False target identification method and device based on fence area, electronic equipment and vehicle

Abstract

The embodiment of the invention discloses a false target identification method and device based on a fence area, electronic equipment and a vehicle, relates to the technical field of target identification, and can effectively enhance the identification performance in a scene with complex false target formation reasons, simplify deployment and reduce the operand. The method comprises the following steps: identifying trace points in a preset identification range of the side part of a vehicle, and determining fence trace points in the identification range; wherein the trace points in the identification range are detected by a millimeter wave radar on the vehicle; calculating the distance between the fence trace point and the first target; based on the distance between the fence trace points and a first target, judging the number of the fence trace points in a specified range around the first target; and determining whether the first target is a false target based on the number of the fence trace points in the specified range. The method is suitable for false target identification based on the fence area.

Description

Technical Field

[0001] The present invention relates to the technical field of target recognition, and in particular to a false target recognition method, device, electronic equipment and vehicle based on a fence area. Background Art

[0002] In the field of intelligent driving, millimeter-wave radar is widely used in automobile safety systems due to its advantages such as easy installation, no influence from weather, strong anti-interference ability in complex environments, and low cost.

[0003] In high-speed scenarios, due to multipath reflection, speed deambiguation errors and other reasons, false motion points are easily formed at fences or near locations. When the speeds of the false points are close, they are easily associated to form false tracks. If the false tracks are not eliminated, it may affect the driving safety of the vehicle. Therefore, the radar is required to detect the ghost tracks formed by false motion points at the fences in high-speed scenarios and suppress them.

[0004] In the related technologies, the suppression of false targets is mainly focused on multipath reflection scenarios. Multipath reflection scenarios require deep learning, and the method of using deep learning to identify false targets is difficult to deploy; for some false targets caused by angle measurement errors and other reasons, their properties may not be different from normal targets, and false targets cannot be well distinguished; when identifying false targets, the curvature radius of the own vehicle and the target vehicle is used. The curvature radius of the target vehicle itself requires more than one sensor, and a single radar cannot suppress false targets and needs to rely on information from other sensors, resulting in a large amount of calculation. Summary of the invention

[0005] In view of this, the embodiments of the present invention provide a false target recognition method, device, electronic device and vehicle based on a fence area, which can improve the accuracy of false target recognition, simplify deployment and reduce the amount of calculation.

[0006] In a first aspect, an embodiment of the present invention provides a false target identification method based on a fence area, comprising: identifying points within a preset identification range on the side of a vehicle, and determining fence points within the identification range; wherein the points within the identification range are detected by a millimeter-wave radar on the vehicle; calculating the distance between the fence points and a first target; based on the distance between the fence points and the first target, determining the number of fence points within a specified range around the first target; and determining whether the first target is a false target based on the number of fence points within the specified range.

[0007] According to a specific implementation method of an embodiment of the present application, the point traces within a preset recognition range on the side of the vehicle are identified to determine the fence point traces within the recognition range, including: calculating the relative position gradient between two adjacent point traces within the preset recognition range on the side of the vehicle; if the relative position gradient between the two adjacent point traces is less than a preset gradient threshold, then determining that the two adjacent point traces are fence point traces.

[0008] According to a specific implementation of the embodiment of the present application, the preset gradient threshold is determined according to the following formula: ;in, represents the distance coefficient, Indicates The radial distance of each fence point relative to the vehicle, Indicates a fixed distance threshold.

[0009] According to a specific implementation method of an embodiment of the present application, the calculation of the relative position gradient between two adjacent point traces within a preset recognition range on the side of the vehicle includes: calculating the absolute value of the difference between the horizontal coordinates of two adjacent point traces within the preset recognition range on the side of the vehicle in the vehicle coordinate system, and obtaining the lateral position gradient between the two adjacent point traces; and / or calculating the absolute value of the difference between the vertical coordinates of two adjacent point traces within the preset recognition range on the side of the vehicle in the vehicle coordinate system, and obtaining the longitudinal position gradient between the two adjacent point traces; wherein, if the relative position gradient between the two adjacent point traces is less than a preset gradient threshold, then the two adjacent point traces are determined to be fence point traces, including: if the lateral position gradient and / or longitudinal position gradient between the two adjacent point traces is less than a preset gradient threshold, then the two adjacent point traces are determined to be fence point traces.

