Vehicle braking method and system based on false cut-out identification

Abstract

The invention discloses a vehicle braking method and system based on false cut-out recognition, and the method comprises the steps: obtaining the transverse speed of a front vehicle in a lane where a target vehicle is located, and obtaining a cut-out behavior recognition result of the front vehicle according to the transverse speed, the longitudinal relative vehicle speed, the transverse distance and the longitudinal distance between the front vehicle and the target vehicle are obtained in real time; the acceleration of an adjacent vehicle located in front of the front vehicle in an adjacent lane is obtained according to the cut-out behavior recognition result; according to the longitudinal relative vehicle speed, the transverse distance, the longitudinal distance and the acceleration, a false cut-out behavior recognition result of the front vehicle is obtained; and the longitudinal acceleration of the front vehicle is obtained according to the false cut-out behavior recognition result, so that the target vehicle is controlled to execute emergency braking operation according to the longitudinal acceleration and the transverse distance, and the vehicle driving safety is remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of vehicle control technology, and specifically to a vehicle braking method and system based on false cut-out recognition. Background Technology

[0002] When a vehicle is traveling straight in its own lane, if a vehicle in front is preparing to cut out or is in the process of cutting out, under normal circumstances, if the longitudinal speed and lateral speed of the vehicle in front remain constant during the cutting-out process, the vehicle will usually not brake and will continue to move forward at the same longitudinal speed as the vehicle cutting out. However, if the vehicle in front of the vehicle cutting out (whose front has entered the adjacent lane) slows down or brakes suddenly, the vehicle cutting out will also brake after a certain reaction time. At this point, the vehicle in front has not completed its cut-out and the parking space is still within the vehicle's lane. This scenario of incomplete cut-out (false cut-out) is very likely to cause a rear-end collision.

[0003] Existing methods for identifying false cut-out behavior mainly rely on assessing collision risk based on the longitudinal and lateral distances to the vehicle in front. However, after the vehicle in front has had a certain reaction time, the time left for the vehicle to make risk judgments is very limited, which can easily lead to rear-end collisions. This makes it impossible to identify false cut-out behavior in a timely manner, seriously affecting the safety of vehicle driving. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention discloses a vehicle braking method based on false cut-out recognition, which significantly improves vehicle driving safety.

[0005] To achieve the above objectives, this invention discloses a vehicle braking method based on false cut-out recognition, comprising: The lateral velocity of the vehicle in front in the lane where the target vehicle is located is obtained, so as to obtain the cut-out behavior recognition result of the vehicle in front based on the lateral velocity; Based on the cut-out behavior recognition results, the longitudinal relative speed, lateral distance and longitudinal distance between the vehicle in front and the target vehicle are obtained in real time. Based on the cut-out behavior recognition result, obtain the acceleration of the adjacent vehicle in the adjacent lane that is in front of the vehicle in front; The false cut-out behavior identification result of the vehicle in front is obtained based on the longitudinal relative vehicle speed, the lateral distance, the longitudinal distance and the acceleration; Based on the false cut-out behavior recognition result, the longitudinal acceleration of the vehicle in front is obtained, and the target vehicle is controlled to perform an emergency braking operation based on the longitudinal acceleration and the lateral distance.

[0006] This invention discloses a vehicle braking method based on false cut-out detection. By comprehensively acquiring the lateral velocity of the vehicle ahead, its longitudinal relative speed to the target vehicle, its lateral distance, its longitudinal distance, and the acceleration of vehicles in adjacent lanes, it can promptly and accurately identify false cut-out behavior by the vehicle ahead. Therefore, upon confirming the risk of a false cut-out, by combining the longitudinal acceleration and lateral distance of the vehicle ahead, it can quickly control the target vehicle to perform emergency braking, effectively avoiding rear-end collisions caused by false cut-outs and improving vehicle driving safety.

[0007] As a preferred example, the step of obtaining the lateral velocity of the vehicle in front in the lane where the target vehicle is located, and obtaining the cut-out behavior recognition result of the vehicle in front based on the lateral velocity, includes: According to a preset sampling frequency, a sequence of images of the vehicles ahead is acquired in real time by an in-vehicle image acquisition device preset in the target vehicle; wherein, the sequence of images of the vehicles ahead includes multiple consecutive images of the vehicles ahead. For any of the images of the vehicles ahead, obtain the coordinates of the rear center point of the vehicle ahead in the image of the vehicle ahead, so as to obtain the rear horizontal angle and rear pitch angle of the vehicle center point in the image of the vehicle ahead relative to the vehicle image acquisition device based on the rear image coordinates and the intrinsic parameter matrix of the vehicle image acquisition device. The longitudinal distance between the rear of the vehicle and the target vehicle in the image of the vehicle ahead is obtained based on the installation height of the vehicle-mounted image acquisition device and the rear pitch angle of the vehicle. The lateral distance between the rear of the vehicle and the target vehicle in the image of the vehicle in front is obtained based on the longitudinal distance of the rear of the vehicle and the horizontal angle. The difference in the lateral distance between the rear of the vehicle is obtained between any two adjacent images of the vehicle in front, so as to obtain the lateral velocity of the rear of the vehicle at each sampling time based on the difference in the lateral distance between the rear of the vehicle and the sampling frequency. Based on the preset window length, multiple lateral velocities of the vehicle rear within each time window are obtained to obtain a sequence of lateral velocities of the vehicle rear. When the difference between any two adjacent lateral velocities of the rear of the vehicle in the lateral velocity sequence is less than or equal to a preset difference threshold, it is determined that the vehicle in front has cut out.

[0008] The above-described scheme acquires a sequence of images of the vehicle ahead in real time using an onboard image acquisition device at a preset sampling frequency, ensuring the continuity and timeliness of data acquisition. Then, by accurately calculating the coordinates, horizontal angle, and pitch angle of the vehicle's rear center point in the image, and combining this with the intrinsic parameter matrix and installation height of the onboard image acquisition device, the longitudinal and lateral distances between the vehicle ahead and the target vehicle can be accurately calculated. Based on this, by analyzing the changes in the lateral distance between adjacent images and combining this with the sampling frequency, the lateral velocity of the vehicle ahead can be calculated in real time. Furthermore, by constructing a lateral velocity sequence within a preset window length and determining whether the difference between any two adjacent velocities in the sequence is less than or equal to a preset difference threshold, the scheme can effectively identify whether the vehicle ahead is attempting to cut out. This method, based on image sequences and geometric calculations, avoids the recognition delays and misjudgments caused by the lack of real-time and accurate lateral velocity information in traditional methods. This provides timely and reliable input for subsequent false cut-out behavior identification and emergency braking operations, significantly improving vehicle driving safety.

[0009] As a preferred example, the step of obtaining the longitudinal relative speed, lateral distance, and longitudinal distance between the vehicle in front and the target vehicle in real time based on the cut-out behavior recognition result includes: When it is determined that the vehicle in front is cutting out, the target cutting lane of the vehicle in front is located, and the rear end point of the side of the vehicle in front that is far away from the target cutting lane is taken as the proximal end point of the vehicle in front. For any of the images of the vehicle ahead, obtain the target bounding box of the vehicle ahead in the image, and determine the endpoint image coordinates of the near end point based on the target bounding box and the image coordinates of the rear of the vehicle. The horizontal and vertical angles of the near endpoint relative to the vehicle image acquisition device in the image of the vehicle ahead are obtained based on the endpoint image coordinates and the intrinsic parameter matrix. The near longitudinal distance between the near endpoint of the forward vehicle image and the target vehicle is obtained based on the installation height of the vehicle-mounted image acquisition device and the pitch angle. The proximal lateral distance between the proximal endpoint and the target vehicle is obtained based on the proximal longitudinal distance and the horizontal angle. Obtain the near-side longitudinal distance difference between any two adjacent images of the vehicle ahead, and obtain the longitudinal relative speed between the vehicle ahead and the target vehicle at each sampling time based on the near-side longitudinal distance difference and the sampling frequency.

