Vehicle adaptive cruise target selection method, device, equipment and storage medium

Abstract

This application provides a method, apparatus, device, and storage medium for target selection in adaptive cruise control, applicable to the field of adaptive cruise control technology. The method includes: acquiring image detection data of the area in front of the vehicle using an onboard camera, and acquiring an image tracking target and the vehicle's current lane based on the image detection data; acquiring radar detection data of the area in front of the vehicle using an onboard radar, and acquiring a radar tracking target based on the radar detection data; fusing the image tracking target and the radar tracking target to obtain a fused target; selecting the fused target that continuously meets target selection criteria within a preset time period as a potential following target; wherein the target selection criteria include pixel-level selection criteria and world-level selection criteria; and selecting the potential following target with the smallest longitudinal distance to the vehicle as the cruise following target. This application's method, by combining pixel-level selection and world-level selection, effectively improves the accuracy of adaptive cruise control target selection.

Description

Technical Field

[0001] This application relates to the field of vehicle adaptive cruise technology, and in particular to a vehicle adaptive cruise target selection method, apparatus, device, and storage medium. Background Technology

[0002] Adaptive Cruise Control (ACC) is an intelligent automatic control system designed to improve driving comfort and safety. ACC is developed from traditional cruise control technology, allowing the vehicle's cruise control system to adjust speed to adapt to traffic conditions.

[0003] During the operation of a vehicle's adaptive cruise control system, target selection is a crucial step, determining which vehicles or obstacles will be considered by the vehicle's cruise control system as objects that need to be tracked or at a safe distance.

[0004] However, existing target selection methods suffer from low accuracy in cruise control. This is because current methods rely on the relative position between the target vehicle and the cruise vehicle. In complex road conditions, vehicles are prone to frequent bumps and vibrations, causing changes in vehicle position and even the mounting position of onboard sensors (cameras, radars). This leads to discrepancies between the target vehicle and lane line data after coordinate transformation and the actual position, causing the ACC to misfollow or miss targets, or even result in collisions. This significantly impacts the accuracy of adaptive cruise control and the safety of the system. Summary of the Invention

[0005] This application provides a vehicle adaptive cruise target selection method, apparatus, device, and storage medium to solve the problem that existing target selection methods have low accuracy in selecting targets while following another vehicle.

[0006] According to the first aspect disclosed in this application, this application provides an adaptive cruise target selection method, including:

[0007] The vehicle acquires image detection data of the front of the vehicle using an onboard camera, and obtains the image tracking target and the current lane where the vehicle is located based on the image detection data;

[0008] The vehicle-mounted radar acquires radar detection data from the front of the vehicle, and the radar tracking target is obtained based on the radar detection data.

[0009] The image-tracked target and the radar-tracked target are fused to obtain the fused target;

[0010] Filtration targets that continuously meet the target screening conditions within a preset time period are selected as potential vehicle-following targets; wherein, the target screening conditions include pixel-level screening conditions and world-level screening conditions;

[0011] Select the potential vehicle with the smallest longitudinal distance from your own vehicle as the cruise following target.

[0012] In one feasible implementation, the pixel layer screening conditions include the fusion target being within the current lane and the invasive rate of the fusion target entering the current lane being greater than a first judgment threshold.

[0013] The world layer screening criteria include the fact that the distance between the fusion target and the centerline of the current lane is less than the tightening width.

[0014] In one feasible implementation, the pixel layer screening conditions include the fusion target being within the current lane and the invasive rate of the fusion target entering the current lane being greater than a second judgment threshold.

[0015] The world layer filtering criteria include the minimum value among the lateral distance between the fused target and the vehicle being less than the minimum turning radius filtering threshold and the lane width of the current lane.

[0016] In one feasible implementation, obtaining the image tracking target and the current lane where the vehicle is located based on the image detection data includes:

[0017] The image detection data is input into the detection and recognition model to obtain image tracking target and lane line point data; wherein, the detection and recognition model includes a tracking target recognition sub-model and a lane recognition sub-model obtained based on a neural network model training;

[0018] The image tracking target is tracked based on the target tracking model, and the target bounding box and target motion attributes of the image tracking target are obtained; wherein, the target tracking model is constructed based on the visual target tracking algorithm, and the target motion attributes include the longitudinal and lateral distances from the vehicle, and the longitudinal and lateral velocities relative to the vehicle;

[0019] The lane line point data is subjected to curve fitting and filtering tracking processing to obtain the lane lines and lane information of the current lane; wherein, the lane information includes the lane width and the distance of the image tracking target from the center line of the current lane.

[0020] In one feasible implementation, the method further includes:

[0021] If any point on the bottom edge of the target bounding box is within the lane line of the current lane, then the image tracking target is determined to be within the current lane, and the intrusion rate of the image tracking target into the current lane is obtained; wherein, the intrusion rate is the ratio between the area of ​​the target bounding box that intrudes into the current lane and the area of ​​the target bounding box.

