model. This algorithm fuses and analyzes environmental perception data collected from multiple sources, extracting pixel-level features of lane lines in the image, 3D coordinate features of LiDAR point clouds, and boundary reflection signal features of millimeter-wave radar. From these, core lane line attribute parameters are selected and quantized to form a standardized parameter set containing start-point intercept, slope, curvature, rate of change of curvature, length coefficient, and type identifier. After integration according to a preset data format, a lane line attribute dataset matching the current driving road is generated.

[0039] It should be noted that the raw sensor data is processed using a deep learning-based lane perception algorithm (such as the improved LaneNet semantic segmentation model). This involves fusing and analyzing the pixel-level features of lane lines in the image, the 3D coordinate features of the LiDAR point cloud, and the boundary reflection signal features of the millimeter-wave radar to extract core lane line attribute parameters (including the starting point intercept coefficient C0, starting point slope C1, curvature C2, rate of change of curvature C3, length coefficient L, and type identifier T), forming a standardized set of lane line attribute coefficients. This coefficient set is transmitted to the downstream lane line information judgment module via in-vehicle Ethernet (supporting ETH gigabit bandwidth) according to the ISO 21784 intelligent driving data interaction protocol, ensuring real-time data transmission (latency ≤10ms) and compatibility.

[0040] The technical solution of this invention compensates for the information shortcomings of single sensor data through multi-feature fusion analysis, ensuring the comprehensiveness and accuracy of lane line attribute extraction.

[0041] S130. Determine the lane line type corresponding to each lane line of the driving road based on the lane line attribute dataset.

[0042] Among them, lane lines can be all lane lines to be determined within the driving road, which can be understood as the four boundary lane lines corresponding to the vehicle and the three adjacent lanes.

[0043] Specifically, based on the lane line attribute dataset, the lane line type corresponding to each lane line on the driving road is determined. By retrieving the lane line attribute datasets corresponding to all lane lines to be determined on the driving road, the integrity of the core attribute coefficients of each lane line is checked to determine the overall data integrity status. If the data is partially complete, it is marked as partially missing; if the data is completely missing, it is marked as completely missing. When the data is completely complete, the geometric relationship between lane lines is analyzed based on the attribute dataset to determine whether it is parallel or non-parallel. At the same time, the lateral distance between the target vehicle and the lane line is calculated by combining environmental perception data. If the vehicle's driving position is close to the edge of the lane line, it is directly marked as a narrow-space lane, thus completing the type matching and marking of each lane line within the driving road.

[0044] S140. Based on the lane line type and lane line attribute dataset, the driving road lane lines are reconstructed to obtain the target lane lines.

[0045] The lane line type can be a lane line status label based on coefficient verification and geometric relationship determination, including five categories: parallel, non-parallel, partially missing, completely missing, and narrow-spaced lanes. The target lane line can be understood as lane line data that has been reconstructed and optimized according to the scenario, and is used to provide lane line input for intelligent driving decision-making and planning.

[0046] Specifically, the target lane lines are obtained by reconstructing lane lines based on lane line type and lane line attribute dataset. For example, according to the lane line type of the driving road, the corresponding scenario-based reconstruction strategy is matched, and the reconstruction operation is performed in combination with the core coefficients of the lane line attribute dataset. For parallel types, the original lane line data is directly used; for non-parallel types, the lane line closest to the vehicle is selected as the axle lane line, the C0 coefficient of other lane lines is retained, and the non-C0 coefficient of the axle lane line is assigned; for partially missing types, the missing C0 is filled in by predicting the historical C0 and lane width coefficient, and then the axle lane line coefficient is shifted; for all missing types, the coefficients are derived based on the vehicle trajectory and steering wheel angle, and the lane lines are virtually generated; for narrow lane types, the near axle lane line is retained, the C0 of the opposite lane line is set to the minimum passage value, and other coefficients of the axle lane line are reused. After reconstruction, the lane lines are also fine-tuned and optimized by lengthening and high-speed extension, and finally the target lane lines are obtained.

[0047] It should be noted that the reconstructed lane line data is standardized, encapsulated, verified, and stored before being output to backend decision planning, chassis control, and other modules. The reconstructed lane line attribute coefficient set (including C0, C1, C2, C3, L, T, and status flags) is encapsulated according to the Protobuf data serialization protocol, and a structured data format is defined (including metadata such as lane line ID, lane number, valid timestamp, and confidence level (≥95%)) to ensure interface compatibility with downstream modules. The encapsulated reconstructed data is then sent to the backend decision planning, chassis control, and other modules.