[0010] The above scheme ensures that the acquired points accurately reflect the potential risks posed by the vehicle cutting out of the target lane during the lane change by locating the target vehicle's lane departure point and selecting the rear endpoint furthest from that lane as the near endpoint. Using the target bounding box and rear-end image coordinates to determine the image coordinates of the near endpoint improves the accuracy of endpoint positioning and avoids potential deviations caused by relying solely on the rear center point. Based on the endpoint image coordinates and intrinsic parameter matrix, the horizontal and pitch angles are calculated, transforming two-dimensional image information into angular information in three-dimensional space, providing a reliable foundation for subsequent distance calculations. Combining the installation height and pitch angle of the onboard image acquisition device to calculate the near longitudinal distance, and using the near longitudinal distance and horizontal angle to calculate the near lateral distance, real-time perception of the precise longitudinal and lateral positional relationship between the vehicle ahead and the target vehicle is achieved. Finally, the longitudinal relative speed is calculated by combining the difference in near longitudinal distance between adjacent images with the sampling frequency, ensuring continuous real-time updates of speed information. These precise and real-time parameter acquisition methods significantly improve the timeliness and accuracy of false cut-out behavior identification, providing reliable data support for subsequent emergency braking decisions, thereby effectively reducing the risk of rear-end collisions and enhancing vehicle driving safety.

[0011] As a preferred example, obtaining the acceleration of the adjacent vehicle in front of the vehicle ahead in the adjacent lane based on the cut-out behavior recognition result includes: The system acquires the adjacent vehicles in front of the vehicle in the target lane and obtains the coordinates of the center point of the rear of the adjacent vehicles at each sampling time according to a preset sampling frequency. The acceleration of the adjacent vehicle at each sampling moment is obtained based on the difference between the coordinates between two adjacent sampling times and the sampling frequency.

[0012] The aforementioned scheme accurately identifies adjacent vehicles in front of the vehicle cutting out of the target lane and continuously acquires the coordinates of their rear center points at a preset sampling frequency, providing high-precision, real-time raw data for subsequent acceleration calculations. Based on this, by calculating the coordinate differences between adjacent sampling times and combining this with the sampling frequency, the acceleration of adjacent vehicles at each sampling time can be quickly and effectively derived. This real-time acceleration information acquisition can promptly capture the deceleration or emergency braking behavior of adjacent vehicles, significantly shortening the risk assessment and response time in false lane-cutting scenarios. Compared to existing technologies that rely solely on longitudinal and lateral distances for collision risk assessment, introducing real-time acceleration information from adjacent vehicles allows for earlier prediction of potential hazards, buying valuable time for the target vehicle to perform emergency braking, effectively preventing rear-end collisions and greatly improving vehicle safety.

[0013] As a preferred example, the step of obtaining the false cut-out behavior recognition result of the vehicle ahead based on the longitudinal relative vehicle speed, the lateral distance, the longitudinal distance, and the acceleration includes: For any sampling time, obtain the first comparison result between the longitudinal distance of the rear of the vehicle and the preset longitudinal distance threshold; Obtain a second comparison result between the acceleration of adjacent vehicles and a preset acceleration threshold; Obtain the third comparison result between the proximal lateral distance and the preset lateral distance threshold; Obtain the fourth comparison result between the longitudinal relative vehicle speed and the preset vehicle speed threshold; Based on the first comparison result, the second comparison result, the third comparison result, and the fourth comparison result, the false cut-out behavior identification result of the vehicle in front at the sampling time is obtained.

[0014] The above scheme evaluates at any sampling moment, ensuring the continuity and real-time nature of identification and avoiding delays in risk assessment due to time constraints. The first comparison between the longitudinal distance to the rear of the vehicle and a longitudinal distance threshold determines whether the longitudinal position exceeds a safe range, aiding in assessing potential collision risk, as longitudinal distance directly reflects the proximity between vehicles. The second comparison between the acceleration of adjacent vehicles and an acceleration threshold detects changes in the behavior of vehicles in adjacent lanes, providing a warning signal, as changes in acceleration may indicate that the vehicle ahead is about to decelerate or brake. The third comparison between the near-side lateral distance and a lateral distance threshold assesses whether the vehicle's lateral position has deviated from its lane, crucial for identifying whether the lane change is complete, as changes in lateral distance indicate whether the vehicle has partially or completely left the lane. The fourth comparison between the longitudinal relative speed and a speed threshold determines whether the relative speed is too high, affecting the urgency level, as relative speed directly determines the collision time window. Finally, the false cut-out behavior identification result is obtained by combining all comparison results, avoiding misjudgment or omission that may be caused by single parameter evaluation. Multi-dimensional threshold comparison ensures more comprehensive and reliable identification, thereby triggering braking operation in time in false cut-out scenarios and reducing the risk of rear-end collision accidents.

[0015] As a preferred example, the step of obtaining the longitudinal acceleration of the vehicle ahead based on the false cut-out behavior recognition result, and controlling the vehicle to perform an emergency braking operation based on the longitudinal acceleration and the lateral distance, includes: When it is determined that the vehicle in front has made a false cut-out, the longitudinal acceleration of the vehicle in front is obtained; When the longitudinal acceleration is less than or equal to a preset acceleration threshold, the target vehicle is controlled to perform an emergency braking operation based on the third comparison result.

[0016] When the above-described scheme detects a false cut-out by a vehicle ahead, it further acquires the vehicle's longitudinal acceleration and combines this acceleration with the lateral distance to make a braking decision. This approach directly quantifies the dynamic deceleration behavior of the vehicle ahead, avoiding braking lag that may result from relying solely on static distance parameters. When the longitudinal acceleration of the vehicle ahead is less than or equal to a preset acceleration threshold, it indicates that it is undergoing significant deceleration. At this point, combining the third comparison result (i.e., the lateral distance comparison result) to determine whether emergency braking is necessary effectively avoids accidental braking when the lateral distance is safe, while ensuring timely response when danger approaches. This significantly improves the accuracy and timeliness of braking decisions, thereby effectively reducing the risk of rear-end collisions in false cut-out scenarios and ensuring vehicle safety.

[0017] On the other hand, the present invention discloses a vehicle braking system based on false cut-out recognition, including a cut-out recognition module, a forward vehicle recognition module, an adjacent vehicle recognition module, a false cut-out recognition module, and a braking module; The cut-out recognition module is used to obtain the lateral speed of the vehicle in front in the lane where the target vehicle is located, so as to obtain the cut-out behavior recognition result of the vehicle in front based on the lateral speed. The forward vehicle recognition module is used to obtain, in real time, the longitudinal relative speed, lateral distance and longitudinal distance between the forward vehicle and the target vehicle based on the cut-out behavior recognition result; The adjacent vehicle recognition module is used to obtain the acceleration of the adjacent vehicle in front of the vehicle in front in the adjacent lane based on the cut-out behavior recognition result; The false cutout recognition module is used to obtain the false cutout behavior recognition result of the vehicle in front based on the longitudinal relative vehicle speed, the lateral distance, the longitudinal distance and the acceleration; The braking module is used to obtain the longitudinal acceleration of the vehicle in front based on the false cut-out behavior recognition result, so as to control the target vehicle to perform an emergency braking operation based on the longitudinal acceleration and the lateral distance.

[0018] This invention discloses a vehicle braking system based on false cut-out detection. By comprehensively acquiring the lateral velocity of the vehicle ahead, its longitudinal relative speed to the target vehicle, its lateral distance, its longitudinal distance, and the acceleration of vehicles in adjacent lanes, it can promptly and accurately identify false cut-out behavior by the vehicle ahead. Therefore, upon confirming the risk of a false cut-out, by combining the longitudinal acceleration and lateral distance of the vehicle ahead, it can quickly control the target vehicle to perform emergency braking, effectively avoiding rear-end collisions caused by false cut-outs and improving vehicle driving safety.