[0022] In one feasible implementation, the method further includes:

[0023] If any point on the bottom edge of the target frame is not within the current lane, then it is determined that the image tracking target is not within the current lane.

[0024] In one feasible implementation, the image-tracked target and the radar-tracked target are fused to obtain a fused target, including:

[0025] The image-tracked target is matched and filtered with the radar-tracked target. If the deviation between the target motion attributes of the image-tracked target and the target motion attributes of the radar-tracked target is within a preset deviation range, the image-tracked target and the radar-tracked target are associated as a fused target.

[0026] According to a second aspect disclosed in this application, this application provides an adaptive cruise target selection device, comprising:

[0027] The image detection module is used to acquire image detection data of the front of the vehicle through the vehicle-mounted camera, and to obtain the image tracking target and the current lane where the vehicle is located based on the image detection data;

[0028] The radar monitoring module is used to acquire radar detection data from the front of the vehicle through the vehicle-mounted radar, and to acquire radar tracking targets based on the radar detection data.

[0029] The target fusion module is used to fuse the image-tracked target and the radar-tracked target to obtain the fused target;

[0030] The target filtering module is used to filter fused targets that continuously meet the target filtering conditions within a preset time period as potential following targets; wherein, the target filtering conditions include world layer filtering conditions and pixel layer filtering conditions; and to filter potential following targets with the smallest longitudinal distance to the vehicle as cruise following targets.

[0031] According to a third aspect disclosed in this application, this application provides an electronic device, including a processor and a memory communicatively connected to the processor;

[0032] The memory stores computer-executed instructions;

[0033] The processor executes computer execution instructions stored in the memory to implement the method described in any one of the first aspects.

[0034] According to the fourth aspect disclosed in this application, this application provides a computer-readable storage medium storing computer-executable instructions, which, when executed, are used to implement the method described in any one of the first aspects.

[0035] According to the fifth aspect disclosed in this application, this application provides a computer program product, including a computer program, which, when executed, is used to implement the method described in any one of the first aspects.

[0036] Compared with the prior art, this application has the following advantages:

[0037] This application provides a vehicle adaptive cruise target selection method, apparatus, device, and storage medium that utilizes the characteristic that the relative positional relationship of targets at the pixel level in an image is not affected by the coordinate system calibration transformation accuracy. It combines pixel-level selection and world-level selection in a complementary manner, and performs multiple selection judgments through pixel-level selection and world-level selection, which effectively improves the accuracy of adaptive cruise target selection. Attached Figure Description

[0038] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0039] Figure 1 A flowchart illustrating an adaptive cruise target selection method provided in this application embodiment;

[0040] Figure 2 A flowchart illustrating another adaptive cruise target selection method provided in this application embodiment;

[0041] Figure 3 This is a schematic diagram illustrating the target intrusion determination provided in an embodiment of this application.

[0042] Figure 4 A schematic diagram of the target screening area provided in the embodiments of this application;

[0043] Figure 5 This is a schematic diagram of the structure of an adaptive cruise target selection device provided in an embodiment of this application;

[0044] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0045] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0046] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0047] Adaptive Cruise Control (ACC) is an intelligent automatic control system designed to improve driving comfort and safety. Developed from traditional cruise control technology, ACC allows the vehicle's cruise control system to adjust speed to adapt to traffic conditions. The system relies on onboard sensors (cameras, radar, etc.) mounted at the front of the vehicle to gather information. When the road ahead is clear, ACC maintains the set maximum cruise speed; when there is a vehicle ahead, ACC reduces speed as needed to maintain a distance from the vehicle ahead based on a selected time interval until a suitable cruise speed is reached. As a result, adaptive cruise control can significantly reduce driver fatigue during long-distance driving.

[0048] During the operation of a vehicle's adaptive cruise control system, target selection is a crucial step, determining which vehicles or obstacles will be considered by the vehicle's cruise control system as objects that need to be tracked or at a safe distance.

[0049] However, existing target selection methods suffer from low accuracy in cruise control target selection. This is because existing methods rely on the relative position between the target vehicle and the cruise vehicle for target selection. Before the target data and lane line data acquired by the vehicle's onboard sensors (onboard cameras, onboard radar, etc.) are output to the target selection module of the vehicle's cruise control system, they generally need to undergo coordinate transformation steps to convert the position of the target vehicle and lane lines to a world-level coordinate system with the cruise vehicle's front position (or the onboard sensor) as the origin, in order to obtain the relative position between the target vehicle and the cruise vehicle.

[0050] In this process, the accuracy of the vehicle sensor calibration largely determines the accuracy of target selection. In complex road conditions, vehicles are prone to frequent bumps and vibrations, causing changes in vehicle position and even the installation position of vehicle sensors. This results in a deviation between the data of the target vehicle and lane lines after coordinate transformation and the actual position, causing ACC to misfollow or miss targets, or even cause collisions. This significantly affects the accuracy of adaptive cruise control and the safety of the system.