[0010] According to a specific implementation method of an embodiment of the present application, the calculation of the distance between the fence point trace and the first target includes: rotating the self-vehicle coordinate system around the origin of the self-vehicle coordinate system so that the direction of the longitudinal axis of the self-vehicle coordinate system is consistent with the direction of the longitudinal axis of the coordinate system of the first target, thereby obtaining the rotated self-vehicle coordinate system; in the rotated self-vehicle coordinate system, calculating the distance between the fence point trace and the first target.

[0011] According to a specific implementation method of an embodiment of the present application, the method of determining the number of fence points within a specified range around a first target based on the distance between the fence points and the first target includes determining the number of fence points within a specified range in front of and behind the first target based on the distance between the fence points and the first target.

[0012] According to a specific implementation method of an embodiment of the present application, determining whether the first target is a false target based on the number of fence points within the specified range includes: counting the number of fence points within the specified range in the current frame to see whether it is greater than a preset threshold, and if the number of fence points within the specified range in the current frame is greater than the preset threshold, determining that the first target is a false target; or counting the number of fence points within the specified range in N consecutive frames to see whether they are all greater than the preset threshold, and if the number of fence points within the specified range in the N consecutive frames is all greater than the preset threshold, determining that the first target is a false target; or if kf>m in N consecutive frames, determining that the first target is a false target; wherein k is the number of frames in which the number of fence points within the specified range is greater than the preset threshold in the N consecutive frames; f is the number of frames in which the number of fence points within the specified range is 0 in the N consecutive frames; and m is the preset threshold.

[0013] According to a specific implementation method of an embodiment of the present application, after determining whether the first target is a false target based on the number of fence points within the specified range, the method also includes: determining the points obscured by the second target detected by the millimeter-wave radar; and suppressing the output of the points obscured by the second target.

[0014] In a second aspect, an embodiment of the present invention further provides a false target identification device based on a fence area, comprising: a first determination unit, for identifying points within a preset identification range on the side of a vehicle, and determining fence points within the identification range; wherein the points within the identification range are detected by a millimeter-wave radar on the vehicle; a calculation unit, for calculating the distance between the fence points and a first target; a judgment unit, for judging the number of fence points within a specified range around the first target based on the distance between the fence points and the first target; and a second determination unit, for determining whether the first target is a false target based on the number of fence points within the specified range.

[0015] According to a specific implementation method of an embodiment of the present application, the first determination unit includes: a first calculation module, used to calculate the relative position gradient between two adjacent point traces within a preset identification range of the side of the vehicle; a determination module, used to determine that the two adjacent point traces are fence point traces if the relative position gradient between the two adjacent point traces is less than a preset gradient threshold.

[0016] According to a specific implementation method of an embodiment of the present application, the first calculation module includes: a calculation submodule, which is specifically used to calculate the absolute value of the difference between the horizontal coordinates of two adjacent point traces within a preset identification range on the side of the vehicle in the vehicle coordinate system, and obtain the lateral position gradient between the two adjacent point traces; and / or calculate the absolute value of the difference between the vertical coordinates of two adjacent point traces within a preset identification range on the side of the vehicle in the vehicle coordinate system, and obtain the longitudinal position gradient between the two adjacent point traces; wherein, the determination module includes: a determination submodule, which is specifically used to determine that the two adjacent point traces are fence point traces if the lateral position gradient and / or longitudinal position gradient between the two adjacent point traces is less than a preset gradient threshold.

[0017] According to a specific implementation method of an embodiment of the present application, the calculation unit includes: an obtaining module, used to rotate the self-vehicle coordinate system around the origin of the self-vehicle coordinate system so that the longitudinal axis direction of the self-vehicle coordinate system is consistent with the longitudinal axis direction of the coordinate system of the first target, thereby obtaining the rotated self-vehicle coordinate system; a second calculation module, used to calculate the distance between the fence point trace and the first target in the rotated self-vehicle coordinate system.

[0018] According to a specific implementation of an embodiment of the present application, the judgment unit includes: a judgment module, which is used to judge the number of fence points within a specified range in front and behind the first target based on the distance between the fence points and the first target.

