Vehicle target detection method and device and vehicle

Abstract

The embodiment of the invention discloses a vehicle target detection method and device and a vehicle, relates to the technical field of automatic driving, and aims to reduce mistaken deceleration of the vehicle. The detection method comprises the following steps: acquiring a first target detected by a first sensor, and a second target and a third target detected by a second sensor; if the first target and the second target conform to the fusion condition and the third target does not conform to the fusion condition, fusing the first target and the second target into a main target, and determining a splitting suppression region in front of the main target; and if the third target is in the split suppression region and the longitudinal speed difference between the main target and the third target is in the speed difference threshold range, suppressing the output of the third target. The method is suitable for automatic driving and other scenes.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and in particular to a vehicle target detection method, device and vehicle. Background Art

[0002] In the field of autonomous driving, correct identification of targets is crucial. However, due to sensor limitations and the influence of the external environment, misidentification often occurs.

[0003] There are often large vehicles on the road that are more than 10 meters or even more than 20 meters long, such as large trucks or large buses. When ordinary millimeter-wave radar detects such vehicles, the number of point clouds is sparse and incoherent, which easily causes a large vehicle to be identified as two or even more targets that are separated from each other. This causes the large target to split. When the target split from the large target (also called a false target) invades the vehicle's lane, it is easy to cause the vehicle to decelerate by mistake. Summary of the invention

[0004] In view of this, an embodiment of the present invention provides a vehicle target detection method, device and vehicle, which are helpful to reduce the occurrence of erroneous deceleration of the vehicle.

[0005] In a first aspect, an embodiment of the present invention provides a vehicle target detection method, which detects a neighboring vehicle target traveling in the same direction in an adjacent lane in front of a vehicle by using a first sensor and a second sensor; wherein the first sensor and the second sensor are sensors on the vehicle;

[0006] Acquire the first target detected by the first sensor, the second target and the third target detected by the second sensor; determine whether the first target, the second target and the third target meet the fusion condition; if the first target and the second target meet the fusion condition, and the third target does not meet the fusion condition, fuse the first target and the second target into a main target, and determine a split inhibition zone in front of the main target; determine whether the third target is within the split inhibition zone, and whether the longitudinal speed difference between the main target and the third target is within the speed difference threshold range; if the third target is within the split inhibition zone, and the longitudinal speed difference between the main target and the third target is within the speed difference threshold range, suppress the output of the third target.

[0007] In one feasible implementation, the first sensor is a visual sensor, and the second sensor is a radar; wherein, detecting the neighboring vehicle target traveling in the same direction in the adjacent lane in front of the vehicle by using the first sensor and the second sensor includes: collecting the image of the adjacent lane in front of the vehicle by using the visual sensor, and determining the neighboring vehicle target traveling in the same direction in the adjacent lane in front of the vehicle based on image recognition; and performing target detection on the adjacent lane in front of the vehicle by using the radar, and determining the neighboring vehicle target traveling in the same direction in the adjacent lane in front of the vehicle.

[0008] In an achievable implementation, the determining whether the first target, the second target and the third target meet the fusion conditions includes: calculating the association between the second target and the first target based on the Mahalanobis distance between the second target and the first target, and if the association between the second target and the first target is greater than a preset threshold, determining that the second target and the first target meet the fusion conditions; determining the association between the third target and the first target or the second target based on the Mahalanobis distance between the third target and the first target or the second target, and if the association between the third target and the first target or the second target is greater than a preset threshold, determining that the third target meets the fusion conditions with the first target or the second target.

[0009] In an achievable implementation, the calculating the association degree between the second target and the first target based on the Mahalanobis distance between the second target and the first target includes: calculating the association degree F between the second target and the first target according to the following formula:

[0010] ;

[0011] in, ;

[0012] D M1 is the Mahalanobis distance calculated based on the longitudinal distances of the neighboring vehicle relative to the vehicle measured by the visual sensor and the radar, respectively, D M2 is the Mahalanobis distance calculated based on the lateral distance of the neighboring vehicle relative to the vehicle measured by the visual sensor and the radar, respectively, D M3 is the Mahalanobis distance calculated based on the longitudinal speed of the neighboring vehicle relative to the vehicle measured by the visual sensor and the radar, respectively, D M4 is the Mahalanobis distance calculated according to the lateral speed of the neighboring vehicle relative to the own vehicle measured by the visual sensor and the radar respectively, and σ is an empirical coefficient.

[0013] In a feasible implementation, the determining of the split inhibition zone in front of the main target includes: determining a current longitudinal distance between the main target and the own vehicle, and a current longitudinal absolute speed of the main target; determining the length of the split inhibition zone in front of the main target based on the current longitudinal distance between the main target and the own vehicle, the current longitudinal absolute speed of the main target, and a pre-set mapping relationship between the longitudinal distance between the main target and the own vehicle, the longitudinal absolute speed of the main target, and the length of the split inhibition zone.

[0014] In an achievable implementation, the determining the split inhibition zone in front of the main target includes: determining a width of the split inhibition zone in front of the main target based on a width of the main target.

[0015] In a feasible implementation, before determining the split inhibition zone in front of the main target, the method also includes: determining the current traffic scene, and determining whether to activate the split inhibition zone in front of the main target based on the current traffic scene; wherein the current traffic scene includes the current road scene, the current longitudinal speed of the vehicle, the vehicle type of the main target, the longitudinal absolute speed of the third target, and at least one of the vehicle type of the third target.

[0016] In one feasible implementation, the step of determining the current traffic scene and determining whether to activate the split inhibition zone in front of the main target based on the current traffic scene includes: if the current road scene is a highway, the current longitudinal speed of the vehicle is above 60kph, the vehicle type of the main target is a large vehicle, the longitudinal absolute speed of the third target is greater than 5m / s, and the vehicle type of the third target is a small vehicle, then the step of determining the split inhibition zone in front of the main target is initiated.

[0017] In a second aspect, an embodiment of the present invention further provides a vehicle target detection device for detecting a neighboring vehicle target traveling in the same direction in an adjacent lane in front of a vehicle, the device comprising:

[0018] A target acquisition module, used to acquire a first target detected by a first sensor, a second target and a third target detected by a second sensor; wherein the first sensor and the second sensor are sensors on the vehicle;

[0019] A target fusion module is used to determine whether the first target, the second target and the third target meet the fusion conditions; if the first target and the second target meet the fusion conditions, and the third target does not meet the fusion conditions, the first target and the second target are fused into a main target, and a split inhibition zone in front of the main target is determined; determine whether the third target is in the split inhibition zone, and whether the longitudinal speed difference between the main target and the third target is within the speed difference threshold range; if the third target is in the split inhibition zone, and the longitudinal speed difference between the main target and the third target is within the speed difference threshold range, suppress the output of the third target.