[0051] To address the aforementioned technical issues, this application proposes an adaptive cruise target selection method that effectively improves the accuracy of adaptive cruise target selection by combining pixel-level selection and world-level selection.

[0052] The technical solution for adaptive cruise target selection provided in this application will be described in detail below through specific embodiments. It should be noted that the following embodiments may exist independently or in combination with each other, and the same or similar content may not be described again in different embodiments.

[0053] It should be noted that the adaptive cruise target selection method provided in this application embodiment is executed by the vehicle cruise control system, and correspondingly, the adaptive cruise target selection device is also set in the vehicle cruise control system.

[0054] Figure 1 This is a flowchart illustrating an adaptive cruise target selection method provided in an embodiment of this application. (See attached diagram.) Figure 1 In some embodiments, the adaptive cruise target selection method includes the following steps:

[0055] S101 acquires image detection data of the front of the vehicle through the vehicle-mounted camera, and obtains the image tracking target and the current lane of the vehicle based on the image detection data.

[0056] When the vehicle activates adaptive cruise control, the onboard camera can collect image data in front of the vehicle and use the collected image data as detection data to detect and track target objects (vehicles, pedestrians, obstacles, etc.) in front of the vehicle, as well as detect the current lane in which the vehicle is traveling.

[0057] Specifically, vehicle cameras are typically installed at the front of the vehicle to capture images of the road and environment in front of the vehicle.

[0058] Specifically, after acquiring image detection data, the vehicle-mounted camera can also preprocess the acquired images, such as denoising and enhancing contrast, to improve image quality.

[0059] S102 acquires radar detection data from the front of the vehicle through the vehicle-mounted radar, and acquires radar tracking targets based on the radar detection data.

[0060] When the vehicle activates adaptive cruise control, the onboard radar can collect radar detection data in front of the vehicle and use this data to sense and track target objects in front of the vehicle.

[0061] Specifically, vehicle-mounted radar can be millimeter-wave radar. Millimeter-wave radar transmits millimeter waves through an antenna. These millimeter waves are reflected back when they encounter a target object, and the radar antenna then receives these reflected signals. Based on the time difference between the transmitted and received signals, as well as the frequency changes of the signals, the radar system can calculate information such as the target's distance, speed, and angle.

[0062] Meanwhile, the vehicle-mounted radar system can process the received signals, identifying the type (such as vehicles, pedestrians, etc.), position, and speed of objects ahead. After identifying an object, the radar system begins tracking it. The tracking process involves continuous observation and position updates of the object to ensure the system can monitor its motion in real time.

[0063] S103 fuses the image-tracked target and the radar-tracked target to obtain the fused target.

[0064] Image-based target tracking provides rich visual information, such as the target's shape, color, and texture, which is beneficial for accurate target identification, classification, and tracking. However, image recognition technology is easily affected by factors such as lighting, shadows, and occlusion, leading to a decrease in the accuracy and stability of target tracking. Radar technology, on the other hand, is unaffected by lighting and weather conditions, can operate stably in various environments, and has a long detection range and high-precision measurement capabilities. However, radar data is typically sparse, lacking detailed visual information about the target, which may result in insufficient accuracy in target identification and tracking in certain situations.

[0065] Therefore, by fusing image-tracked targets from image recognition and radar-tracked targets from radar recognition, we can fully leverage the advantages of both images and radar to improve the accuracy and robustness of target detection and tracking.

[0066] S104, filter out fusion targets that continuously meet the target filtering conditions within a preset time period as potential car-following targets; wherein, the target filtering conditions include pixel layer filtering conditions and world layer filtering conditions.

[0067] Among them, the pixel layer refers to the pixel coordinate system level on the image detection data, while the world layer refers to the world coordinate system level constructed with the vehicle camera as the origin. Therefore, under complex road conditions, the world layer coordinates may deviate due to bumps and vibrations, resulting in a decrease in the accuracy of target selection.

[0068] However, at the pixel level, the relative positions of the tracked target, lane lines, and cruise vehicle in the image remain unaffected. Therefore, by combining pixel-level and world-level filtering conditions, pixel-level filtering can be performed at the pixel level of the image detection data. The filtering results are unaffected by the calibration of the onboard sensors, thereby improving the accuracy and robustness of target selection.

[0069] Specifically, potential following targets should be selected by prioritizing those in the same lane as the cruise vehicle. Vehicles in the same lane typically have similar speeds and directions, making the following process more stable. Furthermore, the relative position and speed changes between vehicles in the same lane and the cruise vehicle are more direct and closely observed. Following vehicles in the same lane makes it easier to predict their trajectories and potential actions (such as braking, acceleration, lane changes, etc.), allowing for proactive reactions and ensuring driving safety.

[0070] S105 filters out potential following targets with the smallest longitudinal distance from the vehicle and uses them as the cruise following targets.