[0019] According to a specific implementation method of an embodiment of the present application, the second determination unit includes: a statistical module, used to count whether the number of fence points within the specified range in the current frame is greater than a preset threshold. If the number of fence points within the specified range in the current frame is greater than the preset threshold, the first target is determined to be a false target; or count whether the number of fence points within the specified range in N consecutive frames is greater than the preset threshold. If the number of fence points within the specified range in the N consecutive frames is greater than the preset threshold, the first target is determined to be a false target; or if kf>m in N consecutive frames, the first target is determined to be a false target; wherein k is the number of frames in which the number of fence points within the specified range is greater than the preset threshold in the N consecutive frames; f is the number of frames in which the number of fence points within the specified range is 0 in the N consecutive frames; and m is the preset threshold.

[0020] In a third aspect, an embodiment of the present invention further provides an electronic device, comprising: a processor and a memory, the memory being used to store executable program code; the processor runs a program corresponding to the executable program code by reading the executable program code stored in the memory, so as to execute the false target identification method based on the fence area provided in any embodiment of the present invention.

[0021] In a fourth aspect, an embodiment of the present invention further provides a vehicle, on which is installed a false target identification device based on a fence area, and the false target identification device based on a fence area is the false target identification device based on a fence area provided by any embodiment of the present invention.

[0022] The false target recognition method, device, electronic device and vehicle based on the fence area provided by the embodiment of the present invention recognize the points within the preset recognition range of the side of the vehicle, and determine the fence points within the recognition range; wherein the points within the recognition range are detected by the millimeter-wave radar on the vehicle; the distance between the fence points and the first target is calculated; based on the distance between the fence points and the first target, the number of fence points within the specified range around the first target is determined; based on the number of fence points within the specified range, it is determined whether the first target is a false target. In this way, the points within the recognition range are detected by the millimeter-wave radar on the vehicle, and other sensors for external environment perception are not relied on, which reduces the amount of calculation; based on the distance between the fence points and the first target, the number of fence points within the specified range around the first target is determined, and based on the number of fence points within the specified range, it is determined whether the first target is a false target. This process does not require deep learning, and becomes simple in deployment. It is suitable for the recognition of false targets caused by various reasons, and enhances the recognition performance in scenarios with complex reasons for the formation of false targets. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0024] Figure 1 A flow chart of a false target identification method based on a fence area provided by an embodiment of the present invention;

[0025] Figure 2 A schematic diagram of the vehicle side recognition range in an embodiment of the present invention;

[0026] Figure 3 A schematic diagram of a designated range around a first target in an embodiment of the present invention;

[0027] Figure 4 It is a flow chart of fence dot marking in an embodiment of the present invention;

[0028] Figure 5a Figure 5b and are schematic diagrams before and after the vehicle coordinate system is converted in an embodiment of the present invention;

[0029] Figure 6 A schematic diagram of an occlusion scene in an embodiment of the present invention;

[0030] Figure 7 A schematic diagram of another occlusion scene in an embodiment of the present invention;

[0031] Figure 8 This is a flow chart of determining occlusion point traces by vector cross multiplication in an embodiment of the present invention;

[0032] Fig. 9 A detailed flow chart of a false target identification method based on a fence area provided by an embodiment of the present invention;

[0033] Fig.10 A schematic diagram of a structure of a false target identification device based on a fence area provided by an embodiment of the present invention;

[0034] Fig.11 A schematic structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0035] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0036] It should be clear that the described embodiments are only some embodiments of the present invention, not all 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.

[0037] In a first aspect, an embodiment of the present invention provides a false target recognition method based on a fence area, which can effectively enhance the recognition performance in scenarios where the causes of false target formation are complex, simplify deployment, and reduce the amount of computation.

[0038] Figure 1 A flow chart of a false target identification method based on a fence area provided by an embodiment of the present invention, such as Figure 1 As shown, an embodiment of the present invention provides a false target recognition method based on a fence area, comprising:

[0039] S11, identifying points within a preset identification range on the side of the vehicle to determine fence points within the identification range; wherein the points within the identification range are detected by a millimeter wave radar on the vehicle;

[0040] In the embodiment of the present invention, when the millimeter wave radar identifies false targets, it must first determine a certain range and search for them conditionally. Fence false targets mostly appear when the vehicle is moving straight or turning at a small angle. This requires limiting the yaw angular velocity of the vehicle and not detecting false targets when the vehicle turns at a large amplitude. Therefore, according to the limitations of the characteristics and functional requirements of the fence false targets and the steering amplitude of the vehicle, the identification range can be determined. In some examples, the identification range is as follows: Figure 2 As shown, the radial distance for false track judgment can be within 80m, the yaw rate limit of the ego vehicle can be selected to be 0.15rad / s, and the lateral position threshold can be selected to be greater than 1.2m.