[0019] As a preferred example, the cutout recognition module includes an image sequence unit, an angle recognition unit, a distance calculation unit, and a behavior recognition unit; The image sequence unit is used to acquire a sequence of images of the vehicles ahead in real time using an in-vehicle image acquisition device pre-installed in the target vehicle, according to a preset sampling frequency; wherein the sequence of images of the vehicles ahead includes multiple consecutive images of the vehicles ahead. The angle recognition unit is used to obtain the rear image coordinates of the rear center point of the vehicle in any image of the vehicle in front, so as to obtain the rear horizontal angle and rear pitch angle of the rear center point in the image of the vehicle in front relative to the vehicle image acquisition device based on the rear image coordinates and the intrinsic parameter matrix of the vehicle image acquisition device. The distance calculation unit is used to obtain the longitudinal distance between the front vehicle and the target vehicle in the front vehicle image based on the installation height of the vehicle-mounted image acquisition device and the rear vehicle pitch angle; and to obtain the lateral distance between the front vehicle and the target vehicle in the front vehicle image based on the longitudinal distance and the horizontal angle. The behavior recognition unit is used to obtain the difference in the lateral distance between the rear of any two adjacent images of the vehicle in front, so as to obtain the lateral speed of the rear of the vehicle at each sampling time according to the difference in the lateral distance between the rear of the vehicle and the sampling frequency; according to a preset window length, it obtains multiple lateral speeds of the rear of the vehicle within each time window to obtain a sequence of lateral speeds of the rear of the vehicle; when the difference between any two adjacent lateral speeds of the rear of the vehicle in the sequence of lateral speeds of the rear of the vehicle is less than or equal to a preset difference threshold, it is determined that the vehicle in front has cut out behavior.

[0020] The above-described scheme acquires a sequence of images of the vehicle ahead in real time using an onboard image acquisition device at a preset sampling frequency, ensuring the continuity and timeliness of data acquisition. Then, by accurately calculating the coordinates, horizontal angle, and pitch angle of the vehicle's rear center point in the image, and combining this with the intrinsic parameter matrix and installation height of the onboard image acquisition device, the longitudinal and lateral distances between the vehicle ahead and the target vehicle can be accurately calculated. Based on this, by analyzing the changes in the lateral distance between adjacent images and combining this with the sampling frequency, the lateral velocity of the vehicle ahead can be calculated in real time. Furthermore, by constructing a lateral velocity sequence within a preset window length and determining whether the difference between any two adjacent velocities in the sequence is less than or equal to a preset difference threshold, the scheme can effectively identify whether the vehicle ahead is attempting to cut out. This method, based on image sequences and geometric calculations, avoids the recognition delays and misjudgments caused by the lack of real-time and accurate lateral velocity information in traditional methods. This provides timely and reliable input for subsequent false cut-out behavior identification and emergency braking operations, significantly improving vehicle driving safety.

[0021] As a preferred example, the forward vehicle recognition module includes an endpoint positioning unit and an operating parameter recognition unit; The endpoint positioning unit is used to locate the target lane of the vehicle ahead when it is determined that the vehicle ahead is cutting out, so as to take the rear endpoint of the vehicle ahead on the side away from the target lane as the near endpoint of the vehicle ahead; for any image of the vehicle ahead, the target box of the vehicle ahead in the image is obtained, so as to determine the endpoint image coordinates of the near endpoint based on the target box and the rear image coordinates. The operating parameter identification unit is used to obtain the horizontal angle and pitch angle of the near endpoint relative to the vehicle-mounted image acquisition device in the image of the vehicle ahead based on the endpoint image coordinates and the intrinsic parameter matrix; to obtain the near longitudinal distance between the near endpoint and the target vehicle in the image of the vehicle ahead based on the installation height of the vehicle-mounted image acquisition device and the pitch angle; to obtain the near lateral distance between the near endpoint and the target vehicle based on the near longitudinal distance and the horizontal angle; and to obtain the near longitudinal distance difference between any two adjacent images of the vehicle ahead, so as to obtain the longitudinal relative speed between the vehicle ahead and the target vehicle at each sampling time based on the near longitudinal distance difference and the sampling frequency.

[0022] The above scheme ensures that the acquired points accurately reflect the potential risks posed by the vehicle cutting out of the target lane during the lane change by locating the target vehicle's lane departure point and selecting the rear endpoint furthest from that lane as the near endpoint. Using the target bounding box and rear-end image coordinates to determine the image coordinates of the near endpoint improves the accuracy of endpoint positioning and avoids potential deviations caused by relying solely on the rear center point. Based on the endpoint image coordinates and intrinsic parameter matrix, the horizontal and pitch angles are calculated, transforming two-dimensional image information into angular information in three-dimensional space, providing a reliable foundation for subsequent distance calculations. Combining the installation height and pitch angle of the onboard image acquisition device to calculate the near longitudinal distance, and using the near longitudinal distance and horizontal angle to calculate the near lateral distance, real-time perception of the precise longitudinal and lateral positional relationship between the vehicle ahead and the target vehicle is achieved. Finally, the longitudinal relative speed is calculated by combining the difference in near longitudinal distance between adjacent images with the sampling frequency, ensuring continuous real-time updates of speed information. These precise and real-time parameter acquisition methods significantly improve the timeliness and accuracy of false cut-out behavior identification, providing reliable data support for subsequent emergency braking decisions, thereby effectively reducing the risk of rear-end collisions and enhancing vehicle driving safety.

[0023] As a preferred example, the adjacent vehicle identification module includes a vehicle positioning unit and an acceleration identification unit; The vehicle positioning unit is used to obtain the adjacent vehicles located in front of the vehicle in the target lane, and to obtain the coordinates of the center point of the rear of the adjacent vehicles at each sampling time according to a preset sampling frequency. The acceleration identification unit is used to obtain the acceleration of the adjacent vehicle at each sampling time based on the difference between the coordinates between two adjacent sampling times and the sampling frequency.

[0024] The aforementioned scheme accurately identifies adjacent vehicles in front of the vehicle cutting out of the target lane and continuously acquires the coordinates of their rear center points at a preset sampling frequency, providing high-precision, real-time raw data for subsequent acceleration calculations. Based on this, by calculating the coordinate differences between adjacent sampling times and combining this with the sampling frequency, the acceleration of adjacent vehicles at each sampling time can be quickly and effectively derived. This real-time acceleration information acquisition can promptly capture the deceleration or emergency braking behavior of adjacent vehicles, significantly shortening the risk assessment and response time in false lane-cutting scenarios. Compared to existing technologies that rely solely on longitudinal and lateral distances for collision risk assessment, introducing real-time acceleration information from adjacent vehicles allows for earlier prediction of potential hazards, buying valuable time for the target vehicle to perform emergency braking, effectively preventing rear-end collisions and greatly improving vehicle safety. Attached Figure Description

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] Figure 1 This is a schematic flowchart of a vehicle braking method based on false cutout recognition disclosed in an embodiment of the present invention; Figure 2 This is a schematic diagram of a vehicle braking system based on false cut-out recognition disclosed in an embodiment of the present invention; Figure 3 This is a schematic diagram of a vehicle driving state disclosed in another embodiment of the present invention; Figure 4 This is a schematic diagram of a vehicle cutting-out behavior state disclosed in another embodiment of the present invention. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] Example 1 Reference Figure 1 To improve the timeliness of recognizing false cut-out behavior of vehicles ahead, this embodiment discloses a vehicle braking method based on false cut-out recognition, including: Step 101: Obtain the lateral speed of the vehicle in front in the lane where the target vehicle is located, so as to obtain the cut-out behavior recognition result of the vehicle in front based on the lateral speed.