[0071] Maintaining lane stability is crucial for driving safety during adaptive cruise control. Selecting the vehicle with the shortest longitudinal distance as the following target ensures the cruise control system responds promptly to changes in the vehicle ahead, such as sudden braking or acceleration, thus preventing potential collisions. Simultaneously, the cruise control system intelligently adjusts the vehicle speed to maintain a safe distance, avoiding rear-end collisions caused by following too closely. Furthermore, it helps the vehicle stay stable within the current lane, reducing safety hazards associated with lane changes.

[0072] In this embodiment, taking advantage of the fact that the relative positional relationship of targets at the pixel level in an image is not affected by the accuracy of coordinate system calibration transformation, pixel-level screening and world-level screening are combined and complemented. Through multiple screening judgments by pixel-level screening and world-level screening, the accuracy of target screening in adaptive cruise is effectively improved.

[0073] exist Figure 1 Based on the embodiments shown, the following is combined with Figure 2 The technical solution of the above-mentioned adaptive cruise target selection method will be further introduced.

[0074] Figure 2 A flowchart illustrating another adaptive cruise target selection method provided in this application embodiment is shown below. Figure 2 In some embodiments, the adaptive cruise target selection method includes the following steps:

[0075] S201 acquires image detection data from the front of the vehicle using an onboard camera.

[0076] S202, input the image detection data into the detection and recognition model to obtain image tracking target and lane line point data; wherein, the detection and recognition model includes a tracking target recognition sub-model and a lane recognition sub-model obtained based on a neural network model training.

[0077] Specifically, by collecting a large amount of information on vehicles, pedestrians, and lane lines on real roads, a target recognition sub-model and a lane recognition sub-model are trained using a neural network model for target recognition and lane line recognition, respectively.

[0078] Specifically, the target tracking and recognition sub-model can use models such as YOLO and R-CNN.

[0079] Specifically, the lane recognition sub-model can use the CIrNet model, which obtains the initial lane line point data of the world layer after preprocessing and world coordinate transformation.

[0080] S203, Track the target in the image based on the target tracking model, and obtain the target bounding box and target motion attributes of the target in the image; wherein, the target tracking model is constructed based on the visual target tracking algorithm, and the target motion attributes include the longitudinal and lateral distances from the vehicle, and the longitudinal and lateral velocities relative to the vehicle.

[0081] The target tracking model is constructed using a visual target tracking algorithm. It first employs a tracking method combining a matching loss matrix and Hungarian matching. When the target detection output is stable, target tracking is stable. Since it does not involve an AI inference model, the module is simple and easy to use. Target distance estimation uses UKF Kalman filtering. Based on different conditions, it uses a hybrid ranging method of width measurement and landing point measurement for distance calculation. Then, UKF Kalman filtering is applied to the same target for position and velocity filtering, outputting the final target result. The main outputs include the category, position, and velocity of the tracked target (such as pedestrians or vehicles), and the target bounding box and its size are annotated on the image detection data.

[0082] Specifically, for the target motion attributes of image tracking targets, X represents the lateral distance to the vehicle, and Y represents the longitudinal distance to the vehicle. This indicates the lateral velocity relative to the vehicle itself. This indicates the longitudinal speed relative to the vehicle.

[0083] S204, Perform curve fitting and filtering tracking processing on the lane line point data to obtain the lane lines and lane information of the current lane; wherein, the lane information includes the lane width and the distance of the image tracking target from the center line of the current lane.

[0084] After obtaining the lane line point data, curve fitting is performed on the lane line point data to obtain the fitted lane lines. At the same time, relevant information about the lane lines is further obtained based on the lane lines, and the lane lines are displayed on the image of the image detection data.

[0085] Specifically, the curve fitting employs a cubic curve algorithm. This algorithm can accurately fit data points, especially when the number of data points is large and their distribution is complex; it maintains a good fit even then. By solving for the cubic polynomial coefficients between each data point, it ensures that the curve passes through all data points, thus achieving a high-precision fit and obtaining a high-accuracy lane line. The fitted lane line satisfies the following formula:

[0086]

[0087] Where y represents the ordinate of the lane line and x represents the abscissa of the lane line. Represents the polynomial coefficients.

[0088] Specifically, let's assume the lane lines on the left and right sides of the current lane are respectively... Then the curve formula for the two lane lines will be... Adding them together, we can further obtain the lane width as follows: , and the distance d from the image tracking target to the centerline of the current lane.

[0089] Specifically, filtering tracking involves using filtering algorithms to predict and update the target, thereby achieving continuous tracking. These filtering algorithms include Kalman filtering, Meanshift tracking, correlation filtering, particle filtering, and deep learning-based filtering tracking algorithms.

[0090] S205, if any point on the bottom edge of the target box is within the lane line of the current lane, then it is determined that the image tracking target is within the current lane, and the intrusion rate of the image tracking target intruding into the current lane is obtained; where the intrusion rate is the ratio between the area of ​​the target box intruding into the current lane and the area of ​​the target box.

