A vehicle lateral ranging optimization method, device and electronic equipment

Abstract

The present disclosure provides a vehicle lateral ranging optimization method, device and electronic equipment. By introducing a target vehicle lane attribute prediction and cross-frame timing fusion mechanism, and jointly determining the lane attribute information and the lateral ranging result, the accuracy of the target vehicle line pressure state recognition can be effectively improved, the misjudgment risk caused by the lateral ranging jitter can be reduced, and the stability and safety of the vehicle back-end control decision can be improved.

Description

Technical Field

[0001] This disclosure relates to the field of vehicle environmental perception technology, and more specifically, to a vehicle lateral ranging optimization method, device, and electronic device. Background Technology

[0002] With the continuous development of Advanced Driver Assistance Systems (ADAS) and autonomous driving technologies, the accuracy requirements for vehicles' perception of surrounding traffic participants are constantly increasing. Among these requirements, accurately determining the lateral position of a target vehicle and its relationship to the vehicle's lane is a crucial foundation for functions such as lane keeping control, adaptive cruise control, and lane change decision-making. In practical applications, onboard visual perception systems typically acquire the two-dimensional bounding box or three-dimensional position of the target vehicle through object detection algorithms, and combine this with lane line detection results to calculate the lateral distance of the target vehicle relative to the vehicle, thereby determining whether the target vehicle poses a risk of encroaching on the vehicle's lane.

[0003] However, existing technologies relying solely on lateral ranging results for lane relationship determination are susceptible to misjudgments due to various factors. For example, when lane line detection accuracy is insufficient, road markings are blurred, in curved scenarios, or when vehicles cross lane lines, the lateral distance to the nearest point of the target vehicle may fluctuate significantly, leading to unstable lane determination. Furthermore, in multi-target scenarios, due to jitter or tracking loss issues in target detection results, single-frame determination methods cannot guarantee the continuity and reliability of lane relationship determination, thus affecting the decision stability of the backend control module. Summary of the Invention

[0004] This disclosure provides at least one vehicle lateral ranging optimization method, device, and electronic device. By introducing a target vehicle lane attribute prediction and cross-frame temporal fusion mechanism, and jointly determining the lane attribute information and lateral ranging results, the accuracy of target vehicle lane crossing status identification can be effectively improved, the risk of misjudgment caused by lateral ranging jitter can be reduced, and the stability and safety of vehicle back-end control decisions can be improved.

[0005] This disclosure provides a method for optimizing lateral distance measurement of vehicles, including: Acquire road images and target vehicle detection results, collected from lane-related lane line parameters; Based on the road image and the target vehicle detection results, determine the lane attribute information corresponding to the target vehicle; Cross-frame tracking is performed on the target vehicle, and the lane attribute information is subjected to temporal fusion processing based on the tracking results to determine the final lane attribute value of the target vehicle. Obtain the lateral distance measurement result of the target vehicle, and determine the lane marking position corresponding to the target vehicle based on the final lane attribute value, the lateral distance measurement result, and the lane line parameters; The target information of the target vehicle and the line-crossing marker are output for the vehicle's back-end control module to make decisions.

[0006] In one optional implementation, the lane attribute information corresponding to the target vehicle is determined based on the road image and the target vehicle detection result, specifically including: The target vehicle in the road image is detected based on the target detection model, and the two-dimensional bounding box information and three-dimensional position parameters of the target vehicle are obtained. Based on the target detection model, lane attribute prediction is performed on the target vehicle to obtain the lane attribute probability array corresponding to the target vehicle. The initial lane attribute value of the target vehicle in the current frame is determined based on the lane attribute probability array.

[0007] In one optional implementation, the target detection model includes a lane attribute prediction branch, which is set in parallel with the target position regression branch; The feature map extracted by the target detection model is input into the lane attribute prediction branch, and the probability value of the target vehicle belonging to different lane attribute categories is output. Lane attributes include at least one of the following: lane attribute, lane-crossing attribute, and adjacent lane attribute; The attribute corresponding to the category with the highest probability in the lane attribute probability array is determined as the initial lane attribute value of the target vehicle in the current frame.

[0008] In one optional implementation, cross-frame tracking is performed on the target vehicle, and temporal fusion processing is performed on the lane attribute information based on the tracking results to determine the final lane attribute value of the target vehicle, specifically including: Create a lane attribute trajectory array for the tracking trajectory corresponding to each target vehicle; Based on the intersection-union ratio matching result or appearance feature matching result between the target vehicle in the current frame and the historical tracking trajectory, the association relationship between the target vehicle in the current frame and the historical tracking trajectory is determined; When the association is successful, the initial lane attribute value of the target vehicle in the current frame is written into the lane attribute trajectory array of the corresponding historical tracking trajectory; When the association fails, a new tracking trajectory is created for the target vehicle in the current frame, and a corresponding lane attribute trajectory array is established. The lane attribute trajectory array is a fixed-length array. When the length of the lane attribute trajectory array exceeds the preset length after a new initial lane attribute value is written, the earliest written lane attribute value is deleted to keep the length of the lane attribute trajectory array from exceeding the preset length. Write the initial lane attribute value of the target vehicle in the current frame into the lane attribute trajectory array that matches it; Based on multiple historical attribute values ​​in the lane attribute trajectory array, the final lane attribute value of the target vehicle is determined.

[0009] In one optional implementation, the final lane attribute value of the target vehicle is determined based on multiple historical attribute values ​​in the lane attribute trajectory array, specifically including: Count the number of occurrences of each lane attribute value in the lane attribute trajectory array; The lane attribute value that appears most frequently is determined as the final lane attribute value of the target vehicle; When there are multiple candidate lane attribute values ​​that appear the same number of times, the preliminary lane attribute value corresponding to the most recent frame is determined as the final lane attribute value.

[0010] In one optional implementation, the lateral distance measurement result of the target vehicle is obtained, and the lane line parameters are used to determine the lane marking position corresponding to the target vehicle, specifically including: Obtain the lateral distance measurement result of the nearest point of the target vehicle; Obtain the lateral reference parameters of the lane line corresponding to the lane boundary; The lateral distance measurement result of the nearest point is compared with the lateral reference parameters of the lane line, and combined with the final lane attribute value, the lane marking position corresponding to the target vehicle is determined.