[0020] In a feasible implementation, the first sensor is a visual sensor, and the second sensor is a radar; the visual sensor is used to collect images of the adjacent lane in front of the vehicle, and the neighboring vehicle target traveling in the same direction in the adjacent lane in front of the vehicle is determined based on image recognition; the radar is used to perform target detection on the adjacent lane in front of the vehicle, and the neighboring vehicle target traveling in the same direction in the adjacent lane in front of the vehicle is determined.

[0021] In an achievable implementation, the target fusion module includes: a fusion condition judgment submodule, used to judge whether the first target, the second target and the third target meet the fusion condition; a target fusion submodule, used to fuse the first target and the second target into a main target if the first target and the second target meet the fusion condition and the third target does not meet the fusion condition; an inhibition zone determination submodule, used to determine the split inhibition zone in front of the main target; a target output inhibition submodule, used to judge whether the third target is in the split inhibition zone, and whether the longitudinal speed difference between the main target and the third target is within the speed difference threshold range; if the third target is in the split inhibition zone and the longitudinal speed difference between the main target and the third target is within the speed difference threshold range, the output of the third target is suppressed.

[0022] In an achievable implementation, the fusion condition judgment submodule is specifically used to: calculate the association degree between the second target and the first target based on the Mahalanobis distance between the second target and the first target, and if the association degree between the second target and the first target is greater than a preset threshold, determine that the second target and the first target meet the fusion condition; based on the Mahalanobis distance between the third target and the first target or the second target, judge the association degree between the third target and the first target or the second target, and if the association degree between the third target and the first target or the second target is greater than a preset threshold, determine that the third target and the first target or the second target meet the fusion condition.

[0023] In an achievable implementation, the inhibition zone determination submodule is specifically used to: determine the current longitudinal distance between the main target and the ego vehicle, and the current longitudinal absolute speed of the main target; determine the length of the split inhibition zone in front of the main target based on the current longitudinal distance between the main target and the ego vehicle, the current longitudinal absolute speed of the main target, and a pre-set mapping relationship between the longitudinal distance between the main target and the ego vehicle, the longitudinal absolute speed of the main target, and the length of the split inhibition zone.

[0024] In an achievable implementation, the inhibition zone determination submodule is further configured to determine a width of a split inhibition zone in front of the main target based on a width of the main target.

[0025] In an achievable implementation, the target fusion module further includes: a scene determination submodule, used to determine the current traffic scene; wherein the current traffic scene includes the current road scene, the current longitudinal speed of the vehicle, the vehicle type of the main target, the longitudinal absolute speed of the third target, and at least one of the vehicle types of the third target.

[0026] In one feasible implementation, the step of determining the current traffic scene and determining whether to activate the split inhibition zone in front of the main target based on the current traffic scene includes: if the current road scene is a highway, the current longitudinal speed of the vehicle is above 60kph, the vehicle type of the main target is a large vehicle, the longitudinal absolute speed of the third target is greater than 5m / s, and the vehicle type of the third target is a small vehicle, then the step of determining the split inhibition zone in front of the main target is initiated.

[0027] In a third aspect, an embodiment of the present invention further provides a vehicle, comprising: a vehicle body; a first sensor and a second sensor, which are arranged on the vehicle body and are used to detect a neighboring vehicle target traveling in the same direction in an adjacent lane in front of the vehicle; a target fusion module, which is used to determine whether the first target detected by the first sensor, the second target detected by the second sensor, and the third target meet the fusion conditions; if the first target and the second target meet the fusion conditions, and the third target does not meet the fusion conditions, the first target and the second target are fused into a main target, and a split inhibition zone in front of the main target is determined; it is determined whether the third target is in the split inhibition zone, and whether the longitudinal speed difference between the main target and the third target is within the speed difference threshold range; if the third target is in the split inhibition zone, and the longitudinal speed difference between the main target and the third target is within the speed difference threshold range, the third target is suppressed from being output to the electronic control unit in the intelligent driving assistance system.

[0028] The vehicle target detection method, device and vehicle provided in the embodiment of the present application detect the neighboring vehicle target traveling in the same direction in the adjacent lane in front of the vehicle through the first sensor and the second sensor; obtain the first target detected by the first sensor, the second target and the third target detected by the second sensor; if the first target and the second target meet the fusion condition, and the third target does not meet the fusion condition, then fuse the first target and the second target into the main target, and determine the split inhibition zone in front of the main target; if the third target is in the split inhibition zone, and the longitudinal speed difference between the main target and the third target is within the speed difference threshold range, then suppress the output of the third target, that is, do not output the third target to the electronic control unit in the intelligent driving assistance system. In this way, the interference of the third target on the electronic control unit can be avoided, and the occurrence of the erroneous deceleration of the vehicle caused by the third target invading the lane of the vehicle can be reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0030] Figure 1 A schematic diagram of a process flow of an implementation method of a vehicle target detection method of the present invention;

[0031] Figure 2a A schematic diagram of vehicle target splitting in one implementation of the vehicle target detection method of the present invention;

[0032] Figure 2b A schematic diagram of a large vehicle target splitting scene in one implementation of a vehicle target detection method of the present invention;

[0033] Figure 3 A schematic diagram of determining a main target forward suppression area in an embodiment of a vehicle target detection method of the present invention;

[0034] Figure 4 A schematic diagram of not outputting radar targets in a suppression area in one embodiment of a vehicle target detection method of the present invention;

[0035] Figure 5 A flow chart of determining the forward suppression range of a main target in one embodiment of the vehicle target detection method of the present invention;

[0036] Figure 6 A flow chart of a process for suppressing the splitting of a large vehicle target in another embodiment of the vehicle target detection method of the present invention;

[0037] Figure 7The figure is a functional structure diagram of an implementation manner of a vehicle target detection device of the present invention. DETAILED DESCRIPTION

[0038] The embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. It should be clear that the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0039] An embodiment of the present invention provides a vehicle target detection method to reduce erroneous deceleration of a vehicle equipped with an intelligent driving assistance system.

[0040] In a first aspect, an embodiment of the present invention provides a vehicle target detection method, which is applicable to a vehicle detecting a neighboring vehicle traveling in the same direction in an adjacent lane in front of the vehicle. The detection method comprises the following steps:

[0041] S11. Detecting a neighboring vehicle target traveling in the same direction in an adjacent lane in front of the vehicle by using a first sensor and a second sensor.

[0042] The first sensor and the second sensor are sensors installed on the vehicle. The first sensor and the second sensor can be installed at the front of the vehicle to detect targets in front of the vehicle, including neighboring vehicle targets traveling in the same direction in the adjacent lane in front of the vehicle, so as to provide data basis for the auxiliary driving system of the vehicle.