[0029] Step 102: Based on the cut-out behavior recognition result, obtain in real time the longitudinal relative speed, lateral distance and longitudinal distance between the vehicle in front and the target vehicle.

[0030] Step 103: Based on the cut-out behavior recognition result, obtain the acceleration of the adjacent vehicle in the adjacent lane that is in front of the vehicle in front.

[0031] Step 104: Obtain the false cut-out behavior recognition result of the vehicle in front based on the longitudinal relative vehicle speed, the lateral distance, the longitudinal distance and the acceleration.

[0032] Step 105: Obtain the longitudinal acceleration of the vehicle in front based on the false cut-out behavior recognition result, and control the target vehicle to perform an emergency braking operation based on the longitudinal acceleration and the lateral distance.

[0033] In this embodiment, to address the issue of rear-end collisions caused by a vehicle suddenly cutting halfway into the adjacent lane and then immediately braking or stopping abruptly, the system identifies the driving behavior of vehicles in the adjacent lane ahead in a timely manner. This allows for timely prevention and determination of whether the vehicle's cut-out is a false one, and initiates active braking when necessary to avoid a rear-end collision. Specifically, by real-time monitoring and calculation of the acceleration of the vehicle in the adjacent lane ahead (the vehicle ahead of the vehicle preparing to cut into the adjacent lane, referred to here as: the vehicle ahead of the vehicle in the adjacent lane), when braking behavior is detected in the vehicle ahead of the vehicle in the adjacent lane, the cut-out behavior is determined to be a failure, indicating a false cut-out. The vehicle then immediately triggers emergency braking upon recognizing the braking of the vehicle ahead to avoid a rear-end collision.

[0034] In this embodiment, refer to Figure 3The driving state of the target vehicle shown is as follows. The target vehicle refers to the vehicle that is being controlled to perform the braking operation. Figure 3 The vehicle in front refers to the vehicle in the lane where the target vehicle is located, that is, the vehicle in front of the target vehicle. Figure 3 The preceding vehicle; lateral speed refers to the lateral movement speed of the preceding vehicle relative to the centerline of its current lane, and its changing trend can be used to determine whether the vehicle is changing lanes; lane-cutting behavior recognition result refers to the judgment result based on parameters such as the lateral speed of the preceding vehicle, determining whether the preceding vehicle is cutting out of the current lane or is about to do so; longitudinal relative speed refers to the speed difference between the preceding vehicle and the target vehicle along the direction of travel, reflecting the rate at which the two vehicles approach or move away; lateral distance refers to the distance between the preceding vehicle and the target vehicle perpendicular to the direction of travel; longitudinal distance refers to the distance between the preceding vehicle and the target vehicle in the direction of travel; adjacent vehicles refer to vehicles in the lane adjacent to the preceding vehicle that are in front of the preceding vehicle, i.e. Figure 3 The preceding vehicle in the adjacent lane; acceleration refers to the rate of change of vehicle speed over time, including longitudinal acceleration and lateral acceleration; false cut-out behavior identification result refers to the judgment result that when a vehicle in front shows a cutting-out tendency, it will be interrupted by obstacles or deceleration in front, thus causing its rear end to remain in the target vehicle's lane, based on the dynamic information of other vehicles; longitudinal acceleration refers to the acceleration of a vehicle along the direction of travel, reflecting the trend of vehicle acceleration or deceleration; emergency braking operation refers to the braking behavior taken quickly by the target vehicle with maximum braking capacity when a potential collision risk is detected.

[0035] In this embodiment, by comprehensively acquiring the lateral speed of the vehicle ahead, its longitudinal relative speed to the target vehicle, its lateral distance, its longitudinal distance, and the acceleration of vehicles in adjacent lanes, the system can promptly and accurately identify false cut-out attempts by the vehicle ahead. Therefore, upon confirming the risk of a false cut-out, by combining the longitudinal acceleration and lateral distance of the vehicle ahead, the system can quickly control the target vehicle to perform emergency braking, effectively avoiding rear-end collisions caused by false cut-outs and improving vehicle driving safety.

[0036] In this embodiment, step 101 includes: Step 1011: According to a preset sampling frequency, acquire in real time a sequence of images of the vehicles ahead using an in-vehicle image acquisition device preset in the target vehicle; wherein, the sequence of images of the vehicles ahead includes multiple consecutive images of the vehicles ahead. Step 1012: For any of the images of the vehicles ahead, obtain the coordinates of the rear center point of the vehicle ahead in the image of the vehicle ahead, so as to obtain the rear horizontal angle and rear pitch angle of the vehicle center point in the image of the vehicle ahead relative to the vehicle image acquisition device based on the rear image coordinates and the intrinsic parameter matrix of the vehicle image acquisition device. Step 1013: Based on the installation height of the vehicle-mounted image acquisition device and the rear tilt angle of the vehicle, obtain the longitudinal distance between the rear of the vehicle and the target vehicle in the image of the vehicle in front; Step 1014: Obtain the lateral distance between the rear of the vehicle and the target vehicle in the image of the vehicle in front based on the longitudinal distance of the rear of the vehicle and the horizontal angle; Step 1015: Obtain the difference in the lateral distance between the rear of the vehicle between any two adjacent images of the vehicle in front, so as to obtain the lateral velocity of the rear of the vehicle at each sampling time based on the difference in the lateral distance between the rear of the vehicle and the sampling frequency. Step 1016: Based on the preset window length, obtain multiple rear lateral velocities within each time window to obtain a rear lateral velocity sequence. Step 1017: When the difference between any two adjacent lateral velocities of the rear of the vehicle in the lateral velocity sequence is less than or equal to a preset difference threshold, it is determined that the vehicle in front has cut out.

[0037] In this embodiment, the system monitors in real time whether the target vehicle ahead maintains a stable lateral speed Vlat. The stability of this lateral speed is used to identify whether the vehicle ahead is preparing to cut out, and then to determine whether the cutting-out behavior is a false one, thereby improving vehicle driving safety. Preferably, to improve the accuracy of lateral speed recognition and thus the accuracy of cutting-out behavior recognition, an image sequence of the vehicle ahead can be acquired using an onboard image acquisition device pre-installed on the target vehicle. This image sequence is then used to identify the lateral speed of the vehicle ahead at each sampling moment. First, the longitudinal distance between the target vehicle and the vehicle ahead is calculated, where the longitudinal distance equals the camera height. The pitch angle tangent is calculated; similarly, the lateral distance between the center point of the rear of the vehicle in front and the target vehicle at each sampling time is calculated, where the lateral distance = longitudinal distance * horizontal angle tangent. The vehicle-mounted image acquisition device can be a monocular high-definition camera installed at the front of the target vehicle, used to capture and arrange a series of original image frames corresponding to the vehicle in front in chronological order at continuous time points. The rear center point represents the geometric or visual center of the rear of the vehicle in front. The vehicle in front can be identified by a deep learning-based target detection algorithm, and the bottom center point of its detection box or the geometric center of the entire detection box can be used as the rear center point. The rear image coordinates refer to the two-dimensional coordinates (u, v) of the rear center point on the image plane, expressed in pixels, and are usually directly output by target detection or feature localization algorithms. The intrinsic parameter matrix describes the internal geometric parameters of the vehicle-mounted image acquisition device, such as focal length, principal point coordinates, and distortion coefficients. The intrinsic parameter matrix refers to a matrix describing the internal geometric parameters of the vehicle-mounted image acquisition device, including focal length, principal point coordinates, distortion coefficients, etc. The rear horizontal angle and rear pitch angle represent the horizontal and vertical deflection angles of the rear center point relative to the optical axis of the vehicle-mounted image acquisition device, respectively. These angles can be calculated by inversely projecting the rear image coordinates onto the intrinsic parameter matrix of the vehicle-mounted image acquisition device, yielding the horizontal and vertical deflection angles of the rear center point relative to the camera's optical axis. The installation height refers to the vertical distance of the vehicle-mounted image acquisition device relative to the ground. This installation height can be determined during vehicle design and installation and preset as a fixed value in the system. After obtaining the preset camera parameters, the rear longitudinal distance can be calculated using the tangent relationship between the installation height and the rear pitch angle through triangulation. The rear lateral distance refers to the distance perpendicular to the vehicle's direction of travel between the center point of the rear of the preceding vehicle and the target vehicle's vehicle-mounted image acquisition device. This lateral distance can also be calculated using the tangent relationship between the rear longitudinal distance and the rear horizontal angle through triangulation.