[0041] After determining the recognition range, the millimeter-wave radar is used to detect points within the recognition range of the side of the vehicle in real time. The points may include fence points, other vehicle points and false points, etc. Multiple frames of points detected within a certain distance or a certain time period are recorded, and then the static discrete point data is preprocessed to determine the fence points within the recognition range.

[0042] S12, calculating the distance between the fence point trace and the first target;

[0043] When clustered points appear within the recognition range and it is impossible to determine whether they are false points, the clustered points can be abstracted into a target point, and the target point is used as the first target for determination.

[0044] During the judgment process, the distance between the confirmed fence points within the recognition range and the first screened target is first calculated.

[0045] S13, based on the distance between the fence point trace and the first target, determining the number of fence point traces within a specified range around the first target;

[0046] A specified range is determined around the first target, the specified range including a lateral range and a longitudinal range of the front and rear positions of the first target. In some examples, the specified range is as follows: Figure 3 As shown, the lateral range is selected as 0.5 meters, and the longitudinal range is selected as 5 meters front and back. The specified range can be determined by the actual situation of the vehicle.

[0047] Calculate the distance between each fence point and the first target within the recognition range, determine whether the fence point is within the specified range based on the distance, and count the number of fence points within the specified range around the first target.

[0048] S14: Determine whether the first target is a false target based on the number of fence points within the specified range.

[0049] The judgment condition is set according to the actual situation of the vehicle, and whether the first target is a false target is determined according to the number of fence points within a specified range around the first target. If the first target is determined to be a false target, the first target is deleted.

[0050] The false target identification method based on the fence area provided by the embodiment of the present invention identifies the points within the preset identification range of the side of the vehicle, and determines the fence points within the identification range; wherein the points within the identification range are detected by the millimeter wave radar on the vehicle; the distance between the fence points and the first target is calculated; based on the distance between the fence points and the first target, the number of fence points within the specified range around the first target is determined; based on the number of fence points within the specified range, it is determined whether the first target is a false target. In this way, the points within the identification range are detected by the millimeter wave radar on the vehicle, and other sensors for external environment perception are not relied on, which reduces the amount of calculation; based on the distance between the fence points and the first target, the number of fence points within the specified range around the first target is determined, and based on the number of fence points within the specified range, it is determined whether the first target is a false target. This process does not require deep learning, and becomes simple in deployment. It is suitable for the identification of false targets caused by various reasons, and enhances the identification performance in scenarios with complex causes of false target formation, which is conducive to providing the accuracy of false target identification.

[0051] In some examples, identifying points within a preset recognition range on the side of the vehicle and determining fence points within the recognition range includes: calculating the relative position gradient between two adjacent points within the preset recognition range on the side of the vehicle; if the relative position gradient between the two adjacent points is less than a preset gradient threshold, determining that the two adjacent points are fence points.

[0052] In the process of determining the fence points within the identification range, we must first calculate the relative position gradient between two adjacent points within the preset identification range of the side of the vehicle. We need to sort the lateral position and / or longitudinal position of the points detected by the millimeter-wave radar within the identification range of the side of the vehicle, and calculate the position gradient to obtain a gradient set. For the set of lateral gradients and / or longitudinal gradients that meet the conditions, the points that meet the conditions are marked with the fence point flag.

[0053] The gradient is expressed as follows:

[0054] (1)

[0055] in Representing a collection Middle The position gradient of the stationary point traces, the set The points within the preset recognition range of the side of the vehicle. and Respectively represent and The coordinate positions of adjacent points, Representing a collection The number of stationary points in the.

[0056] If the relative position gradient between two adjacent points is less than the preset gradient threshold, the two adjacent points are determined to be fence points. The gradient of the point satisfies formula (2), then point as the starting point of the fence, continue searching the gradient set, and determine the and The adjacent points are fence points, and the restrictions on the gradient are as follows:

[0057] (2)

[0058] in Indicates The preset gradient threshold of the stationary trace.

[0059] In some examples, the preset gradient threshold is determined according to the following formula: ;in, represents the distance coefficient, Indicates The radial distance of each fence point relative to the vehicle, Indicates a fixed distance threshold.

[0060] When the millimeter-wave radar detects points within the identification range of the side of the vehicle, the points closer to the vehicle are distributed more densely, while the points farther away from the vehicle are distributed more discretely, and the gradient threshold must change accordingly. Therefore, the gradient threshold is improved in accuracy through three physical values: distance coefficient, radial distance and distance threshold, and is adaptively determined according to the distance change.