[0043] The first sensor and the second sensor may be visual sensors (such as cameras) and / or radars. The radar may be a laser radar or a millimeter wave radar. The neighboring vehicle targets traveling in the same direction in the adjacent lane refer to one or more vehicle targets in the lane adjacent to the lane currently traveled by the vehicle, traveling in the same direction as the vehicle, and relatively close to the vehicle.

[0044] S12: Acquire a first target detected by the first sensor, and a second target and a third target detected by the second sensor.

[0045] The first sensor and the second sensor can work independently to respectively detect neighboring vehicle targets traveling in the same direction in the adjacent lane in front of the vehicle, thereby obtaining the first target detected by the first sensor and the second target and the third target detected by the second sensor.

[0046] Limited by the detection accuracy of the second sensor, the second target and the third target detected by the second sensor may be the same neighboring vehicle (for example, the body of the neighboring vehicle is longer and the track of the neighboring vehicle is split), or may be different neighboring vehicles.

[0047] S13: Determine whether the first target, the second target, and the third target meet the fusion conditions.

[0048] Determine whether the first target, the second target, and the third target meet the fusion condition, so as to determine whether at least two of the first target, the second target, and the third target can be fused into one target, that is, determine whether at least two of the first target, the second target, and the third target are a neighboring vehicle.

[0049] The fusion condition may be whether the distance and / or longitudinal speed difference (along the driving mode of the vehicle) of the two targets meets the preset condition. If the first target, the second target and the third target all meet the fusion condition, the first target, the second target and the third target are fused into one target. If the first target and the second target meet the fusion condition, but the third target does not meet the fusion condition, step S14 is executed.

[0050] S14, merging the first target and the second target into a main target, and determining a split inhibition zone in front of the main target.

[0051] The first target and the second target are merged into the main target, specifically, a maximum bounding rectangle frame of the area where the first target and the second target are located is established, and the maximum bounding rectangle frame is used as the main target.

[0052] Figure 2a It is a schematic diagram of merging the first target and the second target into the main target in one embodiment of the present invention. Figure 2a In the figure, the front radar target point is the target point detected by the radar installed at the front of the vehicle (which can be called the front radar, i.e. the second sensor); the forward-looking target point is the target point detected by the visual sensor installed at the front of the vehicle (which can be called the forward-looking, i.e. the first sensor). Figure 2a Two third targets that do not meet the fusion conditions are shown in FIG. 2 , that is, split targets.

[0053] Figure 2b It is a schematic diagram of merging the first target and the second target into the main target in an actual scene in one embodiment of the present invention. Figure 2b In the figure, the target formed by the single front radar is the split target formed when the front radar detects the neighboring vehicle. Figure 2b A third target that does not meet the fusion conditions is shown in FIG. 1 , which is a split target.

[0054] The split inhibition zone can be determined based on the safe following distance of the main target. For example, the safe following distance of the main target can be used as the length of the split inhibition zone, and the width of the main target can be used as the width of the split inhibition zone. The split inhibition zone can also be determined based on the current longitudinal distance between the main target and the ego vehicle, and the current longitudinal absolute speed of the main target.

[0055] Determining the split inhibition zone in front of the main target is convenient for judging whether the third target is a split target of the neighboring vehicle. The so-called split target of the neighboring vehicle refers to a target formed in front of the main target by part of the traces of the neighboring vehicle that are not associated with the main target.

[0056] S15: Determine whether the third target is in the split inhibition zone, and whether the longitudinal speed difference between the main target and the third target is within a speed difference threshold range.

[0057] Whether the third target is in the split inhibition zone may be determined by judging whether the coordinates of the third target are in the split inhibition zone.

[0058] See also Figure 3 As shown in Figure 1, assume that the length of the main target rectangle is I, the length of the suppression zone is L, and the width is W. The longitudinal distance X of the near edge of the suppression zone N = D xFus +I, far edge longitudinal distance X F =X N +L, lateral distance Y from the side of the vehicle N =D yFus -W / 2, lateral distance Y from the side of the vehicle F =D yFus +W / 2; where D xFus is the longitudinal distance of the main target relative to the vehicle, D yFus The lateral distance of the center of the main target relative to the vehicle.

[0059] If the longitudinal coordinate X and the transverse coordinate Y of the third target meet the following conditions, it can be determined that the coordinates of the third target are within the split inhibition zone:

[0060] X N <x<X F

[0061] Y N <y<Y F

[0062] The longitudinal speeds of the main target and the third target can be determined according to the detection of the second sensor. The longitudinal speed of the main target can be determined according to the longitudinal speed of the second target fused into the main target, that is, the longitudinal speed of the second target can be used as the longitudinal speed of the main target.

[0063] The longitudinal speed difference between the main target and the third target is too large, which basically means that the main target and the third target correspond to different neighboring vehicles. Therefore, it is necessary to determine a suitable speed difference threshold. When the longitudinal speed difference between the main target and the third target is within the speed difference threshold range, the third target may be a split target of the neighboring vehicle corresponding to the main target.

[0064] If the third target is in the splitting inhibition zone, and the longitudinal speed difference between the main target and the third target is within the speed difference threshold range, it can be basically determined that the third target is the splitting target of the neighboring vehicle corresponding to the main target, and step S16 can be executed at this time.

[0065] In order to prevent the speed difference between the split target and the main target from fluctuating back and forth near the speed difference threshold, thereby causing the split target to frequently switch between output and non-output, the speed difference between the split target and the main target can be set to exceed the time threshold τ continuously within the speed difference threshold before the output can be suppressed.

[0066] S16. Suppress the output of the third target.

[0067] The output of the third target is suppressed, that is, the third target is not output to the electronic control unit in the intelligent driving assistance system, so that the driving control of the vehicle by the electronic control unit is not affected by the third target.

[0068] The above steps S12-S16 can be executed by a target fusion module in the intelligent driving assistance system. Figure 4 It is a schematic diagram of not outputting the third target in the split inhibition zone according to an embodiment of the present invention.

[0069] The target fusion module in the suppression area will start the track of the third target, but will not output it. When the ego vehicle approaches, the length error of the visually recognized target becomes smaller. When the main target and the split target meet the association conditions, they will be associated, so that one visual target is associated with multiple targets mistakenly identified by the front radar.