[0091] At the pixel level, the system determines whether the target has intruded into the current lane of the cruising vehicle by judging whether any point on the bottom edge of the target box is within the lane line of the current lane, and calculates the intrusion rate by the intrusion area.

[0092] For example, see Figure 3Taking the lane between the two middle lane lines as the current lane, if any point on the bottom edge of the target bounding box of the tracked object lies within the convex polygon formed by the four points of the current lane lines, then the tracked object is considered likely to be within the current lane. Then, the intersection point is calculated based on the area formed by the target bounding box points and the lane line points, and the intersection area (i.e., ...) is calculated based on the intersection point. Figure 3 The area of ​​the shadow portion of the target bounding box falling into the current lane is then compared with the area of ​​the target bounding box of the tracked target. Finally, the value of whether the tracked target is in the current lane and the percentage are assigned to the tracked target's attributes.

[0093] Specifically, InLaneFlag is used to indicate whether the tracked target has intruded into the current lane. If the tracked target intrudes into the current lane, InLaneFlag is set to 1; if the tracked target has not intruded into the current lane, InLaneFlag is set to 0. InLaneRate is used to represent the intrusion rate, which is the percentage of the tracked target that has intruded into the current lane. <InLaneRate≤1。

[0094] S206, if any point on the bottom edge of the target box is not in the current lane, then it is determined that the image tracking target is not in the current lane.

[0095] If any point on the bottom edge of the target bounding box is not in the current lane, it means that the image tracking target has not invaded the vehicle's current lane, but is in another lane next to it.

[0096] The S207 acquires radar detection data from the front of the vehicle through its onboard radar and uses this data to track targets.

[0097] S208, matching and filtering the image-tracked target with the radar-tracked target, and when the deviation between the target motion attributes of the image-tracked target and the target motion attributes of the radar-tracked target is within a preset deviation range, the image-tracked target and the radar-tracked target are associated as a fused target.

[0098] The process involves receiving target tracking results from both the vehicle-mounted radar and the vehicle-mounted camera, comparing their timestamps, and selecting the most recent data for matching and filtering. Once the horizontal and vertical positions and velocities of the image-tracked target and the radar-tracked target meet threshold ranges and a certain duration is reached, the image-tracked target and the radar-tracked target are associated as a fused target.

[0099] After target fusion, the relevant attributes of the image-tracked target and the radar-tracked target are also combined, and the fused target also has parameters such as intrusion identification and intrusion rate that the image-tracked target possesses.

[0100] S209, select fusion targets that continuously meet the target selection criteria within a preset time period as potential follow-up targets; among which, the target selection criteria include pixel layer selection criteria and world layer selection criteria.

[0101] Preferably, the pixel layer screening conditions include the fusion target being in the current lane and the invasive rate of the fusion target entering the current lane being greater than a first judgment threshold;

[0102] The world layer screening criteria include the target's distance from the centerline of the current lane being less than the tightening width.

[0103] In complex road conditions, the target location information and lane line parameters at the world layer are easily affected by road bumps, leading to inconsistencies with the actual coordinates in some special cases, resulting in mis-screening. Therefore, in addition to this, pixel layer screening conditions are used to compensate for the world layer screening, so as to screen out potential following targets that are in the same lane as the cruise vehicle as accurately as possible.

[0104] Specifically, the fusion target is located within the current lane, and the intrusion rate of the fusion target into the current lane is greater than the first judgment threshold, which specifically satisfies the following formula:

[0105]

[0106]

[0107] In this context, InLaneFlag represents the intrusion flag, and InLaneRate represents the intrusion rate. This indicates the first judgment threshold.

[0108] Specifically, the distance between the fusion target and the centerline of the current lane is less than the tightening width, satisfying the following formula:

[0109]

[0110] Where d represents the distance of the fusion target from the centerline of the current lane. Indicates tightening the width. Indicates lane width, This represents the tightening coefficient.

[0111] For example, see Figure 4 Assuming the cruise vehicle in the diagram is A, and the distance between the fusion target B and the center line of the current lane of cruise vehicle A is 4 meters, and the current lane width is 3 meters, then the distance between the fusion target B and the center line of the current lane, with a tightening coefficient of 0.8, is 4 meters, which is greater than 3.2 meters. Clearly, the fusion target B is outside the current lane of cruise vehicle A.

[0112] In the diagram, the distance between the merged target C and the center line of the current lane of the cruise vehicle A is 0 meters. Since 0 meters is less than 3.2 meters, the merged target C is clearly within the current lane of the cruise vehicle A.

[0113] Specifically, the tightening coefficient is linearly related to the longitudinal distance between the target and the vehicle. The closer the longitudinal distance, the closer the tightening coefficient is to 1, and the farther the longitudinal distance, the closer the tightening coefficient is to 0.