[0061] No. The preset gradient threshold of the stationary traces As shown in formula (3), Represents the distance coefficient, which refers to the distance coefficient between two adjacent points. Indicates The radial distance of the fence point trace relative to the vehicle. The radial distance is the distance from the coordinate position of the nth fence point trace to the origin of the vehicle in the vehicle coordinate system. Indicates a fixed distance threshold value, which refers to the distance threshold between two adjacent traces.

[0062] (3)

[0063] In some examples, the lateral distance coefficient α = 0.01 and the distance threshold β = 0.6, and the longitudinal coefficient α = 0.25 and the distance threshold β = 8.

[0064] In some examples, the calculation of the relative position gradient between two adjacent points within a preset identification range on the side of the vehicle includes: calculating the absolute value of the difference between the horizontal coordinates of two adjacent points within the preset identification range on the side of the vehicle in the vehicle coordinate system, to obtain the lateral position gradient between the two adjacent points; and / or calculating the absolute value of the difference between the vertical coordinates of two adjacent points within the preset identification range on the side of the vehicle in the vehicle coordinate system, to obtain the longitudinal position gradient between the two adjacent points; wherein, if the relative position gradient between two adjacent points is less than a preset gradient threshold, the two adjacent points are determined to be fence points, including: if the lateral position gradient and / or longitudinal position gradient between two adjacent points is less than a preset gradient threshold, the two adjacent points are determined to be fence points.

[0065] After using the millimeter wave radar to detect the points within the recognition range of the side of the vehicle, the lateral position and / or longitudinal position of the points are sorted, and the position gradient is calculated to obtain a gradient set. When calculating the position gradient, the lateral position gradient between two adjacent points can be obtained by calculating the absolute value of the difference between the horizontal coordinates of two adjacent points in the vehicle coordinate system within the preset recognition range of the side of the vehicle, and / or the longitudinal position gradient between two adjacent points can be obtained by calculating the absolute value of the difference between the vertical coordinates of two adjacent points in the vehicle coordinate system within the preset recognition range of the side of the vehicle.

[0066] After obtaining the position gradient, we start to judge and mark the fence point with the points that meet the conditions. Figure 4 Mark the flow chart for the fence traces, such as Figure 4 As shown in FIG. 1 , if the lateral position gradient and / or the longitudinal position gradient between two adjacent point traces is less than a preset gradient threshold, then the first and For example, suppose the first two adjacent points are fence points. and The coordinates of two adjacent points are (3, 2) and (3, 3), the horizontal distance coefficient α = 0.01 and the distance threshold β = 0.6, the vertical coefficient α = 0.25, and the distance threshold β = 8. According to formula (3), the horizontal gradient threshold can be obtained =0.01× 0.64, vertical gradient threshold + 8 = 8.90, the lateral position gradient between two adjacent points is , the longitudinal position gradient between two adjacent points is 1, then the lateral position gradient and / or longitudinal position gradient between two adjacent point traces is less than the lateral and / or longitudinal gradient threshold between two adjacent point traces, and the first and Two adjacent points are fence points.

[0067] The existing scheme simply searches for the front and rear traces without considering that the traces at a long distance are more scattered, which may result in fence trace recognition errors. In the embodiment of the present invention, in view of the fact that the traces closer to the vehicle are densely distributed and the traces farther from the vehicle are more scattered, the gradient threshold should also change accordingly. Since in formula (3) of this scheme, The gradient threshold of each point trace is determined according to the nth point trace. The coordinates of the point trace take the vehicle as the origin of the coordinate system. The farther the point is from the vehicle, the greater its radial distance relative to the vehicle and the greater the gradient threshold, so that this scheme conforms to the actual situation.

[0068] In some examples, calculating the distance between the fence point trace and the first target includes: rotating the self-vehicle coordinate system around the origin of the self-vehicle coordinate system so that the longitudinal axis direction of the self-vehicle coordinate system is consistent with the longitudinal axis direction of the coordinate system of the first target, thereby obtaining a rotated self-vehicle coordinate system; and calculating the distance between the fence point trace and the first target in the rotated self-vehicle coordinate system.