[0070] The vehicle target detection method implementation method of the present application detects the neighboring vehicle target traveling in the same direction in the adjacent lane in front of the vehicle through the first sensor and the second sensor; obtains the first target detected by the first sensor, the second target and the third target detected by the second sensor; if the first target and the second target meet the fusion condition, and the third target does not meet the fusion condition, the first target and the second target are fused into the main target, and the split inhibition zone in front of the main target is determined; if the third target is in the split inhibition zone, and the longitudinal speed difference between the main target and the third target is within the speed difference threshold range, the output of the third target is suppressed, that is, the third target is not output to the electronic control unit in the intelligent driving assistance system. In this way, the interference of the third target to the electronic control unit can be avoided, and the occurrence of the vehicle's erroneous deceleration caused by the third target invading the vehicle's lane can be reduced.

[0071] In some embodiments, the first sensor is a visual sensor (such as a camera), and the second sensor is a radar, specifically a millimeter wave radar.

[0072] Accordingly, the step of detecting the neighboring vehicle target traveling in the same direction in the adjacent lane in front of the vehicle by using the first sensor and the second sensor (step S11) includes:

[0073] Through the visual sensor, the image of the adjacent lane in front of the vehicle is collected, and the neighboring vehicle target traveling in the same direction in the adjacent lane in front of the vehicle is determined based on image recognition; through the radar, the target of the adjacent lane in front of the vehicle is detected, and the neighboring vehicle target traveling in the same direction in the adjacent lane in front of the vehicle is determined.

[0074] Using visual sensors, the position of neighboring vehicles relative to the lane line can be determined based on image information, and the lane status of neighboring vehicles can be confirmed. In addition, a large number of models of different models can be trained through visual sensors, and the target type recognition accuracy is high, which can more accurately confirm whether the target type is a large vehicle (a large truck with a body length of more than ten meters or even more than twenty meters).

[0075] The visual sensor can process the acquired image information into target information, which includes the distance of the neighboring vehicle relative to the vehicle, the speed, the type of the neighboring vehicle, scene information, etc.

[0076] Millimeter-wave radar detects neighboring vehicles through radio waves, and uses the Doppler effect principle and high-resolution measurement technology to achieve high ranging and speed measurement accuracy. Millimeter-wave radar can process the reflected wave signal into target information, which may include the distance and speed of the neighboring vehicle relative to the vehicle. The speed information output by the millimeter-wave radar is the relative speed of the neighboring vehicle relative to the vehicle, which can be converted into an absolute speed.

[0077] In some embodiments, the relative speed is converted into an absolute speed as follows:

[0078] Taking the center of the rear axle of the vehicle as the origin of the coordinate system, calculate the distance R from the target to the origin and the angle θ with the x-axis of the vehicle coordinate system:

[0079] ;

[0080] θ = artan(dy / dx);

[0081] Superimpose the lateral and longitudinal velocities caused by the yaw angle of the vehicle:

[0082] V xhost =V x -ω×R×sinθ

[0083] V yhost =V y +ω×R×cosθ

[0084] Among them, V xhost is the longitudinal velocity of the vehicle after superposition; V yhost is the lateral velocity of the vehicle after superposition; ω is the yaw angular velocity of the vehicle; Vx is the longitudinal velocity of the vehicle; Vy is the lateral velocity of the vehicle;

[0085] According to the relative speed of the radar and the lateral and longitudinal speeds of the vehicle, the absolute speed of the radar target can be calculated:

[0086] V xFR =V xRel +V xhost ;

[0087] V yFR =V yRel +V yhost ;

[0088] Among them, V xFR V is the absolute longitudinal velocity of the front radar target; yFR is the absolute lateral velocity of the radar target.

[0089] In this embodiment, the main target is the fusion result of the target detected by the visual sensor (first sensor) installed in front of the vehicle (can be called the forward-looking target) and the target detected by the radar (second sensor, can be called the front radar) installed in front of the vehicle (can be called the front radar target). The split target is the point at other positions on the same neighboring vehicle output by the front radar of the vehicle that cannot be clustered due to the large distance difference.

[0090] In this embodiment, multimodal data analysis can be performed based on the detection data of the visual sensor and the millimeter-wave radar, which facilitates the fusion processing of the targets detected by the visual sensor and the targets detected by the radar, and can greatly improve the accuracy of target detection.

[0091] In some implementations, the determining whether the first target, the second target, and the third target meet the fusion condition (step S13) may include:

[0092] S131, based on the Mahalanobis distance between the second target and the first target, calculating the correlation between the second target and the first target, if the correlation between the second target and the first target is greater than a preset threshold, determining that the second target and the first target meet the fusion condition;

[0093] S132. Based on the Mahalanobis distance between the third target and the first target or the second target, determine the correlation between the third target and the first target or the second target. If the correlation between the third target and the first target or the second target is greater than a preset threshold, determine that the third target and the first target or the second target meet the fusion condition.

[0094] The data measured by the visual sensor can be aligned in time and space with the data measured by the radar. The Mahalanobis distance of the speed, distance and other dimensions of the targets measured by the visual sensor and the radar can be calculated to calculate the correlation of the targets, thereby determining whether the targets measured by the visual sensor (referred to as sensor targets) and the targets measured by the radar (referred to as radar targets) are correlated.

[0095] In some implementations, in the above step S131, calculating the association degree between the second target and the first target based on the Mahalanobis distance between the second target and the first target may include:

[0096] According to the following formula, the correlation F between the second target and the first target is calculated:

[0097] ;

[0098] in, ;

[0099] D M1 D is the Mahalanobis distance calculated based on the longitudinal distance of the target (i.e., the neighboring vehicle) relative to the vehicle measured by the visual sensor and radar respectively. M2 is the Mahalanobis distance calculated based on the lateral distance of the neighboring vehicle relative to the vehicle measured by the visual sensor and radar, respectively. M3 is the Mahalanobis distance calculated based on the longitudinal speed of the neighboring vehicle relative to the vehicle measured by the visual sensor and radar, respectively, D M4 is the Mahalanobis distance calculated based on the lateral speed of the neighboring vehicle relative to the own vehicle measured by the visual sensor and radar respectively, and σ is the empirical coefficient.

[0100] In some examples, D can be calculated according to the following formula M1 :

[0101] ;

[0102] Among them, x is the longitudinal distance of the second target relative to the vehicle measured by the radar (which can be simply referred to as the radar target longitudinal distance), μ is the longitudinal distance of the first target relative to the vehicle measured by the visual sensor (which can be simply referred to as the visual longitudinal distance), and ∑ is the covariance matrix. It should be understood that there is only a single dimension of the longitudinal distance here, so it can be simplified to the sum of the variance of the radar target longitudinal distance and the variance of the forward-looking distance.

[0103] Similarly, the corresponding Mahalanobis distance D can be calculated based on the lateral distance, longitudinal speed, and lateral speed of the target relative to the vehicle measured by the visual sensor and radar respectively. M2 , D M3 , D M4 .