[0114] Preferably, the pixel layer screening conditions include the fusion target being in the current lane and the invasive rate of the fusion target entering the current lane being greater than a second judgment threshold;

[0115] The world layer screening criteria include the minimum of the longitudinal distance between the target vehicle and the vehicle being less than the minimum turning radius screening threshold and the minimum of the current lane width.

[0116] In close-range situations, due to the installation angle of the vehicle-mounted camera, the target bounding box detected at close range will be much larger than the actual size of the target, resulting in errors in the pixel layer screening results. Therefore, the pixel layer screening is compensated by using world layer screening conditions to screen potential following targets in the same lane as the cruise vehicle as accurately as possible.

[0117] Meanwhile, by complementing pixel-level filtering and world-level filtering and making multiple judgments, the probability of misidentifying following a car is reduced when following a car on a curve, and the requirements for the accuracy of lane line algorithm fitting are also reduced.

[0118] Specifically, the fusion target is located within the current lane, and the intrusion rate of the fusion target into the current lane is greater than the second judgment threshold, which specifically satisfies the following formula:

[0119]

[0120]

[0121] In this context, InLaneFlag represents the intrusion flag, and InLaneRate represents the intrusion rate. This indicates the second judgment threshold.

[0122] Specifically, the minimum value between the lateral distance between the target vehicle and the vehicle and the minimum turning radius screening threshold and the current lane width is satisfied, according to the following formula:

[0123]

[0124] in, Represents the absolute value of the horizontal distance. This represents the minimum turning radius screening threshold. Indicates the width of the lane.

[0125] See Figure 4 As can be seen, if the fused target is outside the turning radius area formed by the turning of the cruise vehicle, it indicates that the fused target will not affect the driving of the cruise vehicle and will not be regarded as a potential following target, thus eliminating the possibility of incorrect judgment in pixel layer screening under close distance conditions.

[0126] For the two target screening conditions mentioned above, a target only needs to pass either screening condition and remain so for a certain period of time to be considered a potential follower target.

[0127] Preferably, before target filtering, outlier processing can be performed on the received target data and lane line data. Then, it can be determined whether the lane lines are complete. If the lane lines are incomplete, the electronic fence filtering mode based on steering wheel angle is entered. Only when the lane lines are complete is the filtering mode combining pixel-level filtering and world-level filtering entered.

[0128] Specifically, the electronic fence filtering mode refers to the system first setting one or more predefined areas, which can be key traffic nodes such as roads, intersections, and highway entrances. These areas are marked on an electronic map, and corresponding boundaries and rules are set. When a vehicle enters a predefined area, the cruise control system automatically adjusts the following target based on the current traffic conditions and the set speed and distance.

[0129] S210 filters out potential targets that have the smallest longitudinal distance from the vehicle and uses them as the cruise following targets.

[0130] Specifically, after determining the target vehicle to be followed, the cruise control system calculates the required throttle and brake control values ​​and then transmits the information to the chassis for longitudinal control of the entire vehicle, thereby achieving the adaptive cruise function.

[0131] In this embodiment, the image detection end determines whether the target is in the current lane of the vehicle based on the raw image data and calculates the target's intrusion rate into the current lane. The entire process is not affected by external parameter conversion, which reduces the requirements for the calibration accuracy of the vehicle camera.

[0132] Figure 5 This is a schematic diagram of the structure of an adaptive cruise target selection device provided in an embodiment of this application. (See attached diagram.) Figure 5 The adaptive cruise target selection device includes various functional modules for implementing the aforementioned adaptive cruise target selection method, and any functional module can be implemented by software and / or hardware.

[0133] In some embodiments, the adaptive cruise target selection device 500 includes an image detection module 501, a radar monitoring module 502, a target fusion module 503, and a target selection module 504. Wherein:

[0134] The image detection module 501 is used to acquire image detection data of the front of the vehicle through the vehicle-mounted camera, and to acquire the image tracking target and the current lane where the vehicle is located based on the image detection data;

[0135] The radar monitoring module 502 is used to acquire radar detection data from the front of the vehicle through the vehicle-mounted radar, and to acquire radar tracking targets based on the radar detection data;

[0136] The target fusion module 503 is used to fuse image-tracked targets and radar-tracked targets to obtain fused targets;

[0137] The target selection module 504 is used to select fused targets that continuously meet the target selection conditions within a preset time period as potential following targets; wherein, the target selection conditions include world layer selection conditions and pixel layer selection conditions; and to select potential following targets with the smallest longitudinal distance from the vehicle as cruise following targets.

[0138] In some embodiments, the pixel layer filtering conditions include the fusion target being in the current lane and the invasive rate of the fusion target entering the current lane being greater than a first judgment threshold.

[0139] The world layer screening criteria include the target's distance from the centerline of the current lane being less than the tightening width.

[0140] In some embodiments, the pixel layer filtering conditions include the fusion target being in the current lane and the invasive rate of the fusion target entering the current lane being greater than a second judgment threshold.