[0069] In the process of calculating the distance between the fence point trace and the first target, the millimeter wave radar detects the points within the recognition range of the side of the vehicle using the coordinate system established by the vehicle. The coordinate system takes the vehicle as the origin, the direction of movement as the x-axis, and the direction perpendicular to the direction of movement as the y-axis. Since the first target and the vehicle may have different orientations, it is more complicated to calculate the distance between the fence point trace and the first target. Therefore, converting the vehicle coordinate system is helpful to calculate the positional relationship between the fence point trace and the first target. When the first target and the vehicle have different orientations, the angle between the first target's direction of movement and the vehicle's direction of movement is , the vehicle coordinate system can be rotated around the origin of the vehicle coordinate system Angle, so that the vertical axis direction of the self-vehicle coordinate system is consistent with the vertical axis direction of the coordinate system of the first target, to obtain the rotated self-vehicle coordinate system. For example, Figure 5a and Figure 5b This is a schematic diagram before and after the vehicle coordinate system is transformed. Figure 5a and Figure 5b As shown, with the vehicle as the origin, the rotation Angle, with the first target movement direction as the x-axis, and the direction perpendicular to the first target movement direction as the y-axis. Before and after the vehicle coordinate system is transformed, the first target coordinate transformation is as shown in formulas (4) and (5).

[0070] (4) (5)

[0071] in , represents the lateral and longitudinal distances of the first target in the first target coordinate system, , Represents the lateral and longitudinal distances of the first target in the ego-vehicle coordinate system; is the heading angle of the first target (the angle between the moving direction of the first target and the moving direction of the vehicle), in the coordinate system of the vehicle The axis is 0, left rotation is positive, right rotation is negative, and the value range is ( - π, π] In the rotated ego-vehicle coordinate system, calculate the distance between the fence point trace and the first target.

[0072] In some examples, based on the distance between the fence point trace and the first target, the number of fence point traces within a specified range around the first target is determined, including: based on the distance between the fence point trace and the first target, the number of fence point traces within a specified range in front and behind the first target is determined.

[0073] According to existing statistics, the most significant feature of the false target track is that there are many stationary points within the range close to the false target. Therefore, when dealing with such problems, a method for the position relationship between the stationary point and the first target can be designed to determine whether the first target is a false target. Therefore, after determining the recognition range, the distance between each fence point within the recognition range and the first target can be calculated, and each fence point within all recognition ranges can be looped, such as Figure 3 As shown, the number of fence points falling within the specified range in front of and behind the first target is counted.

[0074] In some examples, determining whether the first target is a false target based on the number of fence points within the specified range includes: counting the number of fence points within the specified range in the current frame to see whether it is greater than a preset threshold, and if the number of fence points within the specified range in the current frame is greater than the preset threshold, determining that the first target is a false target; or counting the number of fence points within the specified range in N consecutive frames to see whether they are all greater than the preset threshold, and if the number of fence points within the specified range in N consecutive frames is all greater than the preset threshold, determining that the first target is a false target; or if kf>m in N consecutive frames, determining that the first target is a false target; wherein k is the number of frames in which the number of fence points within the specified range is greater than the preset threshold in the N consecutive frames; f is the number of frames in which the number of fence points within the specified range is 0 in the N consecutive frames; and m is the preset threshold.

[0075] According to the number of fence points falling in the specified range in front of and behind the first target, it is judged whether the first target is a false target. The judgment conditions are as follows:

[0076] In the current frame, the number of fence points falling within the specified range in front of and behind the first target is counted to see whether it is greater than a preset threshold. The preset threshold includes the number of frames and the number of fence points. In some examples, the minimum number of fence points that meets the fence false target suppression condition is 3, and the minimum number of frames that meets the fence false target suppression condition is 3 frames. The preset threshold can be adjusted according to actual conditions. If the number of fence points falling within the specified range in front of and behind the first target in the current frame is greater than the preset threshold, the first target is determined to be a false target.

[0077] Alternatively, count the number of fence points that fall within a specified range in front of and behind the first target in N consecutive frames to see whether they are all greater than a preset threshold. If the number of fence points that fall within a specified range in front of and behind the first target in N consecutive frames is greater than the preset threshold, then determine that the first target is a false target.