[0104] It should be understood that other conditions, such as the heading angle of the target, etc., may also be added to calculate the Mahalanobis distance to increase the reliability of the association.

[0105] The degree of association F can also be called the association score F. When F≥δ, it indicates that the two targets are associated; where δ is a preset association threshold.

[0106] The method for determining the correlation between the third target and the first target or the second target may be the same as the method for determining the correlation between the second target and the first target, and will not be described in detail.

[0107] The speed and distance of the target detected by each sensor may vary greatly, and the calculated correlation score is small, which will lead to the inability to correlate. In this case, a track can be started for the target that cannot be correlated, forming a single sensor target for output. For extra-long trucks of more than ten meters or even more than twenty meters, there are many radar reflection points that are sparse and incoherent, and the distance of the clustered targets is far apart, which will cause the large target to be split into multiple small targets. Therefore, multiple single radar targets will be formed near the fusion target (main target), such as Figure 2a shown. Figure 2a It is shown that there are two split targets formed near the front of the main target, which are single radar targets (targets formed only by radar).

[0108] The associated target is the fused target (primary target), which combines the following advantages of visual sensors and millimeter-wave radars:

[0109] 1. Longitudinal position and speed follow the front millimeter-wave radar with higher accuracy:

[0110] V xFus =V xFR , D xFus =D xFR

[0111] Among them, V xFus is the fusion target longitudinal velocity, V xFR is the longitudinal absolute velocity of the associated radar target, D xFus is the longitudinal distance of the fusion target relative to the vehicle, D xFR is the longitudinal distance of the associated radar target relative to the vehicle.

[0112] (ii) Lateral position and speed follow the visual sensor with higher accuracy:

[0113] V yFus =V yFR , D yFus =D yFR

[0114] Where V yFus is the fusion target lateral velocity, V yFR The absolute lateral velocity of the associated visual sensor target, D yFus is the lateral distance of the fusion target relative to the vehicle, D yFR is the lateral distance of the associated visual sensor target relative to the ego vehicle.

[0115] If there is no new measurement update in the current frame, that is, the radar detection information is not updated, the longitudinal velocity and longitudinal distance of the fused target can be predicted using the CA motion model (constant acceleration motion model).

[0116] Let state x = (v, d), v represents velocity, d represents distance, and acceleration is a k . Given k-1 states x k-1 =(v k-1 , d k-1 ), then at the next moment x k =(v k , d k )

[0117] d k =d k-1 +Δtv k-1 +1 / 2a k (Δt)²

[0118] v k =v k-1 +a k Δt

[0119] Written in the form of state transfer:

[0120] x k =Fx k-1 +Ba k (1)

[0121] F = : State transfer matrix

[0122] B = : Control matrix

[0123] In reality, there are various disturbances due to external influences, so noise will be mixed in the state transfer process, and this noise cannot be measured. However, it can be considered that this noise obeys a zero-mean Gaussian distribution, so formula (1) can be rewritten as:

[0124] x k =Fx k-1 +Bu k +w k (2)

[0125] u k is the control input, which is a in this embodiment k , w k is the process noise, w k ~N(0,Q k ), Q k is its covariance matrix.

[0126] If there is a new measurement update in the current frame, that is, the radar detection information is updated, the updated measurement of the current frame is used to correct the predicted value and perform iterative calculation using the Kalman filter.

[0127] The observation value of the current frame (i.e. the measurement updated by the sensor) and the state x k Relationship: y k =Hx k +v k (3)

[0128] H is the observation transfer matrix, v k is the observation noise, which follows a Gaussian distribution with a mean of 0, that is, v k ~N(0,R k ), R k is its covariance matrix.

[0129] According to equations (2) and (3), assuming that at time k-1, there is an estimated value for the system k-1 , the covariance of this estimate is P k-1 , using the state transfer equation to get the predicted state k - =F k-1 +Bu k , the covariance P of this predicted state k - =FP k-1 F T +Q k

[0130] Then the Kalman gain K at time k can be derived k

[0131] K k =P k - H T (HP k - H T +R k ) -1

[0132] According to the current prediction value, observation value and Kalman gain, the optimal estimate at this moment can be derived

[0133] k = k - +K k (y k -H k- )

[0134] Based on the predicted value covariance at the current moment, the Kalman gain deduces the optimal estimated value covariance at this moment

[0135] P k = (1-K k H)P k -

[0136] By repeating the prediction update cycle, we can get the optimal estimate at each moment.

[0137] In some embodiments, in step S14, determining the split inhibition zone in front of the main target may include: determining the current longitudinal distance between the main target and the ego vehicle, and the current longitudinal absolute speed of the main target; determining the length of the split inhibition zone in front of the main target based on the current longitudinal distance between the main target and the ego vehicle, the current longitudinal absolute speed of the main target, and a pre-set mapping relationship between the longitudinal distance between the main target and the ego vehicle, the longitudinal absolute speed of the main target, and the length of the split inhibition zone.

[0138] The longitudinal distance between the near edge of the vehicle and the rear axle of the vehicle and the circumscribed rectangular frame of the main target that merges the first target and the second target can be used as the current longitudinal distance between the main target and the vehicle; the longitudinal absolute speed of the second target can be used as the current longitudinal absolute speed of the main target.

[0139] Based on the current longitudinal distance between the main target and the ego vehicle and the current longitudinal absolute speed of the main target, the pre-set mapping relationship table of the longitudinal distance between the main target and the ego vehicle, the longitudinal absolute speed of the main target and the length of the split inhibition zone can be traversed to obtain the length of the split inhibition zone in front of the main target.

[0140] In some embodiments, the mapping relationship between the longitudinal distance between the main target and the ego vehicle, the longitudinal absolute speed of the main target and the length of the splitting inhibition zone can be as follows: Figure 5 shown.

[0141] Figure 5 It is also a flow chart for determining the forward suppression range of the main target in one embodiment of the vehicle target detection method of the present invention, according to Figure 5 As shown, for the longitudinal distance between the current main target and the ego vehicle, if the longitudinal distance between the current main target and the ego vehicle is > 80m, and the longitudinal absolute speed of the main target is > 80kph, then the length of the split inhibition zone in front of the main target (also called the forward direction) is 12m;

[0142] If the current longitudinal distance between the main target and the ego vehicle is >40m, and the longitudinal absolute speed of the main target is >80kph, then the length of the split inhibition zone in front of the main target is 10m;

[0143] If the current longitudinal distance between the main target and the ego vehicle is <40m, and the longitudinal absolute speed of the main target is >80kph, then the length of the split inhibition zone in front of the main target is 8m;

[0144] If the current longitudinal distance between the main target and the ego vehicle is > 80m, and the longitudinal absolute speed of the main target is > 40kph, then the length of the split inhibition zone in front of the main target is 9m;

[0145] If the current longitudinal distance between the main target and the ego vehicle is > 80m, and the longitudinal absolute speed of the main target is > 18kph, then the length of the split inhibition zone in front of the main target is 7m;

[0146] If the current longitudinal distance between the main target and the ego vehicle is > 80m, and the longitudinal absolute speed of the main target is < 18kph, then there is no inhibition range in front of the main target, that is, there is no split inhibition zone.