[0048] The technical solution of this invention achieves scenario-based and adaptive lane line reconstruction by matching a dedicated reconstruction strategy according to the lane line type. It accurately adapts to different road driving conditions, ensures the effectiveness and continuity of lane lines, and the output target lane line can accurately match the decision-making and planning needs of advanced intelligent driving, reduce the risk of path deviation, and improve driving stability.

[0049] In some embodiments, when the lane line type is non-parallel, the target lane line is obtained by reconstructing the lane lines of the driving road based on the lane line type and the lane line attribute dataset. This includes: determining the lateral distance between the target vehicle and the lane line of the driving lane based on environmental perception data, and determining the axle lane line from the lane line of the driving lane based on the lateral distance; obtaining the lane line attribute dataset of the axle lane line, and reconstructing the lane lines of the driving road based on the lane line attribute dataset of the axle lane line to obtain the target lane line.

[0050] In this context, "non-parallel" can refer to a situation where the coefficients of the four lane lines corresponding to a three-lane road are complete, but their geometric relationship is not parallel. Lateral distance can be understood as the lateral distance between the target vehicle's body and the lane lines on both sides of the driving lane, used to determine the relative position of the vehicle and the lane lines. The axle lane line can be a selected baseline lane line in a non-parallel scenario, used to provide coefficient references for reconstructing other lane lines.

[0051] Specifically, when the lane line type is non-parallel, the lateral position information of the target vehicle and the lane lines on both sides of the driving lane is extracted from the environmental perception data, and the corresponding lateral distance is calculated. The lane line with the closer lateral distance to the target vehicle is selected as the axle lane line. The complete lane line attribute dataset of this axle lane line is retrieved, the C0 coefficients of other non-parallel lane lines are retained, and all non-C0 coefficients of the axle lane line, such as C1, C2, C3, L, and T, are assigned to other lane lines. A lane line parallel to the axle lane line is constructed by coefficient translation, and the reconstructed target lane line is obtained after verification.

[0052] The technical solution of this invention selects the axle lane line based on lateral distance, so that the reconstruction benchmark fits the actual driving position of the vehicle, which improves the scene adaptability of lane line reconstruction and provides accurate lane line input for path planning, avoiding the problem of driving trajectory deviation caused by non-parallel lane lines.

[0053] In some embodiments, when the lane line type is partially missing, lane line reconstruction is performed on the driving road based on the lane line type and lane line attribute dataset to obtain the target lane line. This includes: selecting lane lines with complete lane line attribute datasets as axis lane lines, and obtaining historical lane line attribute data for lane lines with missing data, wherein the historical lane line attribute data includes historical lateral distance parameters and historical lane width parameters; calculating the lateral distance prediction parameters for lane lines with missing data based on the historical lane line attribute data; completing the lane line attribute dataset for lane lines with missing data based on the axis lane line attribute dataset and the lateral distance prediction parameters; and reconstructing the driving road based on the completed lane line attribute dataset to obtain the target lane line.

[0054] The historical lane line attribute data includes historical lateral distance parameters and historical lane width parameters. Partially missing data can refer to a situation where the coefficients of the four lane lines corresponding to three lanes are missing and unassigned. The historical lateral distance parameter can be understood as the valid C0 coefficient of the missing lane line in one frame, used as the basis for lateral distance prediction. The historical lane width parameter can be the lane width coefficient during past driving processes, which can be used to assist in estimating the lateral distance. The lateral distance prediction parameter can be understood as the calculated predicted C0 value of the missing lane line.

[0055] Specifically, when the lane line type is partially missing, the system first selects lane lines with complete coefficients from the lane line attribute dataset as axle lane lines. Simultaneously, it retrieves historical lane line attribute data for the missing lane lines, extracting historical lateral distance and lane width parameters. By summing the coefficients and combining these two types of historical parameters, the system derives the predicted lateral distance parameter for the missing lane line. The non-C0 coefficients of the axle lane lines are retained, and the predicted parameter is assigned the C0 coefficient of the missing lane line. This completes the lane line attribute dataset for the missing lane lines. Finally, based on the completed lane line attribute dataset, the system reconstructs the lane lines to obtain the target lane line.