[0038] When calculating lateral velocity, the difference in lateral distance between the rear of the vehicle refers to the change in the lateral distance between the rear of the vehicle in front between two consecutive sampling times. The lateral velocity can then be calculated by dividing the difference in lateral distance by the sampling period (the reciprocal of the sampling frequency). A fixed time length, such as 1 second, 2 seconds, or a specific number of sampling points, is then set to ensure sufficient data to assess the trend and stability of the lateral velocity. Finally, using the time interval (Δt) between two frames and the difference in lateral and longitudinal distances (Δd), the lateral velocity and longitudinal velocity are calculated separately using the formula: v = Δd / Δt.

[0039] In this embodiment, an onboard image acquisition device acquires a sequence of images of the vehicle ahead in real time at a preset sampling frequency, ensuring the continuity and timeliness of data acquisition. Then, by accurately calculating the coordinates, horizontal angle, and pitch angle of the vehicle's rear center point in the image, and combining this with the intrinsic parameter matrix and installation height of the onboard image acquisition device, the longitudinal and lateral distances between the vehicle's rear and the target vehicle can be accurately calculated. Based on this, by analyzing the changes in the lateral distance between adjacent images and combining this with the sampling frequency, the lateral velocity of the vehicle ahead is calculated in real time. Furthermore, by constructing a lateral velocity sequence within a preset window length and determining whether the difference between any two adjacent velocities in the sequence is less than or equal to a preset difference threshold, the presence of a cutting-out behavior by the vehicle ahead can be effectively identified. This method, based on image sequences and geometric calculations, avoids the recognition delays and misjudgments caused by the lack of real-time and accurate lateral velocity information in traditional methods, thus providing timely and reliable input for subsequent false cutting-out behavior identification and emergency braking operations, significantly improving vehicle driving safety.

[0040] In this embodiment, step 102 includes: Step 1021: When it is determined that the vehicle in front is cutting out, locate the target cutting lane of the vehicle in front, and take the rear end point of the side of the vehicle in front that is far away from the target cutting lane as the near end point of the vehicle in front. Step 1022: For any of the images of the vehicle ahead, obtain the target bounding box of the vehicle ahead in the image, and determine the endpoint image coordinates of the near end point based on the target bounding box and the image coordinates of the rear of the vehicle. Step 1023: Based on the endpoint image coordinates and the intrinsic parameter matrix, obtain the horizontal angle and pitch angle of the near endpoint relative to the vehicle image acquisition device in the image of the vehicle in front; Step 1024: Based on the installation height of the vehicle-mounted image acquisition device and the pitch angle, obtain the near longitudinal distance between the near endpoint in the image of the vehicle ahead and the target vehicle; Step 1025: Obtain the proximal lateral distance between the proximal endpoint and the target vehicle based on the proximal longitudinal distance and the horizontal angle; Step 1026: Obtain the near-side longitudinal distance difference between any two adjacent images of the vehicle ahead, so as to obtain the longitudinal relative speed between the vehicle ahead and the target vehicle at each sampling time based on the near-side longitudinal distance difference and the sampling frequency.

[0041] In this embodiment, the lateral distance Plat between the near end point of the vehicle in front and the center point of the front end of the target vehicle is monitored in real time; the longitudinal distance Plgt between the center point of the rear end of the vehicle in front and the center point of the front end of the target vehicle is monitored in real time; and the longitudinal relative speed between the target vehicle and the vehicle in front is monitored in real time. The near end point is the point where the near end of the vehicle in front is closest to the front end of the target vehicle and is most likely to collide / has a potential collision risk. Specifically, as shown... Figure 4 As shown, when the preceding vehicle is about to cut out, the endpoint of the preceding vehicle on the side furthest from the adjacent lane, closest to the rear of the target vehicle, is considered. When acquiring the longitudinal relative speed, lateral distance, and longitudinal distance, for any image of the preceding vehicle, a target bounding box is obtained within that image. This target bounding box provides the approximate position and size information of the preceding vehicle in the image. Subsequently, the endpoint image coordinates of the near-side endpoint are determined based on this target bounding box and the pre-acquired rear-end image coordinates. After obtaining the endpoint image coordinates of the near-side endpoint, the horizontal and pitch angles of the near-side endpoint relative to the vehicle image acquisition device in the preceding vehicle image are calculated using these coordinates and the intrinsic parameter matrix of the vehicle image acquisition device. Finally, based on the installation height of the vehicle image acquisition device and the calculated pitch angle, the near-side longitudinal distance between the near-side endpoint and the target vehicle in the preceding vehicle image is determined. After obtaining the near-side longitudinal distance, the near-side lateral distance between the near-side endpoint and the target vehicle is further determined based on this near-side longitudinal distance and the previously calculated horizontal angle. Similarly, using trigonometric relationships, the near-side lateral distance can be obtained by multiplying the near-side longitudinal distance by tan(horizontal angle). This allows the system to obtain the precise lateral position of the vehicle in front as it cuts out, with the side closest to the target vehicle in real time.

[0042] To obtain the longitudinal relative speed between the vehicle ahead and the target vehicle, the near-side longitudinal distance difference between any two adjacent images of the vehicle ahead is obtained. Based on this near-side longitudinal distance difference and a preset sampling frequency, the longitudinal relative speed between the vehicle ahead and the target vehicle at each sampling time is calculated. The sampling frequency is the rate at which the image acquisition device acquires images, for example, 30 frames per second. The longitudinal relative speed is obtained by dividing the distance difference by the sampling time interval (i.e., the reciprocal of the sampling frequency).

[0043] In this embodiment, by locating the target lane that the vehicle ahead is cutting out of and selecting the rear endpoint furthest from that lane as the near endpoint, the acquired point location is ensured to most accurately reflect the potential risk posed by the vehicle to the target vehicle during the cut-out process. The image coordinates of the near endpoint are determined using the target bounding box and the rear image coordinates, improving the accuracy of endpoint positioning and avoiding potential deviations caused by solely relying on the rear center point. The horizontal and pitch angles are calculated based on the endpoint image coordinates and the intrinsic parameter matrix, transforming two-dimensional image information into angular information in three-dimensional space, providing a reliable foundation for subsequent distance calculations. The near longitudinal distance is calculated by combining the installation height and pitch angle of the vehicle-mounted image acquisition device, and the near lateral distance is calculated using the near longitudinal distance and horizontal angle, achieving real-time perception of the precise longitudinal and lateral positional relationship between the vehicle ahead and the target vehicle. Finally, the longitudinal relative vehicle speed is calculated by combining the difference in near longitudinal distance between adjacent images with the sampling frequency, ensuring continuous real-time updates of speed information. These precise and real-time parameter acquisition methods significantly improve the timeliness and accuracy of false cut-out behavior identification, providing reliable data support for subsequent emergency braking decisions, thereby effectively reducing the risk of rear-end collisions and enhancing vehicle driving safety.

[0044] In this embodiment, step 103 includes: Step 1031: Obtain the adjacent vehicles in front of the vehicle in the target lane, and obtain the coordinates of the center point of the rear of the adjacent vehicles at each sampling time according to the preset sampling frequency; Step 1032: Based on the difference between the coordinates between two adjacent sampling times, obtain the acceleration of the adjacent vehicle at each sampling time according to the difference and the sampling frequency.