[0147] Generally speaking, considering the driver's reaction time and the sense of crisis when following a close-following object, especially in the case of following a large truck, the braking distance of the large truck is longer. Therefore, the higher the speed of the vehicle on the road, the greater the reserved distance between vehicles. Based on this premise, the split suppression zone (also called suppression range) in front of the main target should be set smaller than the following distance on a normal road.

[0148] Figure 5 The illustrated embodiment is mainly used to determine the length of the split inhibition zone in front of the main target.

[0149] In some embodiments, the width of the split inhibition zone may be determined based on the width of the main target. In some examples, the width of the main target may be directly used as the width of the split inhibition zone in front of the main target.

[0150] Because the lateral distance of the target detected by the radar generally has a large error, the width of the split suppression zone can be expanded by a certain distance based on the width of the main target, such as 0.1m, 0.2m, etc. In an example, the split suppression zone after expansion is as follows Figure 3 shown.

[0151] In other embodiments, the width of the split inhibition zone may also be determined according to a certain ratio of the length of the split inhibition zone, such as using one third or one fourth of the length of the split inhibition zone as the width of the split inhibition zone.

[0152] In order to prevent frequent calculation of the split inhibition zone, in some embodiments, before determining the split inhibition zone in front of the main target, the method may further include:

[0153] Determine the current traffic scene, and determine whether to activate the split inhibition zone in front of the main target based on the current traffic scene; wherein the current traffic scene includes at least one of the current road scene, the current longitudinal speed of the vehicle, the vehicle type of the main target, the longitudinal absolute speed of the third target, and the vehicle type of the third target.

[0154] The information detected by the visual sensor can be used to determine the current road scene, such as whether it is a high-speed scene.

[0155] Obtain vehicle body data to determine the current longitudinal speed of the vehicle (also referred to as vehicle speed) and determine whether the current vehicle speed meets the requirements of the current road scene. For example, based on the minimum speed requirement for expressways, determine whether the current vehicle speed is not less than 60 kph.

[0156] In the current road scene, whether the current vehicle speed meets the requirements of the current road scene,

[0157] The models of the main target and the third target (split target) can be determined. Specifically, the model of the main target can be determined by the information detected by the visual sensor, and the model of the third target can be determined by the data detected by the radar. If the model of the main target is determined to be a large vehicle through the information detected by the visual sensor, and the split radar target is determined to be a small vehicle (for example, the body length is 3.5m-4.6m) through the data detected by the radar, the longitudinal speed of the third target is calculated. In other implementation examples, the fusion target type is a vehicle with a length greater than 7m, which is regarded as a large vehicle, and the split target type is a vehicle with a length less than 7m, which is regarded as a small vehicle.

[0158] According to the target speed information given by the sensor, the longitudinal speed of the third target (radar target) is calculated. Since the radar has a low recognition accuracy for low-speed targets, only radar targets with longitudinal speeds greater than the specified speed are suppressed, such as only radar targets with longitudinal speeds greater than 5m / s. If the calculated longitudinal speed of the third target is greater than the specified speed, such as greater than 5m / s, the step of splitting the suppression zone in front of the main target is initiated.

[0159] In some embodiments, the current traffic scene is determined, and the step of determining whether to activate the split inhibition zone in front of the main target based on the current traffic scene includes: if the current road scene is a highway, the current longitudinal speed of the vehicle is above 60kph, the vehicle type of the main target is a large vehicle, the longitudinal absolute speed of the third target is greater than 5m / s, and the vehicle type of the third target is a small vehicle, then the step of determining the split inhibition zone in front of the main target is initiated.

[0160] In some embodiments, the speed difference threshold in step S15 can be determined as follows: a large truck is a rigid object, so the movement speed of each point along the direction of the large truck is the same. The millimeter wave radar can detect the radial distance, radial speed and angle from the target point to the radar. The movement speed of each target point can be calculated based on this information. Due to the noise in the radar measurement and the different angles of different points on the large truck relative to the front radar, the movement speed of each target point finally calculated has errors. Therefore, when |V xFus -V FR When ︱≤ξ, the two can be considered as the same target, where V xFus is the longitudinal velocity of the fused target, which is equal to the velocity of the associated front radar target, V FR is the moving speed of the front radar target associated with the split target, and ξ can obtain the maximum value based on a large number of repeated experiments.

[0161] The vehicle target detection method of the present invention is described below with reference to a specific embodiment.

[0162] See also Figure 6 The vehicle target detection method of this embodiment includes the following steps:

[0163] S601, acquiring perception data of the forward vision and front radar;

[0164] S602, calculating the correlation score between the forward-looking target and the forward radar target;

[0165] S603, determining whether the correlation score F is greater than or equal to a correlation score threshold of 0.8;

[0166] If the correlation score is greater than or equal to the correlation score threshold 0.8, step S604 is executed; otherwise, step S605 is executed.

[0167] S604, determine whether the following conditions are met at the same time: the vehicle speed is greater than or equal to 60 kph, the fusion target type is a large vehicle, the single radar target (i.e., the split target) is a small vehicle, the current scene is a high-speed scene, and the single radar target speed is greater than 5 m / s;

[0168] If the above conditions are met at the same time, execute step S606, otherwise, execute step S610;

[0169] S605, output single vision and single front radar targets;

[0170] S606, calculating the suppression area and the speed difference between the fusion target and the single radar target;

[0171] S607, determining whether the single radar target is within the suppression area;

[0172] If the single radar target is within the suppression area, execute step S608, otherwise, execute step S610;

[0173] S608, determine whether the longitudinal speed difference between the fusion target and the single radar target is within the speed difference threshold 1m / s, and continuously exceeds the predetermined time 0.5s; if so, execute step S609, otherwise, execute step S610;

[0174] S609, suppress single radar target output;

[0175] S610: Output the fusion target and the single radar target.

[0176] In the embodiment of the present invention, by suppressing the output of the split target of the large truck, the track splitting problem of the large truck in the adjacent lane in front of the vehicle and the problem of the split target invading the lane and causing the vehicle to brake incorrectly can be solved. Therefore, the main target type is mainly a large truck in the adjacent lane in front of the vehicle. The target type and the rectangular frame (box) of the target can be obtained and assigned based on the target information perceived by the forward vision (the visual sensor in front of the vehicle), and the relative position of the lane line and the lane can also be known from the forward vision perception target information.