[0141] The world layer screening criteria include the minimum value among the lateral distance between the target vehicle and the vehicle being less than the minimum turning radius screening threshold and the lane width of the current lane.

[0142] In some embodiments, the image detection module 501 is specifically used for:

[0143] Image detection data is input into the detection and recognition model to obtain image tracking target and lane line point data; the detection and recognition model includes a tracking target recognition sub-model and a lane recognition sub-model trained based on a neural network model;

[0144] The target tracking model is used to track the target in the image and obtain the target bounding box and target motion attributes. The target tracking model is built based on the visual target tracking algorithm. The target motion attributes include the longitudinal and lateral distances from the vehicle, and the longitudinal and lateral velocities relative to the vehicle.

[0145] The lane line point data is subjected to curve fitting and filtering tracking to obtain the lane lines and lane information of the current lane; the lane information includes the lane width and the distance of the image tracking target from the center line of the current lane.

[0146] In some embodiments, the image detection module 501 is further configured to:

[0147] If any point on the bottom edge of the target bounding box is within the lane line of the current lane, then the image tracking target is determined to be within the current lane, and the intrusion rate of the image tracking target into the current lane is obtained; where the intrusion rate is the ratio between the area of ​​the target bounding box that intrudes into the current lane and the area of ​​the target bounding box.

[0148] In some embodiments, the image detection module 501 is further configured to:

[0149] If no point on the bottom edge of the target bounding box is within the current lane, then the image tracking target is determined to be outside the current lane.

[0150] In some embodiments, the target fusion module 503 is specifically used for:

[0151] The image-tracked target is matched and filtered with the radar-tracked target. If the deviation between the target motion attributes of the image-tracked target and the target motion attributes of the radar-tracked target is within a preset deviation range, the image-tracked target and the radar-tracked target are associated as a fused target.

[0152] The adaptive cruise target selection device 500 provided in this application embodiment is used to execute the technical solution provided in the aforementioned adaptive cruise target selection method embodiment. Its implementation principle and technical effect are similar to those in the aforementioned method embodiment, and will not be repeated here.

[0153] It should be noted that the division of the various modules in the above device is merely a logical functional division. In actual implementation, they can be fully or partially integrated into a single physical entity, or they can be physically separated. Furthermore, these modules can be implemented entirely in software via processing elements, entirely in hardware, or partially in software via processing elements and partially in hardware. For example, the target screening module 504 can be a separate processing element, or it can be integrated into a chip in the above device. Alternatively, it can be stored as program code in the memory of the above device, and its function can be called and executed by a processing element. The implementation of other modules is similar. Moreover, these modules can be fully or partially integrated together, or they can be implemented independently. The processing element here can be an integrated circuit with signal processing capabilities. In the implementation process, each step of the above method or each of the above modules can be completed through integrated logic circuits in the hardware of the processor element or through software instructions.

[0154] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. (See attached diagram.) Figure 6 The electronic device 600 includes a processor 601 and a memory 602 communicatively connected to the processor 601;

[0155] Memory 602 stores computer-executed instructions;

[0156] The processor 601 executes computer execution instructions stored in the memory 602 to implement the aforementioned technical solution of the adaptive cruise target selection method.

[0157] In the aforementioned electronic device 600, the memory 602 and the processor 601 are electrically connected directly or indirectly to achieve data transmission or interaction. For example, these components can be electrically connected to each other through one or more communication buses or signal lines, such as bus connections. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be classified as address buses, data buses, control buses, etc., but this does not mean that there is only one bus or one type of bus. The memory 602 stores computer execution instructions for implementing the aforementioned adaptive cruise target selection method, including at least one software functional module that can be stored in the memory 602 in the form of software or firmware. The processor 601 executes various functional applications and data processing by running the software programs and modules stored in the memory 602.

[0158] The memory 602 includes at least one type of readable storage medium, not limited to Random Access Memory (RAM), Read Only Memory (ROM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), etc. The memory 602 stores programs, and the processor 601 executes the programs after receiving execution instructions. Furthermore, the software programs and modules within the memory 602 may also include an operating system, which may include various software components and / or drivers for managing system tasks (e.g., memory management, storage device control, power management, etc.) and can communicate with various hardware or software components to provide an operating environment for other software components.

[0159] Processor 601 can be an integrated circuit chip with signal processing capabilities. The aforementioned processor 601 can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), etc. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor, or processor 601 can be any conventional processor.

[0160] The electronic device 600 is used to execute the technical solution provided in the aforementioned adaptive cruise target selection method embodiment. Its implementation principle and technical effect are similar to those in the aforementioned method embodiment, and will not be repeated here.

[0161] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed, are used to implement the technical solution of the aforementioned adaptive cruise target selection method.

[0162] The aforementioned computer-readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The computer-readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0163] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Alternatively, the readable storage medium can be an integral part of the processor. Both the processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components within the control unit of an adaptive cruise target selection device.