[0011] In one optional implementation, the lane attribute value, the lateral distance measurement result, and the lane line parameters are used to determine the lane marking position corresponding to the target vehicle, specifically including: When the final lane attribute value indicates that the target vehicle belongs to the lane, the lane crossing marker is determined to be not crossing the lane. When the final lane attribute value indicates that the target vehicle belongs to the adjacent lane and the lateral distance measurement result indicates that the target vehicle is outside the lane, the line crossing marker is determined to be not crossing the line; When the final lane attribute value indicates that the target vehicle belongs to the adjacent lane and the lateral distance measurement result indicates that the target vehicle is located in the own lane, the lane crossing marker is determined as a possible lane crossing. When the final lane attribute value indicates that the target vehicle is in a lane-crossing state and the lateral distance measurement result indicates that the target vehicle is outside the lane, the lane-crossing marker is determined as a possible lane-crossing. When the final lane attribute value indicates that the target vehicle is in a line-crossing state and the lateral distance measurement result indicates that the target vehicle is located within the lane, the line-crossing marker is determined to be a line-crossing.

[0012] In one optional implementation, the target information of the target vehicle and the line-crossing marker are output for the vehicle's back-end control module to make a decision, specifically including: The target vehicle's two-dimensional frame information, three-dimensional position information, lateral distance measurement results, and line-crossing markers are packaged together. The packaged target information is sent to the vehicle decision-making or control layer according to the preset communication protocol. When the lane marking indicates that the vehicle has not crossed the lane, it is determined that the target vehicle does not pose a risk of intruding into the lane. When the lane marking position indicates lane crossing, it is determined that the target vehicle has the risk of intruding into the lane. When the line-crossing marker indicates a possible line crossing, a comprehensive decision is made by combining the target vehicle's speed, the vehicle's speed, and other environmental information.

[0013] This disclosure also provides a vehicle lateral ranging optimization device, including: The data acquisition module is used to acquire road images and target vehicle detection results, collected from lane-related lane line parameters; The lane attribute information determination module is used to determine the lane attribute information corresponding to the target vehicle based on the road image and the target vehicle detection result. The final attribute value determination module is used to perform cross-frame tracking for the target vehicle and perform temporal fusion processing on the lane attribute information based on the tracking results to determine the final lane attribute value of the target vehicle. The lane marking determination module is used to obtain the lateral distance measurement result of the target vehicle, and determine the lane marking position corresponding to the target vehicle based on the final lane attribute value, the lateral distance measurement result and the lane line parameters. The decision output module is used to output the target information of the target vehicle and the line-crossing marker, so that the vehicle's back-end control module can make a decision.

[0014] This disclosure also provides an electronic device, including: a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the memory via the bus. When the machine-readable instructions are executed by the processor, they perform the steps of the above-described vehicle lateral distance optimization method, or any possible implementation of the above-described vehicle lateral distance optimization method.

[0015] This disclosure also provides a computer-readable storage medium storing a computer program that, when executed by a processor, performs the steps of the above-described vehicle lateral distance optimization method, or any possible implementation thereof.

[0016] This disclosure also provides a computer program product, including a computer program / instructions, which, when executed by a processor, implement the above-described vehicle lateral distance measurement optimization method, or the steps in any possible implementation of the above-described vehicle lateral distance measurement optimization method.

[0017] This disclosure provides a vehicle lateral ranging optimization method, device, and electronic device. By introducing a target vehicle lane attribute prediction and cross-frame temporal fusion mechanism, and jointly determining the lane attribute information and lateral ranging results, it can effectively improve the accuracy of target vehicle lane crossing status identification, reduce the risk of misjudgment caused by lateral ranging jitter, and improve the stability and safety of vehicle back-end control decisions.

[0018] To make the above-mentioned objects, features and advantages of this disclosure more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of this disclosure, the accompanying drawings used in the embodiments will be briefly described below. These drawings are incorporated in and constitute a part of this specification. They illustrate embodiments conforming to this disclosure and, together with the specification, serve to explain the technical solutions of this disclosure. It should be understood that the following drawings only show some embodiments of this disclosure and should not be considered as limiting the scope. Those skilled in the art can obtain other related drawings based on these drawings without creative effort.

[0020] Figure 1 A flowchart of a vehicle lateral distance measurement optimization method provided by an embodiment of this disclosure is shown; Figure 2 A flowchart of a method for determining a final lane attribute value provided by an embodiment of this disclosure is shown; Figure 3 A schematic diagram of a vehicle lateral distance measurement optimization device provided in an embodiment of this disclosure is shown; Figure 4 A schematic diagram of an electronic device provided in an embodiment of the present disclosure is shown. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of the embodiments of this disclosure clearer, the technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this disclosure, and not all of them. The components of the embodiments of this disclosure described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this disclosure provided in the accompanying drawings is not intended to limit the scope of the claimed disclosure, but merely represents selected embodiments of this disclosure. All other embodiments obtained by those skilled in the art based on the embodiments of this disclosure without inventive effort are within the scope of protection of this disclosure.

[0022] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0023] In this document, the term "and / or" merely describes a relationship, indicating that three relationships can exist. For example, A and / or B can represent three cases: A alone, A and B simultaneously, and B alone. Furthermore, the term "at least one" in this document means any combination of at least two of any one or more elements. For example, including at least one of A, B, and C can mean including any one or more elements selected from the set consisting of A, B, and C.

[0024] Research has revealed that existing technologies relying solely on lateral ranging results for lane relationship determination are susceptible to misjudgments due to various factors. For instance, in scenarios with insufficient lane line detection accuracy, blurred road markings, curved sections, or when vehicles cross lane lines, the lateral distance to the nearest point of a target vehicle may fluctuate significantly, leading to unstable lane identification. Furthermore, in multi-target scenarios, due to jitter or tracking loss issues in target detection results, single-frame determination methods struggle to guarantee the continuity and reliability of lane relationship assessments, thus impacting the decision stability of the backend control module.