[0045] In this embodiment, an in-vehicle image acquisition device can be used to acquire image sequences of adjacent vehicles. Target detection and tracking algorithms are then used to identify the adjacent vehicles and estimate the pixel coordinates of their rear center points in the images. These pixel coordinates are then converted into three-dimensional coordinates in either the vehicle coordinate system or the world coordinate system by combining the intrinsic and extrinsic parameter matrices of the in-vehicle image acquisition device. Based on this, the acceleration of the adjacent vehicles at each sampling time is obtained according to the difference between two adjacent sampling time coordinates and the sampling frequency.

[0046] In this embodiment, by accurately identifying adjacent vehicles in front of the vehicle ahead in the target lane, and continuously acquiring the coordinates of their rear center points at a preset sampling frequency, high-precision, real-time raw data is provided for subsequent acceleration calculations. Based on this, by calculating the coordinate differences between adjacent sampling times and combining this with the sampling frequency, the acceleration of the adjacent vehicle at each sampling time can be quickly and effectively derived. This real-time acceleration information acquisition can promptly capture the deceleration or emergency braking behavior of adjacent vehicles, thereby significantly shortening the risk assessment and response time in false lane exit scenarios. Compared to existing technologies that rely solely on longitudinal and lateral distances for collision risk assessment, introducing real-time acceleration information from adjacent vehicles allows for earlier prediction of potential hazards, providing valuable time for the target vehicle to perform emergency braking, effectively preventing rear-end collisions and greatly improving vehicle driving safety.

[0047] In this embodiment, step 104 includes: Step 1041: For any sampling time, obtain the first comparison result between the longitudinal distance of the rear of the vehicle and the preset longitudinal distance threshold; Step 1042: Obtain the second comparison result between the acceleration of adjacent vehicles and the preset acceleration threshold; Step 1043: Obtain the third comparison result between the proximal lateral distance and the preset lateral distance threshold; Step 1044: Obtain the fourth comparison result between the longitudinal relative vehicle speed and the preset vehicle speed threshold; Step 1045: Based on the first comparison result, the second comparison result, the third comparison result, and the fourth comparison result, obtain the false cut-out behavior identification result of the vehicle in front at the sampling time.

[0048] In this embodiment, obtaining the false cut-out behavior identification result of the vehicle ahead at the sampling time aims to comprehensively consider information from multiple dimensions to make a more comprehensive and accurate judgment on the cut-out behavior of the vehicle ahead, avoiding misjudgments that may be caused by a single parameter. Specifically, logical judgment rules can be adopted. For example, when the first comparison result indicates a risk, the third comparison result indicates that the cut-out is not complete, and the second comparison result indicates a sharp deceleration, it is determined that there is a false cut-out behavior. Preferably, a threshold is set for logical judgment, such as determining that the acceleration Aff in the adjacent lane of the adjacent lane is ≤ -1 m / s², and if Plat < 1 / 2 * (the width of the target vehicle) and the longitudinal distance Plgt < 4.x meters, and the longitudinal relative speed is ≥ 0, it is determined that the vehicle ahead has a false cut-out behavior.

[0049] This embodiment evaluates at any sampling moment, ensuring the continuity and real-time nature of the identification and avoiding delays in risk assessment due to time constraints. A first comparison between the longitudinal distance to the rear of the vehicle and a longitudinal distance threshold determines whether the longitudinal position exceeds a safe range, aiding in assessing potential collision risk, as longitudinal distance directly reflects the proximity between vehicles. A second comparison between the acceleration of adjacent vehicles and an acceleration threshold detects changes in the behavior of vehicles in adjacent lanes, providing a warning signal, as changes in acceleration may indicate that the vehicle ahead is about to decelerate or brake. A third comparison between the near-side lateral distance and a lateral distance threshold assesses whether the vehicle's lateral position has deviated from its lane, crucial for identifying whether the lane change is complete, as changes in lateral distance indicate whether the vehicle has partially or completely left the lane. A fourth comparison between the longitudinal relative speed and a speed threshold determines whether the relative speed is too high, affecting the urgency level, as relative speed directly determines the collision time window. Finally, the false cut-out behavior identification result is obtained by combining all comparison results, avoiding misjudgment or omission that may be caused by single parameter evaluation. Multi-dimensional threshold comparison ensures more comprehensive and reliable identification, thereby triggering braking operation in time in false cut-out scenarios and reducing the risk of rear-end collision accidents.

[0050] In this embodiment, step 106 includes: Step 1061: When it is determined that the vehicle in front has made a false cut-out, obtain the longitudinal acceleration of the vehicle in front; Step 1062: When the longitudinal acceleration is less than or equal to a preset acceleration threshold, control the target vehicle to perform an emergency braking operation according to the third comparison result.

[0051] In this embodiment, the longitudinal acceleration Af of the vehicle ahead is continuously monitored. When the longitudinal acceleration Af of the vehicle ahead is ≤ -0.2, the conditions for triggering AEB (Automatic Emergency Braking) are met. If Plat < 1 / 2 * (vehicle width), the vehicle immediately requests the maximum master cylinder pressure for emergency braking. Specifically, when it is determined that the vehicle ahead is engaging in a false cut-out behavior, the longitudinal acceleration of the vehicle ahead is acquired to obtain the dynamic deceleration information of the vehicle ahead in real time. This longitudinal acceleration can be acquired in various ways. For example, the target vehicle can use its onboard radar sensors, lidar sensors, or vision sensors to monitor the motion state of the vehicle ahead in real time. By performing differential calculations on the longitudinal velocity of the vehicle ahead at continuous time points, its longitudinal acceleration can be obtained. Further, when the longitudinal acceleration is less than or equal to a preset acceleration threshold, the target vehicle is controlled to perform an emergency braking operation based on the third comparison result. This step combines the dynamic deceleration behavior of the vehicle ahead and its lateral distance from the target vehicle to make a precise braking decision. The target vehicle's controller continuously monitors the longitudinal acceleration of the vehicle ahead. When this acceleration value is lower than or equal to a preset negative value (e.g., -3 m / s², indicating significant deceleration), the system further checks the third comparison result mentioned above. This third comparison result refers to the comparison between the near-side lateral distance and a preset lateral distance threshold. If both conditions are met—that is, the vehicle ahead is significantly decelerating and its side distance to the target vehicle is too close—emergency braking is triggered.

[0052] This embodiment, upon detecting a false cut-out by a vehicle ahead, further acquires its longitudinal acceleration and combines this acceleration with the lateral distance to make a braking decision. This method directly quantifies the dynamic deceleration behavior of the vehicle ahead, avoiding braking lag that may result from relying solely on static distance parameters. When the longitudinal acceleration of the vehicle ahead is less than or equal to a preset acceleration threshold, it indicates that it is undergoing significant deceleration. At this point, combining the third comparison result (i.e., the comparison result of lateral distance) to determine whether an emergency braking operation is necessary effectively avoids accidental braking when the lateral distance is safe, while ensuring timely response when danger approaches. This significantly improves the accuracy and timeliness of braking decisions, thereby effectively reducing the risk of rear-end collisions in false cut-out scenarios and ensuring vehicle driving safety.

[0053] On the other hand, refer to Figure 2 This embodiment also discloses a vehicle braking system based on false cut-out recognition, including a cut-out recognition module 201, a forward vehicle recognition module 202, an adjacent vehicle recognition module 203, a false cut-out recognition module 204, and a braking module 205.

[0054] The cut-out recognition module 201 is used to obtain the lateral speed of the vehicle in front in the lane where the target vehicle is located, so as to obtain the cut-out behavior recognition result of the vehicle in front based on the lateral speed.

[0055] The forward vehicle recognition module 202 is used to obtain the longitudinal relative speed, lateral distance and longitudinal distance between the forward vehicle and the target vehicle in real time based on the cut-out behavior recognition result.