[0056] The technical solution of this invention provides a reliable parameter reference for filling in missing lane lines by selecting complete coefficient lane lines as axial lane lines, ensuring the consistency of reconstruction. By combining historical lateral distance parameters and historical lane width parameters to calculate predicted values, the reconstructed target lane line parameters are complete and the geometric relationship is standardized, providing continuous and accurate lane line input for intelligent driving path planning and improving the system's adaptability in scenarios where lane lines are partially missing.

[0057] In some embodiments, when the lane line type is completely missing, the target lane line is obtained by reconstructing the lane lines on the driving road based on the lane line type and the lane line attribute dataset. This includes: determining the actual driving trajectory corresponding to the target vehicle based on environmental perception data and obtaining a preset fixed width value; deriving a virtual lane line attribute dataset based on the preset fixed width value and the actual driving trajectory; generating a virtual lane line based on the virtual lane line attribute dataset and using the virtual lane line as the target lane line.

[0058] The "completely missing" type refers to a laneless road segment where all four lane line coefficients for a three-lane road are unassigned. The actual driving trajectory can be the current path data of the target vehicle. The preset fixed width value can be understood as the preset parameter for the standard lane width of the road. The virtual lane line attribute dataset can be the derived set of core lane line coefficients.

[0059] Specifically, when the lane line type is completely missing, the driving path information of the target vehicle is extracted from the environmental perception data to determine the actual driving trajectory of the vehicle. At the same time, the preset standard road lane fixed width value is retrieved. With the actual driving trajectory as the center reference, the lateral boundary of the lane is defined by combining the preset fixed width value. Then, the slope, curvature, and curvature change rate of the lane line are derived from the steering wheel angle data of the target vehicle. All coefficients are integrated to form a standardized virtual lane line attribute dataset. Based on this dataset, a virtual four-lane containing four boundary lines is constructed, and the generated virtual lane line is directly determined as the target lane line.

[0060] The technical solution of this invention generates virtual lane lines based on the actual driving trajectory of the vehicle, allowing the reconstructed lane lines to conform to the actual driving state of the vehicle and adapt to road scenarios without lane lines. By combining preset fixed width values ​​and steering wheel angle derivation coefficients, the geometric regularity and rationality of the virtual lane lines are ensured, meeting the actual requirements of road driving. The generation of virtual four lanes fills the perception gap in laneless road sections, providing effective and stable lane line input for intelligent driving path planning, avoiding decision-making inaccuracies caused by the complete absence of lane lines, and significantly improving the system's adaptability and driving safety in extreme unmarked scenarios.

[0061] In some embodiments, when the lane line type is a narrow-space lane type, the target lane line is obtained by reconstructing the lane lines of the driving road based on the lane line type and the lane line attribute dataset, including: determining the lateral distance between the target vehicle and the lane line of the driving lane based on environmental perception data, and determining the axle lane line from the lane line of the driving lane based on the lateral distance; and reconstructing the lane lines of the driving road based on the minimum lateral distance parameter and the lane line attribute dataset of the axle lane line to obtain the target lane line.

[0062] Among them, the narrow-spaced lane type vehicle refers to a special operating condition where the target vehicle is close to the edge of the lane line while driving. The minimum lateral distance parameter can be the minimum lateral spacing value of the lane lines used to ensure the normal passage of the target vehicle. The axle lane line can be the lane line on the side closest to the target vehicle.

[0063] Specifically, when the lane type is a narrow-space lane, the position information of the target vehicle and the lane lines on both sides of the driving lane is extracted from the environmental perception data, and the corresponding lateral distances are calculated. The lane line closer to the vehicle is identified as the axle lane line, and its complete lane line attribute dataset is retrieved. All attribute parameters of this axle lane line are retained, and the lateral distance parameter of the lane line on the other side of the driving lane is assigned a preset minimum lateral distance parameter. The remaining attribute coefficients are kept consistent with the axle lane line, and invalid parameters of the original other lane line are discarded. In this way, the lane line reconstruction in the narrow-space scenario is completed, and the target lane line adapted for vehicle avoidance and passage is obtained.