[0025] Based on the above research, this disclosure provides a vehicle lateral ranging optimization method, device, and electronic device. By introducing a target vehicle lane attribute prediction and cross-frame temporal fusion mechanism, and jointly determining the lane attribute information and lateral ranging results, the accuracy of target vehicle lane crossing status identification can be effectively improved, the risk of misjudgment caused by lateral ranging jitter can be reduced, and the stability and safety of vehicle back-end control decisions can be improved.

[0026] To facilitate understanding of this embodiment, a detailed description of the vehicle lateral ranging optimization method disclosed in this disclosure is provided first. The execution entity of the vehicle lateral ranging optimization method provided in this disclosure is generally a computer device with certain computing capabilities. This computer device may include, for example, a terminal device, a server, or other processing devices. The terminal device may be a user equipment (UE), mobile device, user terminal, terminal, cellular phone, cordless phone, personal digital assistant (PDA), handheld device, computing device, in-vehicle device, wearable device, etc. In some possible implementations, the vehicle lateral ranging optimization method can be implemented by a processor calling computer-readable instructions stored in memory.

[0027] See Figure 1 The diagram shows a flowchart of a vehicle lateral distance measurement optimization method provided in this embodiment of the present disclosure. The method includes steps S101 to S105, wherein: S101. Acquire road images and target vehicle detection results, collected from lane-related lane line parameters.

[0028] In practice, road images can be acquired by an onboard camera installed at the front of the vehicle. This onboard camera can be a monocular or binocular camera, and the captured image data reflects the traffic participants and lane markings in the road environment ahead of the vehicle. The onboard camera continuously acquires road images at a preset frame rate and transmits the acquired images to the vehicle's perception processing module for further processing.

[0029] Here, after acquiring the road image, the road image is processed using an object detection algorithm to identify target vehicles in the road scene and output the target vehicle detection results. The target vehicle detection results include at least the two-dimensional bounding box information of the target vehicle in the image, which is used to characterize the position range of the target vehicle in the image.

[0030] In some implementations, the target vehicle detection results may further include parameters such as the target vehicle's three-dimensional position information, size information, and orientation information, thereby providing basic data for subsequent lateral distance calculation and lane relationship judgment.

[0031] Preferably, the above-mentioned target detection algorithm can be implemented based on a deep learning model, such as a target detection network built on a convolutional neural network, which extracts features and identifies targets from the input road image and outputs the detection results corresponding to the target vehicle.

[0032] Simultaneously, lane line information in the road image is identified to obtain lane line parameters related to the lane. Specifically, lane line detection algorithms can be used to process the road image to extract the position and shape information of lane lines. Lane line detection algorithms can be implemented based on image processing methods or deep learning methods. By identifying lane line features in the road image, the position of the lane lines in the image is obtained, and the lane line information is further transformed into a vehicle coordinate system to obtain the spatial parameters corresponding to the lane lines.

[0033] In one embodiment, the lane-related lane line parameters include at least left-side lane line parameters and right-side lane line parameters, used to characterize the lateral position and variation trend of the lane boundary in the vehicle coordinate system. For example, lane line parameters may include the lateral offset of the lane line in the vehicle coordinate system, curvature parameters, or polynomial fitting parameters, used to describe the spatial shape of the lane line and its positional relationship relative to the vehicle.

[0034] In this way, by acquiring road images, target vehicle detection results, and lane line parameters related to the lane, the necessary data foundation can be provided for subsequent determination of the target vehicle's lane attribute information and optimization of the target vehicle's lateral distance measurement.

[0035] S102. Based on the road image and the target vehicle detection results, determine the lane attribute information corresponding to the target vehicle.

[0036] In practice, road images are input into a pre-constructed target vehicle detection model to detect, locate, and identify the attributes of target vehicles in the road scene. This outputs both the target vehicle detection results and the lane attribute prediction results for each target vehicle. The target vehicle detection model can be implemented using a 3D target detection model, preferably the Fcos3D model. The target vehicle detection results include at least the 2D bounding box information of the target vehicle to characterize its position range in the image coordinate system; furthermore, it may include one or more of the target vehicle's 3D position parameters, orientation parameters, and size parameters to characterize the target vehicle's spatial position and geometric features in the vehicle coordinate system.

[0037] In a preferred embodiment, to enable the target vehicle detection model to predict lane attributes, a lane attribute prediction head is added after the neck network of the Fcos3D model, and the lane attribute prediction head is set up in parallel with the original 3D position regression head and classification head.

[0038] Here, the lane attribute prediction head receives the feature map output by the target vehicle detection model as input. The feature map can be a feature map from layer P3 to layer P7. Based on the feature map, the lane attribute prediction head outputs the probability values ​​of the target vehicle belonging to multiple lane attribute categories to form the lane attribute probability array lane_score_array of the target vehicle. lane_score_array is preferably a three-dimensional vector lane_score_array=[P0, P1, P2], where P0, P1, and P2 represent the probability of the target vehicle belonging to different lane attribute categories, and satisfy P0+P1+P2=1.

[0039] Furthermore, the lane attribute categories include at least the lane-specific attribute, lane-crossing attribute, and adjacent lane attribute. Specifically, label 0 is used to indicate that the target vehicle belongs entirely to the lane and no part of the vehicle body crosses the lane line; label 1 is used to indicate that the target vehicle partially crosses the lane line and the edge of the vehicle body overlaps with the lane line; label 2 is used to indicate that the target vehicle belongs entirely to the adjacent lane and no part of the vehicle body crosses the lane line.

[0040] After obtaining the lane attribute probability array, the preliminary lane attribute information of the target vehicle in the current frame can be determined according to preset judgment rules. Preferably, the index value corresponding to the category with the highest probability value in lane_score_array is determined as the preliminary lane attribute value of the target vehicle in the current frame. That is, decoding is performed using argmax(lane_score_array), thereby converting the probability result output by the model into a discrete attribute identifier. This preliminary lane attribute value can be used as the output lane attribute information of the target vehicle in the current frame, or it can be further used for subsequent cross-frame tracking and temporal fusion processing.