[0056] The adjacent vehicle recognition module 203 is used to obtain the acceleration of the adjacent vehicle in front of the vehicle in front in the adjacent lane based on the cut-out behavior recognition result.

[0057] The false cutout recognition module 204 is used to obtain the false cutout behavior recognition result of the vehicle in front based on the longitudinal relative vehicle speed, the lateral distance, the longitudinal distance and the acceleration.

[0058] The braking module 205 is used to obtain the longitudinal acceleration of the vehicle in front based on the false cut-out behavior recognition result, so as to control the target vehicle to perform an emergency braking operation based on the longitudinal acceleration and the lateral distance.

[0059] In this embodiment, the cutout recognition module 201 includes an image sequence unit, an angle recognition unit, a distance calculation unit, and a behavior recognition unit.

[0060] The image sequence unit is used to acquire a sequence of images of the vehicles ahead in real time using an in-vehicle image acquisition device pre-installed in the target vehicle, according to a preset sampling frequency; wherein the sequence of images of the vehicles ahead includes multiple consecutive images of the vehicles ahead.

[0061] The angle recognition unit is used to obtain the rear image coordinates of the center point of the rear of the vehicle in any given image of the vehicle in front, so as to obtain the rear horizontal angle and rear pitch angle of the center point of the vehicle in the image of the vehicle in front relative to the vehicle image acquisition device based on the rear image coordinates and the intrinsic parameter matrix of the vehicle image acquisition device.

[0062] The distance calculation unit is used to obtain the longitudinal distance between the front vehicle and the target vehicle in the front vehicle image based on the installation height of the vehicle-mounted image acquisition device and the rear vehicle pitch angle; and to obtain the lateral distance between the front vehicle and the target vehicle in the front vehicle image based on the longitudinal distance and the horizontal angle.

[0063] The behavior recognition unit is used to obtain the difference in the lateral distance between the rear of any two adjacent images of the vehicle in front, so as to obtain the lateral speed of the rear of the vehicle at each sampling time according to the difference in the lateral distance between the rear of the vehicle and the sampling frequency; according to a preset window length, it obtains multiple lateral speeds of the rear of the vehicle within each time window to obtain a sequence of lateral speeds of the rear of the vehicle; when the difference between any two adjacent lateral speeds of the rear of the vehicle in the sequence of lateral speeds of the rear of the vehicle is less than or equal to a preset difference threshold, it is determined that the vehicle in front has cut out behavior.

[0064] In this embodiment, the forward vehicle recognition module 202 includes an endpoint positioning unit and an operating parameter recognition unit.

[0065] The endpoint positioning unit is used to locate the target lane of the vehicle ahead when it is determined that the vehicle ahead is cutting out, so as to take the rear endpoint of the vehicle ahead that is far away from the target lane as the near endpoint of the vehicle ahead; for any image of the vehicle ahead, the target box of the vehicle ahead in the image is obtained, so as to determine the endpoint image coordinates of the near endpoint based on the target box and the rear image coordinates.

[0066] The operating parameter identification unit is used to obtain the horizontal angle and pitch angle of the near endpoint relative to the vehicle-mounted image acquisition device in the image of the vehicle ahead based on the endpoint image coordinates and the intrinsic parameter matrix; to obtain the near longitudinal distance between the near endpoint and the target vehicle in the image of the vehicle ahead based on the installation height of the vehicle-mounted image acquisition device and the pitch angle; to obtain the near lateral distance between the near endpoint and the target vehicle based on the near longitudinal distance and the horizontal angle; and to obtain the near longitudinal distance difference between any two adjacent images of the vehicle ahead, so as to obtain the longitudinal relative speed between the vehicle ahead and the target vehicle at each sampling time based on the near longitudinal distance difference and the sampling frequency.

[0067] In this embodiment, the adjacent vehicle identification module 203 includes a vehicle positioning unit and an acceleration identification unit.

[0068] The vehicle positioning unit is used to acquire adjacent vehicles located in front of the vehicle ahead in the target lane, and to acquire the coordinates of the center point of the rear of the adjacent vehicles at each sampling time according to a preset sampling frequency.

[0069] The acceleration identification unit is used to obtain the acceleration of the adjacent vehicle at each sampling time based on the difference between the coordinates between two adjacent sampling times and the sampling frequency.

[0070] This embodiment discloses a vehicle braking method and system based on false cut-out recognition. In addition to monitoring only the driving behavior of the vehicle in front, it also monitors the driving behavior of the vehicle in the adjacent lane in real time. By combining the driving behavior of the vehicle in the adjacent lane, it determines whether the vehicle in front will cut out (false cut-out). When the system determines that the vehicle in the adjacent lane is braking and the braking of the vehicle in front meets certain conditions, and the cut-out is not completed (the lateral distance is sufficient to meet the cut-out conditions), the vehicle immediately performs emergency braking in advance to avoid rear-end collision with the vehicle in front whose cut-out has been interrupted, thereby improving the safety of vehicle driving.

[0071] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A vehicle braking method based on false cut-out recognition, characterized in that, include: The lateral velocity of the vehicle in front in the lane where the target vehicle is located is obtained, so as to obtain the cut-out behavior recognition result of the vehicle in front based on the lateral velocity; Based on the cut-out behavior recognition results, the longitudinal relative speed, lateral distance and longitudinal distance between the vehicle in front and the target vehicle are obtained in real time. Based on the cut-out behavior recognition result, obtain the acceleration of the adjacent vehicle in the adjacent lane that is in front of the vehicle in front; The false cut-out behavior identification result of the vehicle in front is obtained based on the longitudinal relative vehicle speed, the lateral distance, the longitudinal distance and the acceleration; Based on the false cut-out behavior recognition result, the longitudinal acceleration of the vehicle in front is obtained, and the target vehicle is controlled to perform an emergency braking operation based on the longitudinal acceleration and the lateral distance.

2. The vehicle braking method based on false cut-out recognition according to claim 1, characterized in that, The step of obtaining the lateral velocity of the vehicle in front in the lane where the target vehicle is located, and obtaining the cut-out behavior recognition result of the vehicle in front based on the lateral velocity, includes: According to a preset sampling frequency, a sequence of images of the vehicles ahead is acquired in real time by an in-vehicle image acquisition device preset in the target vehicle; wherein, the sequence of images of the vehicles ahead includes multiple consecutive images of the vehicles ahead. For any of the images of the vehicles ahead, obtain the coordinates of the rear center point of the vehicle ahead in the image of the vehicle ahead, so as to obtain the rear horizontal angle and rear pitch angle of the vehicle center point in the image of the vehicle ahead relative to the vehicle image acquisition device based on the rear image coordinates and the intrinsic parameter matrix of the vehicle image acquisition device. The longitudinal distance between the rear of the vehicle and the target vehicle in the image of the vehicle ahead is obtained based on the installation height of the vehicle-mounted image acquisition device and the rear pitch angle of the vehicle. The lateral distance between the rear of the vehicle and the target vehicle in the image of the vehicle in front is obtained based on the longitudinal distance of the rear of the vehicle and the horizontal angle. The difference in the lateral distance between the rear of the vehicle is obtained between any two adjacent images of the vehicle in front, so as to obtain the lateral velocity of the rear of the vehicle at each sampling time based on the difference in the lateral distance between the rear of the vehicle and the sampling frequency. Based on the preset window length, multiple lateral velocities of the vehicle rear within each time window are obtained to obtain a sequence of lateral velocities of the vehicle rear. When the difference between any two adjacent lateral velocities of the rear of the vehicle in the lateral velocity sequence is less than or equal to a preset difference threshold, it is determined that the vehicle in front has cut out.