[0078] The number of fence dots falling in the specified range in front of and behind the first target in N consecutive frames is counted. There is another way to judge that if the number of fence dots in the specified range in N consecutive frames is greater than the number of frames with a preset threshold minus the number of frames with the number of fence dots in the specified range being 0, which is greater than the preset threshold, the first target is determined to be a false target. For example, let k be the number of frames in which the number of fence dots in the specified range is greater than the preset threshold in the said consecutive N frames; f is the number of frames in which the number of fence dots in the specified range is 0 in the consecutive N frames; m is the preset threshold; if kf>m, the first target is determined to be a false target. False targets are deleted after accumulating multiple frames, which reduces the probability of mistaken deletion.

[0079] For example, in practice, the fence point trace selects stationary points with a radial distance less than 80 meters, and the lateral range is set within 25 meters. The fixed distance thresholds are selected as 0.01 and 0.6 based on empirical values, respectively. The longitudinal distance coefficient The fixed distance thresholds are selected as 0.25 and 8 respectively based on empirical values. When counting fence points, the horizontal range is 0.5 meters, and the vertical range is 5 meters in front and behind the car. The minimum number of fence tracks falling within the front and rear range of the track is set to 3, and the threshold of the cumulative number of frames is 3 frames. Since the vehicle is driving in the center of the road, the moving point track forms an unconfirmed track at the fence. After searching for the points in front and behind the false track for three consecutive frames to meet the requirements, the track is deleted. Under a certain curvature, the fence point track can be successfully marked, which is compatible with scenes of straight driving and large curvature.

[0080] In some examples, after determining whether the first target is a false target based on the number of fence points within the specified range, the method further includes: determining the points obscured by the second target detected by the millimeter wave radar; and suppressing the output of the points obscured by the second target.

[0081] After determining whether the first target is a false target based on the number of fence points within the specified range, the points generated by the occlusion at the fence must be suppressed. In actual operation, the points detected by the millimeter-wave radar include the real vehicle and the fence or false points near the fence that appear on the side of the real vehicle but cannot be observed by the vehicle camera. It is necessary to first determine the occluded points of the real vehicle. In the case of uncertainty, the real vehicle can be used as the second target, and the preset conditions can be used to determine the points detected by the millimeter-wave radar.

[0082] Figure 6 is a schematic diagram of an occlusion scene, such as Figure 6 As shown in the figure, since the point traces blocked by the second target are far away from the vehicle in other lanes, even if there are safety hazards, they are a greater threat to the second target, so they can be directly eliminated, suppressing the output of the point traces blocked by the second target and not used as the starting track.

[0083] Figure 7 This is a schematic diagram of another occlusion scenario. The point trace is located near the fence. It is impossible to determine whether the point trace is a fence point trace or a false point trace near the fence. The point trace can be set as the target point trace P, and the position of the point trace can be determined by vector intersection to determine the point trace occluded by the second target detected by the millimeter-wave radar.

[0084] In the process of solving the vector, the physical model of the occluding scene should be abstracted into a mathematical model for calculation. First, the second target is abstracted into a rectangle, and the length and width of the rectangle are expressed as , express,

[0085] The ego-vehicle coordinate system takes the ego-vehicle as the origin, the moving direction is the x-axis, the direction perpendicular to the moving direction is the y-axis, and H is the angle between the moving direction of the second target and the x-axis of the ego-vehicle coordinate system. Figure 8 FIG. 1 is a flow chart of determining occlusion points by vector cross multiplication in an embodiment of the present invention. Figure 8 As shown, it is necessary to find the maximum and minimum angles of the rectangular frame abstracted by the millimeter-wave radar in the second target, and obtain the line connecting the two points with the maximum angle. For example, Figure 7 As shown in the figure, it is assumed that the maximum and minimum angles of the rectangular frame abstracted by the millimeter-wave radar at the second target are obtained, A and B are the two points with the maximum angle, and the occlusion point trace P and the line connecting A and B form a vector , .

[0086] Before finding the vector, you need to find the maximum angle. You can find the four vertices of the rectangle first, and get the angle value based on the distance between the vertex and the occlusion point trace in combination with the trigonometric function. Then find the maximum angle based on the obtained angle value. Since the second target and the vehicle may have different orientations, it is more complicated to calculate the distance between the fence point trace and the second target vertex and the occlusion point trace. You can convert the coordinate system as in the previous embodiment. Take the reference point (the vertex in the lower left corner of the track frame) as an example:

[0087] (6) (7) = (8) = (9)

[0088] in, TargetX and TargetY is the geometric position, RefP[3][0] 、 RefP[3][1] RefP[5][ 0] and RefP[5][1] Array coordinates. , The length and width of the abstract rectangle of the second target. , , are the sine and cosine values ​​of the target azimuth (the angle between the second target's motion direction and the x-axis of the vehicle's coordinate system), respectively. , is the coordinate of one vertex of the rectangle, and the steps for obtaining the coordinates of other vertices are the same as those for this vertex. The angle is calculated using formula (10):

[0089] (10)

[0090] in, is the opening angle. Similarly, calculate the maximum opening angle MAXAngle and the minimum opening angle MINAngle of the four vertices.