[0177] The farther the target in the adjacent lane is from the longitudinal position of the vehicle, the less target side information obtained by the forward vision, so the error in calculating the length of the target is larger, and there is relatively more time to react when a distant real target intrudes, so the front suppression distance of the distant main target can be increased.

[0178] In a second aspect, an embodiment of the present invention further provides a vehicle target detection device for detecting a neighboring vehicle target traveling in the same direction in an adjacent lane in front of the vehicle, see Figure 7 The vehicle target detection device of this embodiment includes a target acquisition module 201 and a target fusion module 202.

[0179] The target acquisition module 201 is used to acquire a first target detected by a first sensor, a second target detected by a second sensor, and a third target; wherein the first sensor and the second sensor are sensors on the vehicle;

[0180] The target fusion module 202 is used to determine whether the first target, the second target and the third target meet the fusion conditions; if the first target and the second target meet the fusion conditions, and the third target does not meet the fusion conditions, the first target and the second target are fused into a main target, and a split inhibition zone in front of the main target is determined; whether the third target is in the split inhibition zone, and whether the longitudinal speed difference between the main target and the third target is within the speed difference threshold range; if the third target is in the split inhibition zone, and the longitudinal speed difference between the main target and the third target is within the speed difference threshold range, the output of the third target is suppressed.

[0181] This embodiment can implement the vehicle target detection method described in the previous embodiment, and its implementation process and technical effects are basically the same, which will not be repeated here.

[0182] In some embodiments, the first sensor is a visual sensor, and the second sensor is a radar; the visual sensor is used to collect images of the adjacent lane in front of the vehicle, and the neighboring vehicle target traveling in the same direction in the adjacent lane in front of the vehicle is determined based on image recognition; the radar is used to perform target detection on the adjacent lane in front of the vehicle, and the neighboring vehicle target traveling in the same direction in the adjacent lane in front of the vehicle is determined.

[0183] In some embodiments, the target fusion module includes:

[0184] A fusion condition judgment submodule, used to judge whether the first target, the second target and the third target meet the fusion condition;

[0185] A target fusion submodule, configured to fuse the first target and the second target into a main target if the first target and the second target meet a fusion condition and the third target does not meet the fusion condition;

[0186] An inhibition zone determination submodule, used to determine a split inhibition zone in front of the main target;

[0187] The target output suppression submodule is used to determine whether the third target is in the split suppression zone and whether the longitudinal speed difference between the main target and the third target is within the speed difference threshold range; if the third target is in the split suppression zone and the longitudinal speed difference between the main target and the third target is within the speed difference threshold range, the output of the third target is suppressed.

[0188] In some embodiments, the fusion condition judgment submodule is specifically used to: calculate the association between the second target and the first target based on the Mahalanobis distance between the second target and the first target, and if the association between the second target and the first target is greater than a preset threshold, determine that the second target and the first target meet the fusion condition; based on the Mahalanobis distance between the third target and the first target or the second target, judge the association between the third target and the first target or the second target, and if the association between the third target and the first target or the second target is greater than a preset threshold, determine that the third target and the first target or the second target meet the fusion condition.

[0189] Among them, the method for calculating the correlation between the second target and the first target is basically the same as the method in the aforementioned method implementation method, and will not be repeated here.

[0190] In some embodiments, the inhibition zone determination submodule is specifically used to: determine the current longitudinal distance between the main target and the ego vehicle, and the current longitudinal absolute speed of the main target; determine the length of the split inhibition zone in front of the main target based on the current longitudinal distance between the main target and the ego vehicle, the current longitudinal absolute speed of the main target, and a pre-set mapping relationship between the longitudinal distance between the main target and the ego vehicle, the longitudinal absolute speed of the main target, and the length of the split inhibition zone.

[0191] In some embodiments, the inhibition zone determination submodule is further used to determine the width of the splitting inhibition zone in front of the main target based on the width of the main target.

[0192] In some embodiments, the target fusion module also includes: a scene determination submodule, used to determine the current traffic scene; wherein the current traffic scene includes the current road scene, the current longitudinal speed of the vehicle, the vehicle type of the main target, the longitudinal absolute speed of the third target, and at least one of the vehicle types of the third target.

[0193] In some embodiments, if the current road scene is a highway, the current longitudinal speed of the vehicle is above 60 kph, the vehicle type of the main target is a large vehicle, the absolute longitudinal speed of the third target is greater than 5 m / s, and the vehicle type of the third target is a small vehicle, then the step of determining the split inhibition zone in front of the main target is initiated.

[0194] In the implementation manner of the present invention, by suppressing the output of the split target of the large truck, the problem of the target split from the large truck in the adjacent lane in front of the vehicle invading the vehicle's lane and causing the vehicle to brake incorrectly can be solved.

[0195] In a third aspect, an embodiment of the present invention further provides a vehicle, comprising: a vehicle body; a first sensor and a second sensor, which are arranged on the vehicle body and are used to detect a neighboring vehicle target traveling in the same direction in an adjacent lane in front of the vehicle; a target fusion module, which is used to determine whether the first target detected by the first sensor, the second target detected by the second sensor, and the third target meet the fusion conditions; if the first target and the second target meet the fusion conditions, and the third target does not meet the fusion conditions, the first target and the second target are fused into a main target, and a split inhibition zone in front of the main target is determined; it is determined whether the third target is in the split inhibition zone, and whether the longitudinal speed difference between the main target and the third target is within the speed difference threshold range; if the third target is in the split inhibition zone, and the longitudinal speed difference between the main target and the third target is within the speed difference threshold range, the third target is suppressed from being output to the electronic control unit in the intelligent driving assistance system.

[0196] In some embodiments, the target fusion module in this embodiment may be the target fusion module described in any embodiment of the aforementioned detection device.

[0197] The vehicle provided in the embodiment of the present application detects the neighboring vehicle target traveling in the same direction in the adjacent lane in front of the vehicle through the first sensor and the second sensor; obtains the first target detected by the first sensor, the second target and the third target detected by the second sensor; if the first target and the second target meet the fusion condition, and the third target does not meet the fusion condition, the first target and the second target are fused into the main target, and the split inhibition zone in front of the main target is determined; if the third target is in the split inhibition zone, and the longitudinal speed difference between the main target and the third target is within the speed difference threshold range, the output of the third target is suppressed, that is, the third target is not output to the electronic control unit in the intelligent driving assistance system. In this way, the interference of the third target to the electronic control unit can be avoided, and the occurrence of the erroneous deceleration of the vehicle caused by the third target invading the lane of the vehicle can be reduced.