3. The vehicle braking method based on false cut-out recognition according to claim 2, characterized in that, The step of acquiring the longitudinal relative speed, lateral distance, and longitudinal distance between the vehicle in front and the target vehicle in real time based on the cut-out behavior recognition result includes: When it is determined that the vehicle in front is cutting out, the target cutting lane of the vehicle in front is located, and the rear end point of the side of the vehicle in front that is far away from the target cutting lane is taken as the proximal end point of the vehicle in front. For any of the images of the vehicle ahead, obtain the target bounding box of the vehicle ahead in the image, and determine the endpoint image coordinates of the near end point based on the target bounding box and the image coordinates of the rear of the vehicle. The horizontal and vertical angles of the near endpoint relative to the vehicle image acquisition device in the image of the vehicle ahead are obtained based on the endpoint image coordinates and the intrinsic parameter matrix. The near longitudinal distance between the near endpoint of the forward vehicle image and the target vehicle is obtained based on the installation height of the vehicle-mounted image acquisition device and the pitch angle. The proximal lateral distance between the proximal endpoint and the target vehicle is obtained based on the proximal longitudinal distance and the horizontal angle. Obtain the near-side longitudinal distance difference between any two adjacent images of the vehicle ahead, and obtain the longitudinal relative speed between the vehicle ahead and the target vehicle at each sampling time based on the near-side longitudinal distance difference and the sampling frequency.

4. The vehicle braking method based on false cut-out recognition according to claim 3, characterized in that, The step of obtaining the acceleration of the adjacent vehicle in front of the vehicle in front in the adjacent lane based on the cut-out behavior recognition result includes: The system acquires the adjacent vehicles in front of the vehicle in the target lane and obtains the coordinates of the center point of the rear of the adjacent vehicles at each sampling time according to a preset sampling frequency. The acceleration of the adjacent vehicle at each sampling moment is obtained based on the difference between the coordinates between two adjacent sampling times and the sampling frequency.

5. A vehicle braking method based on false cut-out recognition according to any one of claims 2-4, characterized in that, The step of obtaining the false cut-out behavior recognition result of the vehicle in front based on the longitudinal relative vehicle speed, the lateral distance, the longitudinal distance, and the acceleration includes: For any sampling time, obtain the first comparison result between the longitudinal distance of the rear of the vehicle and the preset longitudinal distance threshold; Obtain a second comparison result between the acceleration of adjacent vehicles and a preset acceleration threshold; Obtain the third comparison result between the proximal lateral distance and the preset lateral distance threshold; Obtain the fourth comparison result between the longitudinal relative vehicle speed and the preset vehicle speed threshold; Based on the first comparison result, the second comparison result, the third comparison result, and the fourth comparison result, the false cut-out behavior identification result of the vehicle in front at the sampling time is obtained.

6. A vehicle braking method based on false cut-out recognition according to claim 5, characterized in that, The step of obtaining the longitudinal acceleration of the vehicle ahead based on the false cut-out behavior recognition result, and controlling the vehicle to perform an emergency braking operation based on the longitudinal acceleration and the lateral distance, includes: When it is determined that the vehicle in front has made a false cut-out, the longitudinal acceleration of the vehicle in front is obtained; When the longitudinal acceleration is less than or equal to a preset acceleration threshold, the target vehicle is controlled to perform an emergency braking operation based on the third comparison result.

7. A vehicle braking system based on false cut-out recognition, characterized in that, It includes a cut-out recognition module, a forward vehicle recognition module, an adjacent vehicle recognition module, a false cut-out recognition module, and a braking module; The cut-out recognition module is used to obtain the lateral speed of the vehicle in front in the lane where the target vehicle is located, so as to obtain the cut-out behavior recognition result of the vehicle in front based on the lateral speed. The forward vehicle recognition module is used to obtain, in real time, the longitudinal relative speed, lateral distance and longitudinal distance between the forward vehicle and the target vehicle based on the cut-out behavior recognition result; The adjacent vehicle recognition module is used to obtain the acceleration of the adjacent vehicle in front of the vehicle in front in the adjacent lane based on the cut-out behavior recognition result; The false cutout recognition module is used to obtain the false cutout behavior recognition result of the vehicle in front based on the longitudinal relative vehicle speed, the lateral distance, the longitudinal distance and the acceleration; The braking module is used to obtain the longitudinal acceleration of the vehicle in front based on the false cut-out behavior recognition result, so as to control the target vehicle to perform an emergency braking operation based on the longitudinal acceleration and the lateral distance.

8. A vehicle braking system based on false cut-out recognition according to claim 7, characterized in that, The cutout recognition module includes an image sequence unit, an angle recognition unit, a distance calculation unit, and a behavior recognition unit; The image sequence unit is used to acquire a sequence of images of the vehicles ahead in real time using an in-vehicle image acquisition device pre-installed in the target vehicle, according to a preset sampling frequency; wherein the sequence of images of the vehicles ahead includes multiple consecutive images of the vehicles ahead. The angle recognition unit is used to obtain the rear image coordinates of the rear center point of the vehicle in any image of the vehicle in front, so as to obtain the rear horizontal angle and rear pitch angle of the rear center point in the image of the vehicle in front relative to the vehicle image acquisition device based on the rear image coordinates and the intrinsic parameter matrix of the vehicle image acquisition device. The distance calculation unit is used to obtain the longitudinal distance between the front vehicle and the target vehicle in the front vehicle image based on the installation height of the vehicle-mounted image acquisition device and the rear vehicle pitch angle; and to obtain the lateral distance between the front vehicle and the target vehicle in the front vehicle image based on the longitudinal distance and the horizontal angle. The behavior recognition unit is used to obtain the difference in the lateral distance between the rear of any two adjacent images of the vehicle in front, so as to obtain the lateral speed of the rear of the vehicle at each sampling time according to the difference in the lateral distance between the rear of the vehicle and the sampling frequency; according to a preset window length, it obtains multiple lateral speeds of the rear of the vehicle within each time window to obtain a sequence of lateral speeds of the rear of the vehicle; when the difference between any two adjacent lateral speeds of the rear of the vehicle in the sequence of lateral speeds of the rear of the vehicle is less than or equal to a preset difference threshold, it is determined that the vehicle in front has cut out behavior.

9. A vehicle braking system based on false cut-out recognition according to claim 8, characterized in that, The forward vehicle recognition module includes an endpoint positioning unit and an operating parameter recognition unit; The endpoint positioning unit is used to locate the target lane of the vehicle ahead when it is determined that the vehicle ahead is cutting out, so as to take the rear endpoint of the vehicle ahead on the side away from the target lane as the near endpoint of the vehicle ahead; for any image of the vehicle ahead, the target box of the vehicle ahead in the image is obtained, so as to determine the endpoint image coordinates of the near endpoint based on the target box and the rear image coordinates. The operating parameter identification unit is used to obtain the horizontal angle and pitch angle of the near endpoint relative to the vehicle-mounted image acquisition device in the image of the vehicle ahead based on the endpoint image coordinates and the intrinsic parameter matrix; to obtain the near longitudinal distance between the near endpoint and the target vehicle in the image of the vehicle ahead based on the installation height of the vehicle-mounted image acquisition device and the pitch angle; to obtain the near lateral distance between the near endpoint and the target vehicle based on the near longitudinal distance and the horizontal angle; and to obtain the near longitudinal distance difference between any two adjacent images of the vehicle ahead, so as to obtain the longitudinal relative speed between the vehicle ahead and the target vehicle at each sampling time based on the near longitudinal distance difference and the sampling frequency.

10. A vehicle braking system based on false cut-out recognition according to claim 9, characterized in that, The adjacent vehicle identification module includes a vehicle positioning unit and an acceleration identification unit; The vehicle positioning unit is used to obtain the adjacent vehicles located in front of the vehicle in the target lane, and to obtain the coordinates of the center point of the rear of the adjacent vehicles at each sampling time according to a preset sampling frequency. The acceleration identification unit is used to obtain the acceleration of the adjacent vehicle at each sampling time based on the difference between the coordinates between two adjacent sampling times and the sampling frequency.

Images