[0064] The technical solution of this invention selects the lane line closest to the vehicle as the axis lane line based on the lateral distance, allowing the reconstruction benchmark to fit the actual driving position of the vehicle and adapt to obstacle avoidance scenarios within the lane. The setting of the minimum lateral distance parameter ensures both the normal passage space of the vehicle and the accurate reconstruction of the narrow-distance lane line, avoiding interference from invalid road areas. The reconstruction method of reusing the axis lane line coefficient is simple and efficient, ensuring the real-time processing needs of the intelligent driving system. The reconstructed target lane line is adapted to narrow-distance avoidance scenarios, providing accurate lane boundary references for path planning and improving the driving safety and adaptability of the system in scenarios where the lane is partially obscured.

[0065] The technical solution of this invention involves collecting environmental perception data corresponding to the target vehicle using intelligent driving sensors installed on the target vehicle; generating a lane line attribute dataset corresponding to the driving road based on the environmental perception data; determining the lane line type corresponding to each lane line of the driving road based on the lane line attribute dataset; and reconstructing the lane lines of the driving road based on the lane line types and the lane line attribute dataset to obtain the target lane lines. Based on this technical solution, by determining the lane line attribute dataset based on the environmental perception data and classifying lane line types, and then reconstructing lane lines based on the classified lane line types and the lane line attribute dataset, the accuracy and completeness of lane line data are improved, path deviation problems caused by lane line data defects are avoided, and the stability of the vehicle's driving trajectory is ensured.

[0066] In some embodiments, Figure 2 A flowchart of a lane line reconstruction method for an autonomous vehicle provided in an embodiment of the present invention is shown below. Figure 2 As shown, this embodiment determines the lane line type corresponding to each lane line of the driving road based on the lane line attribute dataset, including:

[0067] S210. Determine the lane line to be judged corresponding to the target vehicle based on environmental perception data.

[0068] The lane lines to be determined include at least four lane lines. These four lane lines are the four boundary lane lines corresponding to the lane where the target vehicle is located and the three adjacent lanes on both sides. Figure 3 As shown, these represent the left boundary of the left lane, the right boundary of the left lane / left boundary of the vehicular lane, the right boundary of the vehicular lane / left boundary of the right lane, and the right boundary of the right lane, respectively. The lane lines to be determined can be a set of road lane lines extracted from environmental perception data that need to be classified and labeled.

[0069] Specifically, from the environmental perception data collected by multi-source sensors, the geometric shape and semantic attribute information of the lane lines of the target vehicle's driving road are extracted. Taking the current lane of the target vehicle as the core, the detection range is expanded to the adjacent left and right lanes to form a three-lane detection range. From this range, the corresponding four boundary lane lines are selected to form a set of lane lines to be judged containing four or more lane lines.

[0070] S220. Determine the data integrity status of the lane line to be judged based on the lane line attribute dataset.

[0071] The data integrity status can refer to the assignment of the core attribute coefficients of the lane line to be determined, including three categories: fully complete, partially missing, and fully missing. The core attribute coefficients can be understood as the key parameters in the lane line attribute dataset, including the starting intercept, slope, curvature, rate of change of curvature, length coefficient, and type identifier, which are used to fully represent the geometric and semantic features of the lane line.

[0072] Specifically, the data integrity status of the lane lines to be judged is determined based on the lane line attribute dataset. For example, by retrieving the lane line attribute dataset corresponding to the lane line to be judged, the integrity of the core attribute coefficients of each lane line is checked, verifying whether key parameters such as C0, C1, C2, C3, L, and T are unassigned. The coefficient assignment results of the four lane lines to be judged are statistically analyzed. If the core coefficients of all lane lines are assigned values, the data is considered to be completely complete; if some lane lines have unassigned coefficients, the data is considered partially missing; if the core coefficients of all lane lines are unassigned, the data is considered completely missing. This is used to determine the final data integrity status of the lane lines to be judged.

[0073] S230. Determine the lane line type for each lane line to be judged based on the data integrity status and lane line attribute dataset.

[0074] The lane line types include parallel, non-parallel, partially missing, completely missing, and narrow-spaced lane types. Lane line type can be understood as a comprehensive status label of the lane line to be determined, combining data integrity, geometric relationships, and the relative position of the vehicle. A parallel lane line can be a lane line state where all data is complete and the four lane lines are geometrically parallel. A non-parallel lane line can be understood as a lane line state where all data is complete but the four lane lines are not geometrically parallel.