[0041] Optionally, to improve the accuracy of target vehicle lane attribute recognition, the target vehicle detection model can be trained on a vehicle detection dataset with lane attribute annotations during the model training phase. Specifically, lane attribute labels can be added to each target in the original vehicle detection dataset to form training samples containing three types of labels: own lane, lane overrun, and adjacent lane. Furthermore, large vehicle samples in adjacent lanes can be added to the training samples, with the proportion of large vehicle samples in adjacent lanes not less than 30%, and the samples should cover complex scenarios such as strong light, rain, and nighttime to enhance the model's adaptability to large vehicle targets and lane attribute recognition in complex environments.

[0042] In one implementation, the target vehicle detection model uses cross-entropy loss to calculate the error between the predicted lane attribute values ​​and the true labels during training. The lane attribute loss (Loss_lane) and the original Fcos3D loss are combined to form the total loss function for joint training. The original Fcos3D loss includes classification loss and 3D regression loss. After training, the model is deployed in the vehicle vision perception module to output the 2D bounding box, 3D position, and corresponding lane_score_array and preliminary lane attribute values ​​for each target vehicle in real time.

[0043] In some implementations, to further improve the consistency between lane attribute determination results and actual road boundary relationships, the acquired lane-related lane line parameters can be combined to verify the lateral relationship of the target vehicle's two-dimensional bounding box position or three-dimensional position. The lane-related lane line parameters preferably include left-side and right-side lane line parameters, specifically lateral reference parameters of the lane lines in the vehicle coordinate system, such as the C0 coefficient. By jointly analyzing the target vehicle's position parameters and lane line parameters, auxiliary constraints can be applied to the lane attribute prediction results, thereby reducing the probability of misjudgment due to occlusion, target detection jitter, or interference from complex scenes.

[0044] S103. Perform cross-frame tracking on the target vehicle, and perform temporal fusion processing on the lane attribute information based on the tracking results to determine the final lane attribute value of the target vehicle.

[0045] In practice, after obtaining the preliminary lane attribute information of the target vehicle in the current frame, cross-frame tracking is performed on each target vehicle in the visual post-processing tracking module to establish the trajectory correspondence of the target vehicle in multiple consecutive frames. The purpose of this step is to eliminate instantaneous noise and jitter in single-frame lane attribute prediction through multi-frame tracking, making the lane attribute determination result of the target vehicle more stable, thereby avoiding fluctuations in subsequent lateral ranging and matching results due to misjudgment in a single frame.

[0046] For details, see Figure 2 The diagram shows a flowchart of a method for determining a final lane attribute value according to an embodiment of this disclosure. The method includes steps S1031 to S1035, wherein: S1031. Establish a lane attribute trajectory array for the tracking trajectory corresponding to each target vehicle.

[0047] S1032. Based on the intersection-union ratio matching result or appearance feature matching result between the target vehicle in the current frame and the historical tracking trajectory, determine the association between the target vehicle in the current frame and the historical tracking trajectory.

[0048] S1033. When the association is successful, write the initial lane attribute value of the target vehicle in the current frame into the lane attribute trajectory array of the corresponding historical tracking trajectory.

[0049] S1034. When the association fails, a new tracking trajectory is created for the target vehicle in the current frame, and a corresponding lane attribute trajectory array is established. The lane attribute trajectory array is a fixed-length array. When the length of the lane attribute trajectory array exceeds the preset length after a new initial lane attribute value is written, the earliest written lane attribute value is deleted to maintain the length of the lane attribute trajectory array from exceeding the preset length.

[0050] S1035. Write the initial lane attribute value of the target vehicle in the current frame into the lane attribute trajectory array that matches it; determine the final lane attribute value of the target vehicle based on multiple historical attribute values ​​in the lane attribute trajectory array.

[0051] In one implementation, a lane attribute trajectory array `lane_id_tracker` is created for each target vehicle's tracking trajectory tracker to store the initial lane attribute values ​​of the target vehicle in the most recent consecutive frames. Each `lane_id_tracker` corresponds one-to-one with the target vehicle's tracking trajectory. As the target vehicle is continuously tracked between adjacent frames, the initial lane attribute values ​​obtained in each frame are sequentially written into the corresponding `lane_id_tracker`, thereby forming a temporal attribute sequence reflecting the change process of the target vehicle's lane attributes.

[0052] Furthermore, `lane_id_tracker` is preferably set to a fixed-length array with a maximum length of 4, used to cache the preliminary lane attribute values ​​of the target vehicle in the most recent 4 frames. When a new preliminary lane attribute value is written, if the current array length exceeds the preset length of 4, the attribute value of the earliest written frame is deleted, so as to maintain the length of `lane_id_tracker` not exceeding 4 in a first-in-first-out manner. By adopting a fixed-length caching mechanism, sufficient historical attribute information can be retained to support stabilization decisions, while avoiding the storage burden and interference from outdated attribute values ​​caused by the infinite growth of the trajectory array.

[0053] When performing cross-frame tracking of a target vehicle in the current frame, each target vehicle detected in the current frame can be associated with historical tracking trajectories. Preferably, the association matching is achieved by combining intersection-union (IU) matching and appearance feature matching. Specifically, firstly, an initial matching relationship is calculated based on the IU of the two-dimensional bounding boxes between the target vehicle in the current frame and the historical tracking trajectory. Then, the matching result is verified by combining the appearance feature similarity of the target vehicle, thereby determining which historical tracking trajectory the target vehicle in the current frame corresponds to.

[0054] Thus, by using the above matching method, the accuracy of trajectory association can be improved even when the target detection box is slightly jittery, partially occluded, or adjacent targets are moving in parallel.

[0055] In one specific implementation, if the target vehicle in the current frame successfully matches a historical tracking trajectory, the initial lane attribute value of the target vehicle in the current frame is added to the lane_id_tracker corresponding to the matched trajectory. If the length of the lane_id_tracker exceeds a preset length of 4 after adding the value, the data from the earliest frame is deleted. If the target vehicle in the current frame fails to match any historical tracking trajectory, the target vehicle is treated as a new target, a new tracking trajectory is created for it, and the corresponding lane_id_tracker is initialized. Simultaneously, the initial lane attribute value of the target vehicle in the current frame is written as the first element into the lane_id_tracker. This allows for adaptation to both continuously tracked targets and newly appearing targets.