[0198] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0199] Each embodiment in this specification is described in a related manner, and the same or similar parts between the embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0200] In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0201] For the convenience of description, the above device is described by dividing the functions into various units / modules. Of course, when implementing the present invention, the functions of each unit / module can be implemented in the same or multiple software and / or hardware.

[0202] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0203] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A vehicle target detection method, characterized in that: include: Detecting a neighboring vehicle target traveling in the same direction in an adjacent lane in front of the vehicle by means of a first sensor and a second sensor; wherein the first sensor and the second sensor are sensors on the vehicle; Acquire a first target detected by the first sensor, a second target and a third target detected by the second sensor; Determining whether the first target, the second target, and the third target meet fusion conditions; If the first target and the second target meet the fusion condition, and the third target does not meet the fusion condition, the first target and the second target are fused into a main target, and a split inhibition zone in front of the main target is determined; determining whether the third target is in the split inhibition zone, and whether a longitudinal speed difference between the main target and the third target is within a speed difference threshold range; If the third target is within the split suppression zone and the longitudinal speed difference between the main target and the third target is within the speed difference threshold range, the output of the third target is suppressed.

2. The vehicle target detection method according to claim 1, characterized in that: The first sensor is a visual sensor, and the second sensor is a radar; The detecting of the neighboring vehicle target traveling in the same direction in the adjacent lane in front of the vehicle by the first sensor and the second sensor includes: The visual sensor collects an image of the adjacent lane in front of the vehicle, and determines the adjacent vehicle target traveling in the same direction in the adjacent lane in front of the vehicle based on image recognition; The radar is used to detect targets in the adjacent lane in front of the vehicle, and determine the adjacent vehicle target traveling in the same direction in the adjacent lane in front of the vehicle.

3. The vehicle target detection method according to claim 2, characterized in that: The determining whether the first target, the second target, and the third target meet the fusion condition includes: Based on the Mahalanobis distance between the second target and the first target, calculating the correlation between the second target and the first target, and if the correlation between the second target and the first target is greater than a preset threshold, determining that the second target and the first target meet the fusion condition; Based on the Mahalanobis distance between the third target and the first target or the second target, the correlation between the third target and the first target or the second target is determined; if the correlation between the third target and the first target or the second target is greater than a preset threshold, it is determined that the third target and the first target or the second target meet the fusion condition.

4. The vehicle target detection method according to claim 3, characterized in that: The calculating the association degree between the second target and the first target based on the Mahalanobis distance between the second target and the first target includes: The correlation degree F between the second target and the first target is calculated according to the following formula: ； in, ; D M1 is the Mahalanobis distance calculated based on the longitudinal distances of the neighboring vehicle relative to the vehicle measured by the visual sensor and the radar, respectively, D M2 is the Mahalanobis distance calculated based on the lateral distances of the neighboring vehicle relative to the vehicle measured by the visual sensor and the radar, respectively, D M3 is the Mahalanobis distance calculated based on the longitudinal speed of the neighboring vehicle relative to the vehicle measured by the visual sensor and the radar, respectively, D M4 is the Mahalanobis distance calculated according to the lateral speed of the neighboring vehicle relative to the own vehicle measured by the visual sensor and the radar respectively, and σ is an empirical coefficient.

5. The vehicle target detection method according to claim 2, characterized in that: The determining of the splitting inhibition zone in front of the main target includes: Determining a current longitudinal distance between the main target and the ego vehicle, and a current longitudinal absolute speed of the main target; The length of the split inhibition zone in front of the main target is determined based on the current longitudinal distance between the main target and the vehicle, the current longitudinal absolute speed of the main target, and the pre-set mapping relationship between the longitudinal distance between the main target and the vehicle, the longitudinal absolute speed of the main target and the length of the split inhibition zone.

6. The vehicle target detection method according to claim 2 or 5, characterized in that: The determining of the splitting inhibition zone in front of the main target includes: Based on the width of the main target, the width of the splitting inhibition zone in front of the main target is determined.

7. The vehicle target detection method according to claim 2, characterized in that: Before determining the splitting inhibition zone in front of the main target, the method further includes: Determine the current traffic scene, and determine whether to activate the split inhibition zone in front of the main target based on the current traffic scene; wherein the current traffic scene includes at least one of the current road scene, the current longitudinal speed of the vehicle, the vehicle type of the main target, the longitudinal absolute speed of the third target, and the vehicle type of the third target.

8. The vehicle target detection method according to claim 7, characterized in that: The step of determining the current traffic scene and determining whether to activate the split inhibition zone in front of the main target based on the current traffic scene includes: If the current road scene is a highway, the current longitudinal speed of the vehicle is above 60kph, the vehicle type of the main target is a large vehicle, the longitudinal absolute speed of the third target is greater than 5m / s, and the vehicle type of the third target is a small vehicle, then the step of determining the split inhibition zone in front of the main target is initiated.

9. A vehicle target detection device, characterized in that: The device is used to detect a neighboring vehicle target traveling in the same direction in an adjacent lane in front of the vehicle, and comprises: A target acquisition module, used to acquire a first target detected by a first sensor, a second target and a third target detected by a second sensor; wherein the first sensor and the second sensor are sensors on the vehicle; A target fusion module is used to determine whether the first target, the second target and the third target meet the fusion conditions; if the first target and the second target meet the fusion conditions, and the third target does not meet the fusion conditions, the first target and the second target are fused into a main target, and a split inhibition zone in front of the main target is determined; determine whether the third target is in the split inhibition zone, and whether the longitudinal speed difference between the main target and the third target is within the speed difference threshold range; if the third target is in the split inhibition zone, and the longitudinal speed difference between the main target and the third target is within the speed difference threshold range, suppress the output of the third target.

10. A vehicle, characterized in that: include: A vehicle body, wherein the vehicle body is equipped with an intelligent driving assistance system; The first sensor and the second sensor are arranged on the vehicle body and are used to detect a neighboring vehicle target traveling in the same direction in an adjacent lane in front of the vehicle; A target fusion module, used to determine whether a first target detected by the first sensor, a second target detected by the second sensor, and a third target meet a fusion condition; If the first target and the second target meet the fusion conditions, and the third target does not meet the fusion conditions, the first target and the second target are fused into a main target, and a split inhibition zone in front of the main target is determined; whether the third target is within the split inhibition zone, and whether the longitudinal speed difference between the main target and the third target is within a speed difference threshold range is determined; if the third target is within the split inhibition zone, and the longitudinal speed difference between the main target and the third target is within the speed difference threshold range, the third target is suppressed from being output to the electronic control unit in the intelligent driving assistance system.

Images