[0075] Specifically, the lane type of each lane to be determined is determined based on the data integrity status and lane attribute dataset. For example, a preliminary classification can be made based on the data integrity status of the lane to be determined: partially missing data is marked as partially missing, and completely missing data is marked as completely missing. When all data is complete, the geometric relationship of the four lanes is analyzed through the lane attribute dataset. If they are parallel, they are marked as parallel; if they are not parallel, they are marked as non-parallel. At the same time, the lateral distance between the target vehicle and the lane is calculated by combining environmental perception data. If the vehicle is close to the edge of the lane, it is directly marked as a narrow-space lane. Finally, the type classification of all lanes to be determined is completed.

[0076] In some embodiments, determining the lane type of each lane to be determined based on the data integrity status and the lane attribute dataset includes: when the lane attribute datasets of the lanes to be determined are all in a complete state, determining the parallel type between each lane to be determined based on the lane attribute datasets; when the lane attribute datasets of the lanes to be determined are only partially in a complete state, determining the lane type as partially missing; when the lane attribute datasets of the lanes to be determined are all missing, determining the lane type as completely missing; and determining the relative position of the target vehicle and the lanes to be determined based on environmental perception data, and determining the lane type as a narrow-space lane when the target vehicle's driving position is close to the edge of the lane.

[0077] Among them, parallel type can be the lane line subdivision type when the data is complete, including parallel and non-parallel types. Relative position can be the spatial relationship between the target vehicle body and the lane line to be judged, used to determine narrow lane type.

[0078] Specifically, a basic classification is first performed based on the completeness of the lane line data to be determined. If only some lane line attribute datasets are complete, they are directly marked as partially missing; if all lane line attribute datasets are missing, they are marked as completely missing. If all datasets are complete, the spatial geometric relationship of the four lane lines to be determined is analyzed based on the geometric coefficients in the lane line attribute datasets to determine whether they are parallel or non-parallel. At the same time, information such as the lateral distance between the vehicle and the lane lines is extracted from the environmental perception data to determine the relative position. If the vehicle's driving position is close to the edge of the lane line, it is directly marked as a narrow-space lane, thus completing the type determination of all lane lines to be determined.

[0079] The technical solution of this invention involves collecting environmental perception data corresponding to the target vehicle using intelligent driving sensors installed on the target vehicle; generating a lane line attribute dataset corresponding to the driving road based on the environmental perception data; determining the lane line type corresponding to each lane line of the driving road based on the lane line attribute dataset; and reconstructing the lane lines of the driving road based on the lane line types and the lane line attribute dataset to obtain the target lane lines. Based on this technical solution, by determining the lane line attribute dataset based on the environmental perception data and classifying lane line types, and then reconstructing lane lines based on the classified lane line types and the lane line attribute dataset, the accuracy and completeness of lane line data are improved, path deviation problems caused by lane line data defects are avoided, and the stability of the vehicle's driving trajectory is ensured.

[0080] Figure 4 This is a schematic diagram of a lane reconfiguration device for an autonomous vehicle provided in an embodiment of the present invention. Figure 4As shown, the device includes: a data acquisition module 410, a dataset generation module 420, a lane line type division module 430, and a lane line reconstruction module 440.

[0081] Data acquisition module 410 is used to collect environmental perception data corresponding to the target vehicle based on intelligent driving sensors installed on the target vehicle;

[0082] The dataset generation module 420 is used to generate a lane line attribute dataset corresponding to the driving road based on environmental perception data;

[0083] Lane type classification module 430 is used to determine the lane type corresponding to each lane of the driving road based on the lane attribute dataset.

[0084] The lane line reconstruction module 440 is used to reconstruct the lane lines of the driving road based on the lane line type and lane line attribute dataset to obtain the target lane lines.

[0085] In some embodiments, the lane line type classification module is used to determine the lane line to be judged corresponding to the target vehicle based on environmental perception data, wherein the lane line to be judged includes at least four lane lines; determine the data integrity status of the lane line to be judged based on the lane line attribute dataset; and determine the lane line type of each lane line to be judged based on the data integrity status and the lane line attribute dataset, wherein the lane line type includes parallel type, non-parallel type, partially missing type, completely missing type, and narrow-spaced lane type.

[0086] In some embodiments, the lane line reconstruction module is used to: determine the lateral distance between the target vehicle and the lane line of the driving lane based on environmental perception data when the lane line type is non-parallel; determine the axle lane line from the lane line of the driving lane based on the lateral distance; obtain the lane line attribute dataset of the axle lane line; and reconstruct the lane line of the driving road based on the lane line attribute dataset of the axle lane line to obtain the target lane line.