[0056] Furthermore, after completing trajectory association and array updates, a temporal fusion process is performed on the target vehicle based on multiple historical attribute values ​​in lane_id_tracker to determine the final lane attribute value of the target vehicle in the current frame. Preferably, for successfully matched tracking trajectories, the frequency of occurrence of each lane attribute value in lane_id_tracker is counted, and the lane attribute value with the most frequent occurrences, i.e., the mode, is determined as the final lane attribute value of the target vehicle in the current frame.

[0057] For example, when lane_id_tracker=[2, 2, 1, 2], the attribute value 2 appears the most times, so the final lane attribute value is determined to be 2, indicating that the target vehicle is ultimately identified as a target in the adjacent lane.

[0058] Furthermore, when there are multiple candidate attribute values ​​with the same frequency in lane_id_tracker, the preliminary lane attribute value corresponding to the most recent frame can be determined as the final lane attribute value. For example, when lane_id_tracker=[2, 2, 1, 1], attribute value 2 and attribute value 1 appear the same number of times. In this case, the preliminary lane attribute value written in the most recent frame is taken as the final value.

[0059] Optionally, for a newly created tracking trajectory, since its lane_id_tracker only contains the preliminary lane attribute values ​​of the current frame, the preliminary lane attribute values ​​can be directly output as the final lane attribute values ​​of the target vehicle in the current frame; as new attribute values ​​are continuously written in subsequent frames, the final lane attribute values ​​are dynamically updated according to the above-mentioned mode decision rule.

[0060] This approach ensures that effective attribute output is obtained upon the initial appearance of a new target, while also guaranteeing that the attribute results gradually stabilize as the number of observation frames increases. Based on this, the final lane attribute value can be used as input for subsequent lane attribute matching and lane marking determination, participating in the target lateral relationship analysis together with the lateral distance measurement results of the target vehicle.

[0061] S104. Obtain the lateral distance measurement result of the target vehicle, and determine the lane marking position corresponding to the target vehicle based on the final lane attribute value, the lateral distance measurement result and the lane line parameters.

[0062] In practical implementation, the key parameters used for determining the lane crossing status are first obtained. These key parameters include the lateral distance measurement result of the target vehicle and the lane line parameters related to the vehicle's own lane. The preferred lateral distance measurement result is the lateral distance measurement result dis_y of the closest point of the target vehicle, i.e., the lateral distance between the target vehicle and the nearest point of the vehicle. The preferred lane line parameter is the C0 coefficient of the right lane line of the vehicle's own lane. The C0 coefficient is used to characterize the lateral position reference of the corresponding lane line in the vehicle coordinate system.

[0063] Here, by comparing the coefficients dis_y and C0, it can be determined whether the target vehicle is located inside or outside the lane in terms of lateral position. For example, when dis_y is greater than C0, it can be indicated that the target vehicle is outside the lane, and when dis_y is less than C0, it can be indicated that the target vehicle is inside the lane.

[0064] In one embodiment, the lateral ranging result is calculated from the three-dimensional position parameters output by the target vehicle detection model, or it can be directly output by the target vehicle detection model. Preferably, the lateral ranging result is the closest point lateral ranging result determined based on the target's three-dimensional position information output by the Fcos3D model. The lane line parameters are output by the vehicle lane line detection module, the visual mapping module, or the lane line fitting module. In addition to the C0 coefficient of the right lane line, they may also include the left lane line parameters or other parameters that can characterize the lateral positional relationship of the lane boundaries. This application does not limit this.

[0065] Furthermore, after obtaining the final lane attribute value, the lateral distance measurement result dis_y, and the lane line parameter C0, the crossing marker position crossing_lane corresponding to the target vehicle is determined according to the preset matching rules.

[0066] Preferably, the crossing lane indicator is a multi-level indicator used to distinguish whether the target vehicle is not crossing the lane, may be crossing the lane, or is crossing the lane. More preferably, the crossing lane indicator includes the following values: crossing_lane=0 indicates not crossing the lane, crossing_lane=1 indicates may be crossing the lane, and crossing_lane=2 indicates crossing the lane.

[0067] In one specific implementation, when the final lane attribute value indicates that the target vehicle belongs to its own lane, the lane crossing flag is directly determined as not crossing the lane. Specifically, it can be set as follows: if the final lane attribute value of the target is 0, then marking crossing_lane=0. The reason is that when the target vehicle is still stably determined to be a target in its own lane after multi-frame temporal fusion, it means that its entire vehicle body is within the range of its own lane and does not show obvious lane crossing characteristics, so it can be directly determined to be in a non-lane crossing state.

[0068] In another specific implementation, when the final lane attribute value indicates that the target vehicle belongs to the adjacent lane, the crossing marker position is further determined by combining the comparison results between the lateral distance measurement result and the lane line parameters. Specifically, this may include: if the final lane attribute value of the target is 2 and dis_y is greater than C0, it means that the lateral distance measurement result and the lane attribute result are consistent, both indicating that the target vehicle is in the adjacent lane. Therefore, crossing_lane is marked as 0, indicating that the vehicle has not crossed the line. If the final lane attribute value of the target is 2, but dis_y is less than C0, it means that the lateral distance measurement result and the lane attribute result are inconsistent. That is, the attribute result indicates that the target is in the adjacent lane, while the lateral distance measurement result indicates that the target has entered the lane. In this case, it is likely a contradictory scenario caused by the distance measurement error of a large vehicle or the boundary recognition deviation. Therefore, crossing_lane is marked as 1, indicating that the vehicle may have crossed the line.

[0069] Furthermore, when the final lane attribute value indicates that the target vehicle is crossing the lane line, the lane crossing flag is also determined by comparing the lateral distance measurement result with the lane line parameters. Specifically, this can include: if the target's final lane attribute value is 1 and dis_y is greater than C0, it indicates that the target has lane crossing characteristics in terms of attributes, but the lateral distance measurement result shows that its nearest point is still outside the lane, so crossing_lane is marked as 1, indicating that it may be crossing the lane line; if the target's final lane attribute value is 1 and dis_y is less than C0, it indicates that both the attribute result and the lateral distance measurement result support the judgment that the target is crossing the lane line and intruding into the lane, so crossing_lane is marked as 2, indicating that it is crossing the lane line.