[0164] This application also provides a computer program product, including a computer program, which, when executed, is used to implement the technical solution of the aforementioned adaptive cruise target selection method.

[0165] In the above embodiments, those skilled in the art will understand that the above method embodiments can be implemented entirely or partially by software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented entirely or partially in the form of a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of this application is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless network, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid-state drives (SSDs)).

[0166] In the above embodiments, the descriptions of each embodiment have their own emphasis. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments. The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0167] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this application are indicated by the appended claims.

[0168] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. An adaptive cruise target selection method, characterized in that, include: The vehicle acquires image detection data of the front of the vehicle using an onboard camera, and obtains the image tracking target and the current lane where the vehicle is located based on the image detection data; The vehicle-mounted radar acquires radar detection data from the front of the vehicle, and the radar tracking target is obtained based on the radar detection data. The image-tracked target and the radar-tracked target are fused to obtain the fused target; Filtration targets that continuously meet the target screening conditions within a preset time period are selected as potential vehicle-following targets; wherein, the target screening conditions include pixel-level screening conditions and world-level screening conditions; Select the potential vehicle with the smallest longitudinal distance from your own vehicle as the cruise following target; The pixel layer filtering conditions include the fusion target being within the current lane and the fusion target's intrusion rate into the current lane being greater than a first judgment threshold; the world layer filtering conditions include the fusion target's distance from the centerline of the current lane being less than the tightening width; Alternatively, the pixel layer filtering conditions include the fusion target being within the current lane and the invasive rate of the fusion target entering the current lane being greater than a second judgment threshold; the world layer filtering conditions include the minimum value between the lateral distance between the fusion target and the vehicle being less than the minimum turning radius filtering threshold and the lane width of the current lane.

2. The method according to claim 1, characterized in that, Based on the image detection data, the image tracking target and the current lane where the vehicle is located are obtained, including: The image detection data is input into the detection and recognition model to obtain image tracking target and lane line point data; wherein, the detection and recognition model includes a tracking target recognition sub-model and a lane recognition sub-model obtained based on a neural network model training; The image tracking target is tracked based on the target tracking model, and the target bounding box and target motion attributes of the image tracking target are obtained; wherein, the target tracking model is constructed based on the visual target tracking algorithm, and the target motion attributes include the longitudinal and lateral distances from the vehicle, and the longitudinal and lateral velocities relative to the vehicle; The lane line point data is subjected to curve fitting and filtering tracking processing to obtain the lane lines and lane information of the current lane; wherein, the lane information includes the lane width and the distance of the image tracking target from the center line of the current lane.

3. The method according to claim 2, characterized in that, The method further includes: If any point on the bottom edge of the target bounding box is within the lane line of the current lane, then the image tracking target is determined to be within the current lane, and the intrusion rate of the image tracking target into the current lane is obtained; wherein, the intrusion rate is the ratio between the area of ​​the target bounding box that intrudes into the current lane and the area of ​​the target bounding box.

4. The method according to claim 3, characterized in that, The method further includes: If any point on the bottom edge of the target frame is not within the current lane, then the image tracking target is determined to be outside the current lane.

5. The method according to claim 2, characterized in that, The image-tracked target and the radar-tracked target are fused to obtain a fused target, including: The image-tracked target is matched and filtered with the radar-tracked target. If the deviation between the target motion attributes of the image-tracked target and the target motion attributes of the radar-tracked target is within a preset deviation range, the image-tracked target and the radar-tracked target are associated as a fused target.

6. An adaptive cruise target selection device, characterized in that, include: The image detection module is used to acquire image detection data of the front of the vehicle through the vehicle-mounted camera, and to obtain the image tracking target and the current lane where the vehicle is located based on the image detection data; The radar monitoring module is used to acquire radar detection data from the front of the vehicle through the vehicle-mounted radar, and to acquire radar tracking targets based on the radar detection data. The target fusion module is used to fuse the image-tracked target and the radar-tracked target to obtain the fused target; The target filtering module is used to filter fused targets that continuously meet the target filtering conditions within a preset time period as potential following targets; wherein, the target filtering conditions include world layer filtering conditions and pixel layer filtering conditions; and to filter potential following targets with the smallest longitudinal distance to the vehicle as cruise following targets. The pixel layer filtering conditions include the fusion target being within the current lane and the fusion target's intrusion rate into the current lane being greater than a first judgment threshold; the world layer filtering conditions include the fusion target's distance from the centerline of the current lane being less than the tightening width; Alternatively, the pixel layer filtering conditions include the fusion target being within the current lane and the invasive rate of the fusion target entering the current lane being greater than a second judgment threshold; the world layer filtering conditions include the minimum value between the lateral distance between the fusion target and the vehicle being less than the minimum turning radius filtering threshold and the lane width of the current lane.

7. An electronic device, characterized in that, Includes a processor and a memory communicatively connected to the processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory to implement the method as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed, are used to implement the method as described in any one of claims 1 to 5.

Images