[0087] In some embodiments, the lane line reconstruction module is used to: select lane lines with complete lane line attribute datasets as axis lane lines when the lane line type is partially missing; obtain historical lane line attribute data for lane lines with missing data, wherein the historical lane line attribute data includes historical lateral distance parameters and historical lane width parameters; calculate the lateral distance prediction parameters for lane lines with missing data based on the historical lane line attribute data; complete the lane line attribute dataset for lane lines with missing data based on the axis lane line attribute dataset and the lateral distance prediction parameters; and reconstruct the lane lines of the driving road based on the completed lane line attribute dataset to obtain the target lane line.

[0088] In some embodiments, the lane line reconstruction module is used to: determine the actual driving trajectory corresponding to the target vehicle based on environmental perception data when the lane line type is completely missing, and obtain a preset fixed width value; derive a virtual lane line attribute dataset based on the preset fixed width value and the actual driving trajectory; generate a virtual lane line based on the virtual lane line attribute dataset, and use the virtual lane line as the target lane line.

[0089] In some embodiments, the lane line reconstruction module is used to, when the lane line type is a narrow lane type, determine the lateral distance between the target vehicle and the lane line of the driving lane based on environmental perception data, and determine the axle lane line from the lane line of the driving lane based on the lateral distance; and reconstruct the lane line of the driving road based on the minimum lateral distance parameter and the lane line attribute dataset of the axle lane line to obtain the target lane line.

[0090] In some embodiments, the lane line type classification module is used to determine the parallel type between each lane line to be determined based on the lane line attribute dataset when all lane line attribute datasets of the lane lines to be determined are complete; determine the lane line type as partially missing when only part of the lane line attribute datasets of the lane lines to be determined are complete; determine the lane line type as completely missing when all lane line attribute datasets of the lane lines to be determined are missing; determine the relative position of the target vehicle and the lane lines to be determined based on environmental perception data, and determine the lane line type as narrow-space lane when the target vehicle's driving position is close to the edge of the lane line.

[0091] The technical solution of this invention involves collecting environmental perception data corresponding to the target vehicle using intelligent driving sensors installed on the target vehicle; generating a lane line attribute dataset corresponding to the driving road based on the environmental perception data; determining the lane line type corresponding to each lane line of the driving road based on the lane line attribute dataset; and reconstructing the lane lines of the driving road based on the lane line types and the lane line attribute dataset to obtain the target lane lines. Based on this technical solution, by determining the lane line attribute dataset based on the environmental perception data and classifying lane line types, and then reconstructing lane lines based on the classified lane line types and the lane line attribute dataset, the accuracy and completeness of lane line data are improved, path deviation problems caused by lane line data defects are avoided, and the stability of the vehicle's driving trajectory is ensured.

[0092] The lane reconstruction device for autonomous vehicles provided in this embodiment of the invention can execute the lane reconstruction method for autonomous vehicles provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method.

[0093] Figure 5A schematic diagram of an electronic device 10, which can be used to implement embodiments of the present invention, 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 processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0094] like Figure 5 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded into the RAM 13 from storage unit 18. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0095] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0096] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as lane line reconstruction methods for autonomous vehicles.

[0097] In some embodiments, the lane reconstruction method for an autonomous vehicle can be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the lane reconstruction method for an autonomous vehicle described above can be performed. Alternatively, in other embodiments, processor 11 can be configured to perform the lane reconstruction method for an autonomous vehicle by any other suitable means (e.g., by means of firmware).

[0098] Various embodiments of the systems and techniques described above herein 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), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0099] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0100] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0101] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device 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 provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, 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 sound input, voice input, or tactile input).

[0102] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0103] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0104] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and no limitation is imposed herein.

[0105] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A lane line reconstruction method for an autonomous vehicle, characterized in that, include: Environmental perception data corresponding to the target vehicle is collected based on intelligent driving sensors installed on the target vehicle. Based on the environmental perception data, a lane line attribute dataset corresponding to the driving road is generated; The lane line type corresponding to each lane line of the driving road is determined based on the lane line attribute dataset. Based on the lane line type and the lane line attribute dataset, the driving road is reconstructed to obtain the target lane line.