[0070] In some implementations, the above-mentioned logic for determining the crossing lane marker can be implemented using a lookup table, conditional branching, or a rule engine. For example, a mapping table between the final lane attribute value, the lateral distance measurement result, and the crossing lane marker can be pre-established. After obtaining the final lane attribute value and the comparison result between dis_y and C0, the corresponding crossing_lane value can be directly output by looking up the table. Alternatively, the above logic can be executed item by item through multi-level conditional statements in the program. As long as it is possible to jointly determine the crossing lane marker based on the final lane attribute value, the lateral distance measurement result, and the lane line parameters, it falls within the protection scope of this application.

[0071] Optionally, after determining the line crossing marker, the target information of the target vehicle is associated with the line crossing marker and output. The target information preferably includes the two-dimensional bounding box information, three-dimensional position information, and lateral distance measurement result dis_y of the target vehicle, and the line crossing marker is preferably crossing_lane.

[0072] By packaging and distributing 2DBbox, 3D position, dis_y, and crossing_lane together, more complete target semantic and geometric information can be provided to the backend decision-making layer. For example, when crossing_lane=0, even if dis_y is slightly less than C0, the backend can still prioritize determining that the target is located in the adjacent lane and there is no risk of intrusion; when crossing_lane=2, it can be determined that the target has indeed crossed the line and intruded, triggering braking; when crossing_lane=1, a comprehensive judgment can be made by combining the target speed, the vehicle's speed, and other environmental information.

[0073] In this way, the lateral distance measurement results of the target vehicle can be obtained through the above method. Based on the final lane attribute value, the lateral distance measurement results and lane line parameters, the corresponding lane crossing mark position of the target vehicle can be determined. This effectively combines the multi-frame semantic attribute results of the target vehicle with the single-frame geometric distance measurement results, improves the accuracy of lane crossing recognition and intrusion judgment, reduces the risk of misjudgment caused by the lateral distance measurement deviation of large vehicles, and provides a more reliable decision-making basis for back-end functions such as ACC and AEB.

[0074] S105. Output the target information of the target vehicle and the line-crossing marker for the vehicle's back-end control module to make a decision.

[0075] In practice, after determining the target vehicle's line-crossing marker, the visual perception module correlates and outputs the target information and line-crossing marker for each target vehicle, and sends it to the vehicle's back-end control module according to a preset data format. The purpose of this step is to uniformly distribute the target geometric information, lateral distance measurement information, and line-crossing semantic information obtained from the front-end perception side to the back-end, so that the back-end control module can make driving control decisions based on a more complete target representation.

[0076] In one implementation, the target information includes at least the target vehicle's two-dimensional bounding box (2DBbox), three-dimensional position information, and lateral ranging result (dis_y). The line-crossing marker is preferably the crossing_lane. That is, the visual perception module packages the 2DBbox, 3D position, dis_y, and crossing_lane corresponding to each target and sends them to the vehicle's decision-making layer according to a standard protocol.

[0077] Furthermore, the packaging process can be organized according to a preset field order and data structure. For example, the target vehicle identification information, two-dimensional bounding box coordinates, three-dimensional position parameters, lateral distance measurement results, and line-crossing markers can be sequentially written into the data packet and then sent to the back-end control module via the vehicle communication bus or in-vehicle network.

[0078] It should be noted that the standard protocol can be any data communication protocol already deployed within the vehicle. This application does not limit this, as long as it enables the standardized distribution of target information and line-crossing markers, it is applicable to this application. This improves the data compatibility and interface adaptability between the perception results and the decision-making module.

[0079] In a preferred embodiment, the vehicle back-end control module includes an ACC (Adaptive Cruise Control) module; further, it may also include an AEB (Automatic Emergency Braking) module or other control modules related to the risk assessment of the target vehicle. After receiving the target information and the lane marking, the back-end control module prioritizes the lane marking to determine whether the target vehicle poses a risk of intruding into the lane, rather than relying solely on the lateral distance measurement result dis_y.

[0080] Specifically, in one implementation, when crossing_lane=0, it indicates that the target vehicle has not crossed the lane. In this case, even if dis_y is slightly less than C0, the back-end control module can still prioritize determining that the target vehicle is located in the adjacent lane and has no risk of intrusion, thus not triggering braking or performing avoidance control for the intruding target. This approach is particularly suitable for scenarios where large vehicles are traveling in adjacent lanes and, due to lateral distance measurement deviation, the closest point mistakenly enters the lane side, effectively reducing the probability of false braking.

[0081] Furthermore, when `crossing_lane=2`, it indicates that the target vehicle is crossing the lane line and has intruded into the lane. In this case, the backend control module can determine that the target vehicle is a target with a clear risk of intrusion and trigger corresponding control operations, such as braking control, risk warning, or other driving safety control operations. By distinguishing and processing this type of target from those that are not crossing the lane line, the accuracy of the backend control module's response to truly dangerous targets can be improved.

[0082] In another implementation, when crossing_lane=1, it indicates that the target vehicle is in a state where it may cross the lane. This state generally corresponds to a situation where there is a certain contradiction between the target's final lane attribute value and the lateral distance measurement result, or the target is in a critical boundary area. In this case, the back-end control module does not immediately make a deterministic judgment based on a single parameter, but instead makes a comprehensive judgment by combining the target vehicle speed, the vehicle's own speed, and other environmental information to avoid false braking due to relying solely on dis_y.

[0083] It should be noted that other environmental information may also include target vehicle heading change information, relative speed information, road curvature information, adjacent lane traffic status information, etc., but this application does not limit this.

[0084] Optionally, the process of outputting target information and lane markings may also include filtering, sorting, or prioritizing target vehicles. For example, target vehicles located in front of the vehicle's direction of travel and with a high collision relevance may be prioritized; targets may also be prioritized based on target distance, relative speed, and lane markings so that the backend control module can prioritize high-risk targets.

[0085] It should be noted that as long as the processing does not change the core scheme of associating the target information with the pressure mark and outputting it for the back-end control module to make decisions, it can be used as an optional implementation method of this application.

[0086] In this way, the target information of the target vehicle and the lane crossing marker can be output to the vehicle's back-end control module. When making ACC, AEB or other driving control decisions, the back-end control module not only refers to the target vehicle's two-dimensional bounding box, three-dimensional position and lateral distance measurement results, but also refers to the lane crossing marker that reflects the risk of the target lane intrusion. This improves the accuracy of identification and decision-making for large vehicles in adjacent lanes, lane crossing targets, and lateral distance fluctuation scenarios, and reduces the risk of false braking and missed braking.

[0087] This disclosure provides a vehicle lateral ranging optimization method that introduces a target vehicle lane attribute prediction and cross-frame temporal fusion mechanism, and jointly determines the lane attribute information and lateral ranging results. This effectively improves the accuracy of target vehicle lane crossing status identification, reduces the risk of misjudgment caused by lateral ranging jitter, and enhances the stability and safety of vehicle back-end control decisions.

[0088] Those skilled in the art will understand that, in the above-described method of the specific implementation, the order in which each step is written does not imply a strict execution order and does not constitute any limitation on the implementation process. The specific execution order of each step should be determined by its function and possible internal logic.

[0089] Based on the same inventive concept, this disclosure also provides a vehicle lateral distance measurement optimization device corresponding to the vehicle lateral distance measurement optimization method. Since the principle of the device in this disclosure for solving the problem is similar to the vehicle lateral distance measurement optimization method described above in this disclosure, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be described again.

[0090] Please see Figure 3 , Figure 3 This is a schematic diagram of a vehicle lateral distance measurement optimization device provided in an embodiment of this disclosure. Figure 3 As shown in the figure, the vehicle lateral ranging optimization device 300 provided in this embodiment includes: The data acquisition module 310 is used to acquire road images and target vehicle detection results, which are collected from lane-related lane line parameters.

[0091] The lane attribute information determination module 320 is used to determine the lane attribute information corresponding to the target vehicle based on the road image and the target vehicle detection result.

[0092] The final attribute value determination module 330 is used to perform cross-frame tracking for the target vehicle and perform temporal fusion processing on the lane attribute information based on the tracking results to determine the final lane attribute value of the target vehicle.

[0093] The lane marking determination module 340 is used to obtain the lateral distance measurement result of the target vehicle, and determine the lane marking position corresponding to the target vehicle based on the final lane attribute value, the lateral distance measurement result and the lane line parameters.

[0094] The decision output module 350 is used to output the target information of the target vehicle and the line-crossing marker position for the vehicle back-end control module to make a decision.

[0095] The processing flow of each module in the device and the interaction flow between each module can be referred to the relevant descriptions in the above method embodiments, and will not be detailed here.

[0096] This disclosure provides a vehicle lateral ranging optimization device that introduces a target vehicle lane attribute prediction and cross-frame temporal fusion mechanism, and jointly determines the lane attribute information and lateral ranging results. This effectively improves the accuracy of target vehicle lane crossing status recognition, reduces the risk of misjudgment caused by lateral ranging jitter, and enhances the stability and safety of vehicle back-end control decisions.

[0097] Corresponding to Figure 1 In the vehicle lateral ranging optimization method, this disclosure also provides an electronic device 400, such as... Figure 4The diagram shown is a structural schematic of an electronic device 400 provided in an embodiment of this disclosure, including: Processor 41, memory 42, and bus 43; memory 42 is used to store execution instructions, including main memory 421 and external memory 422; the main memory 421, also called internal memory, is used to temporarily store the computational data in processor 41, as well as the data exchanged with external memory 422 such as hard disk. Processor 41 exchanges data with external memory 422 through main memory 421. When the electronic device 400 is running, processor 41 and memory 42 communicate through bus 43, enabling processor 41 to execute... Figure 1 The steps of the vehicle lateral distance measurement optimization method.

[0098] This disclosure also provides a computer-readable storage medium storing a computer program that, when executed by a processor, performs the steps of the vehicle lateral distance optimization method described in the above-described method embodiments. The storage medium can be a volatile or non-volatile computer-readable storage medium.

[0099] This disclosure also provides a computer program product, which includes computer instructions. When the computer instructions are executed by a processor, they can perform the steps of the vehicle lateral distance measurement optimization method described in the above method embodiments. For details, please refer to the above method embodiments, which will not be repeated here.

[0100] The aforementioned computer program product can be implemented through hardware, software, or a combination thereof. In one optional embodiment, the computer program product is specifically embodied in a computer storage medium; in another optional embodiment, the computer program product is specifically embodied in a software product, such as a software development kit (SDK), etc.

[0101] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the device described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here. In the several embodiments provided in this disclosure, it should be understood that the disclosed device and method can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Another point is that the displayed or discussed mutual coupling or direct coupling or communication connection may be through some communication interfaces; the indirect coupling or communication connection of devices or units may be electrical, mechanical, or other forms.

[0102] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0103] In addition, the functional units in the various embodiments of this disclosure can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0104] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of this disclosure, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this disclosure. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0105] Finally, it should be noted that the above-described embodiments are merely specific implementations of this disclosure, used to illustrate the technical solutions of this disclosure, and not to limit it. The protection scope of this disclosure is not limited thereto. Although this disclosure has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features, within the scope of the technology disclosed in this disclosure. Such modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this disclosure, and should all be covered within the protection scope of this disclosure. Therefore, the protection scope of this disclosure should be determined by the protection scope of the claims.

Claims

1. A method for optimizing lateral distance measurement of vehicles, characterized in that, include: Acquire road images and target vehicle detection results, collected from lane-related lane line parameters; Based on the road image and the target vehicle detection results, determine the lane attribute information corresponding to the target vehicle; Cross-frame tracking is performed on the target vehicle, and the lane attribute information is temporally fused based on the tracking results to determine the final lane attribute value of the target vehicle. Obtain the lateral distance measurement result of the target vehicle, and determine the lane marking position corresponding to the target vehicle based on the final lane attribute value, the lateral distance measurement result, and the lane line parameters; The target information of the target vehicle and the line-crossing marker are output for the vehicle's back-end control module to make decisions.