2. The method according to claim 1, characterized in that, The step of determining the lane line type corresponding to each lane line of the driving road based on the lane line attribute dataset includes: Based on the environmental perception data, a lane line to be determined corresponding to the target vehicle is determined, wherein the lane line to be determined includes at least four lane lines. Determine the data integrity status of the lane line to be judged based on the lane line attribute dataset; Based on the data integrity status and the lane line attribute dataset, the lane line type of each lane line to be judged is determined, wherein the lane line type includes parallel type, non-parallel type, partially missing type, completely missing type, and narrow-spaced lane type.

3. The method according to claim 1, characterized in that, When the lane line type is non-parallel, the process of reconstructing the lane lines of the driving road based on the lane line type and the lane line attribute dataset to obtain the target lane line includes: Based on the environmental perception data, the lateral distance between the target vehicle and the lane lines of the driving lane is determined, and the axle lane line is determined from the lane lines of the driving lane based on the lateral distance. Obtain the lane line attribute dataset of the axle lane line, and reconstruct the lane lines of the driving road based on the lane line attribute dataset to obtain the target lane line.

4. The method according to claim 1, characterized in that, In the case where the lane line type is partially missing, the process of reconstructing the lane lines of the driving road based on the lane line type and the lane line attribute dataset to obtain the target lane lines includes: Select lane lines with complete lane line attribute datasets as axis lane lines, and obtain historical lane line attribute data for lane lines with missing data. The historical lane line attribute data includes historical lateral distance parameters and historical lane width parameters. The lateral distance prediction parameters for lane lines with missing data are calculated based on the historical lane line attribute data. The lane line attribute dataset based on the axle lane line and the lane line attribute dataset of lane lines with missing data completed by the lateral distance prediction parameter; The target lane lines are obtained by reconstructing the lane lines on the driving road based on the completed lane line attribute dataset.

5. The method according to claim 1, characterized in that, When the lane line type is completely missing, the process of reconstructing the lane lines of the driving road based on the lane line type and the lane line attribute dataset to obtain the target lane lines includes: Based on the environmental perception data, the actual driving trajectory corresponding to the target vehicle is determined, and a preset fixed width value is obtained; A virtual lane line attribute dataset is obtained by data derivation based on the preset fixed width value and the actual driving trajectory. Virtual lane lines are generated based on the virtual lane line attribute dataset, and these virtual lane lines are used as the target lane lines.

6. The method according to claim 1, characterized in that, When the lane line type is a narrow-space lane type, the process of reconstructing the lane lines of the driving road based on the lane line type and the lane line attribute dataset to obtain the target lane line includes: Based on the environmental perception data, the lateral distance between the target vehicle and the lane lines of the driving lane is determined, and the axle lane line is determined from the lane lines of the driving lane based on the lateral distance. The target lane line is obtained by reconstructing the lane line based on the minimum lateral distance parameter and the lane line attribute dataset of the axle lane line.

7. The method according to claim 2, characterized in that, The process of determining the lane type of each lane to be judged based on the data integrity status and the lane line attribute dataset includes: When the lane line attribute datasets of the lane lines to be determined are all in a complete state, the parallel type between each lane line to be determined is determined based on the lane line attribute datasets. If the lane line attribute dataset of the lane line to be determined is only partially complete, the lane line type is determined to be partially missing. If all lane line attribute datasets of the lane lines to be determined are missing, the lane line type is determined to be completely missing. Based on the environmental perception data, the relative position of the target vehicle and the lane line to be determined is determined. When the target vehicle is driving close to the edge of the lane line, the lane line type is determined to be a narrow lane type.

8. A lane reconfiguration device for an autonomous vehicle, characterized in that, include: The data acquisition module is used to collect environmental perception data corresponding to the target vehicle based on the intelligent driving sensors installed on the target vehicle; The dataset generation module is used to generate a lane line attribute dataset corresponding to the driving road based on the environmental perception data. The lane line type classification module is used to determine the lane line type corresponding to each lane line of the driving road based on the lane line attribute dataset. The lane line reconstruction module is used to reconstruct the lane lines of the driving road based on the lane line type and the lane line attribute dataset to obtain the target lane line.

9. A vehicle, characterized in that, The vehicle is equipped with: At least one processor; and a memory communicatively connected to the at least one processor; The memory stores a computer program that can be executed by the at least one processor, which is then executed by the at least one processor to enable the at least one processor to perform the lane line reconstruction method for an autonomous vehicle according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed by a processor, implement the lane line reconstruction method for an autonomous vehicle according to any one of claims 1-7.

Images