2. The method according to claim 1, characterized in that, Based on the road image and the target vehicle detection results, the lane attribute information corresponding to the target vehicle is determined, specifically including: The target vehicle in the road image is detected based on the target detection model, and the two-dimensional bounding box information and three-dimensional position parameters of the target vehicle are obtained. Based on the target detection model, lane attribute prediction is performed on the target vehicle to obtain the lane attribute probability array corresponding to the target vehicle. The initial lane attribute value of the target vehicle in the current frame is determined based on the lane attribute probability array.

3. The method according to claim 2, characterized in that: The target detection model includes a lane attribute prediction branch, which is set in parallel with the target position regression branch. The feature map extracted by the target detection model is input into the lane attribute prediction branch, and the probability value of the target vehicle belonging to different lane attribute categories is output. Lane attributes include at least one of the following: lane attribute, lane-crossing attribute, and adjacent lane attribute; The attribute corresponding to the category with the highest probability in the lane attribute probability array is determined as the initial lane attribute value of the target vehicle in the current frame.

4. The method according to claim 1, characterized in that, Cross-frame tracking is performed on the target vehicle, and temporal fusion processing is performed on the lane attribute information based on the tracking results to determine the final lane attribute value of the target vehicle, specifically including: Create a lane attribute trajectory array for the tracking trajectory corresponding to each target vehicle; Based on the intersection-union ratio matching result or appearance feature matching result between the target vehicle in the current frame and the historical tracking trajectory, the association relationship between the target vehicle in the current frame and the historical tracking trajectory is determined; When the association is successful, the initial lane attribute value of the target vehicle in the current frame is written into the lane attribute trajectory array of the corresponding historical tracking trajectory; When the association fails, a new tracking trajectory is created for the target vehicle in the current frame, and a corresponding lane attribute trajectory array is established. The lane attribute trajectory array is a fixed-length array. When the length of the lane attribute trajectory array exceeds the preset length after a new initial lane attribute value is written, the earliest written lane attribute value is deleted to keep the length of the lane attribute trajectory array from exceeding the preset length. Write the initial lane attribute value of the target vehicle in the current frame into the lane attribute trajectory array that matches it; Based on multiple historical attribute values ​​in the lane attribute trajectory array, the final lane attribute value of the target vehicle is determined.

5. The method according to claim 4, characterized in that, Based on multiple historical attribute values ​​in the lane attribute trajectory array, the final lane attribute value of the target vehicle is determined, specifically including: Count the number of occurrences of each lane attribute value in the lane attribute trajectory array; The lane attribute value that appears most frequently is determined as the final lane attribute value of the target vehicle; When there are multiple candidate lane attribute values ​​that appear the same number of times, the preliminary lane attribute value corresponding to the most recent frame is determined as the final lane attribute value.

6. The method according to claim 1, characterized in that, Obtain the lateral distance measurement result of the target vehicle, and determine the lane marking position corresponding to the target vehicle based on the final lane attribute value, the lateral distance measurement result, and the lane line parameters, specifically including: Obtain the lateral distance measurement result of the nearest point of the target vehicle; Obtain the lateral reference parameters of the lane line corresponding to the lane boundary; The lateral distance measurement result of the nearest point is compared with the lateral reference parameters of the lane line, and combined with the final lane attribute value, the lane marking position corresponding to the target vehicle is determined.

7. The method according to claim 6, characterized in that, Based on the final lane attribute value, the lateral distance measurement result, and the lane line parameters, the lane marking position corresponding to the target vehicle is determined, specifically including: When the final lane attribute value indicates that the target vehicle belongs to the lane, the lane crossing marker is determined to be not crossing the lane. When the final lane attribute value indicates that the target vehicle belongs to the adjacent lane and the lateral distance measurement result indicates that the target vehicle is outside the lane, the line crossing marker is determined to be not crossing the line; When the final lane attribute value indicates that the target vehicle belongs to the adjacent lane and the lateral distance measurement result indicates that the target vehicle is located in the own lane, the lane crossing marker is determined as a possible lane crossing. When the final lane attribute value indicates that the target vehicle is in a lane-crossing state and the lateral distance measurement result indicates that the target vehicle is outside the lane, the lane-crossing marker is determined as a possible lane-crossing. When the final lane attribute value indicates that the target vehicle is in a line-crossing state and the lateral distance measurement result indicates that the target vehicle is located within the lane, the line-crossing marker is determined to be a line-crossing.

8. The method according to claim 1, characterized in that, Output the target information of the target vehicle and the line-crossing marker for the vehicle's back-end control module to make decisions, specifically including: The target vehicle's two-dimensional frame information, three-dimensional position information, lateral distance measurement results, and line-crossing markers are packaged together. The packaged target information is sent to the vehicle decision-making or control layer according to the preset communication protocol. When the lane marking indicates that the vehicle has not crossed the lane, it is determined that the target vehicle does not pose a risk of intruding into the lane. When the lane marking position indicates lane crossing, it is determined that the target vehicle has the risk of intruding into the lane. When the line-crossing marker indicates a possible line crossing, a comprehensive decision is made by combining the target vehicle's speed, the vehicle's speed, and other environmental information.

9. A vehicle lateral distance measurement optimization device, characterized in that, include: The data acquisition module is used to acquire road images and target vehicle detection results, collected from lane-related lane line parameters; The lane attribute information determination module is used to determine the lane attribute information corresponding to the target vehicle based on the road image and the target vehicle detection result. The final attribute value determination module is used to perform cross-frame tracking for the target vehicle and perform temporal fusion processing on the lane attribute information based on the tracking results to determine the final lane attribute value of the target vehicle. The lane marking determination module is used to obtain the lateral distance measurement result of the target vehicle, and determine the lane marking position corresponding to the target vehicle based on the final lane attribute value, the lateral distance measurement result and the lane line parameters. The decision output module is used to output the target information of the target vehicle and the line-crossing marker, so that the vehicle's back-end control module can make a decision.

10. An electronic device, characterized in that, include: The device includes a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the memory via the bus. When the machine-readable instructions are executed by the processor, the steps of the vehicle lateral distance optimization method as described in any one of claims 1 to 8 are performed.

Images