[0091] According to the obtained fixed point position ( , ), the line connecting the two points with the largest angle forms a vector , , , after vector cross multiplication , if the cross product result is less than 0, then in the vector On the outside, the target point track P can be determined to be an occluded point track. For the point track that meets the occlusion, the track output processing is not performed, and the output of the point track occluded by the second target is suppressed, which can solve the track near the fence that is not deleted by the fence point track strategy. In this embodiment of the present invention, the occluded point track is determined by vector cross multiplication. Various false point tracks caused by causes such as angle measurement errors and multipaths that appear at the fence can be detected.

[0092] The false target recognition method based on the fence area provided by the embodiment of the present invention is described in detail below through a specific embodiment.

[0093] like Fig. 9 As shown, the false target identification method based on the fence area provided by the embodiment of the present invention includes the following steps:

[0094] S201, identifying points within a preset identification range on the side of the vehicle, and determining fence points within the identification range; wherein the points within the identification range are detected by a millimeter wave radar on the vehicle;

[0095] S202, calculating the distance between the fence point trace and the first target;

[0096] S203, based on the distance between the fence point trace and the first target, determining the number of fence point traces within a specified range around the first target;

[0097] S204, determining whether the first target is a false target based on the number of fence points within the specified range;

[0098] S205, determining a point trace blocked by the second target detected by the millimeter wave radar;

[0099] S206: Suppress the output of the point trace blocked by the second target.

[0100] In a second aspect, an embodiment of the present invention provides a false target recognition device based on a fence area, which can effectively enhance the recognition performance in scenarios where the causes of false target formation are complex, simplify deployment, and reduce the amount of computation.

[0101] like Fig.10 As shown, an embodiment of the present invention further provides a false target identification device based on a fence area, comprising:

[0102] The first determination unit 31 is used to identify points within a preset identification range on the side of the vehicle, and determine fence points within the identification range; wherein the points within the identification range are detected by a millimeter wave radar on the vehicle;

[0103] A calculation unit 32, used for calculating the distance between the fence point trace and the first target;

[0104] A judging unit 33, configured to judge the number of fence points within a specified range around the first target based on the distance between the fence points and the first target;

[0105] The second determination unit 34 is configured to determine whether the first target is a false target based on the number of fence points within the specified range.

[0106] The false target identification device based on the fence area provided by the embodiment of the present invention identifies the points within the preset identification range of the side of the vehicle, and determines the fence points within the identification range; wherein the points within the identification range are detected by the millimeter-wave radar on the vehicle; the distance between the fence points and the first target is calculated; based on the distance between the fence points and the first target, the number of fence points within the specified range around the first target is determined; based on the number of fence points within the specified range, it is determined whether the first target is a false target. In this way, the points within the identification range are detected by the millimeter-wave radar on the vehicle, and other sensors for external environment perception are not relied on, thereby reducing the amount of calculation; based on the distance between the fence points and the first target, the number of fence points within the specified range around the first target is determined, and based on the number of fence points within the specified range, it is determined whether the first target is a false target. This process does not require deep learning, and becomes simple in deployment. It is suitable for the identification of false targets caused by various reasons, and enhances the identification performance in scenarios with complex reasons for the formation of false targets.

[0107] According to a specific implementation method of an embodiment of the present application, the first determination unit includes: a first calculation module, used to calculate the relative position gradient between two adjacent point traces within a preset identification range of the side of the vehicle; a determination module, used to determine that the two adjacent point traces are fence point traces if the relative position gradient between the two adjacent point traces is less than a preset gradient threshold.

[0108] According to a specific implementation method of an embodiment of the present application, the first calculation module includes: a calculation submodule, which is specifically used to calculate the absolute value of the difference between the horizontal coordinates of two adjacent point traces within a preset identification range on the side of the vehicle in the vehicle coordinate system, and obtain the lateral position gradient between the two adjacent point traces; and / or calculate the absolute value of the difference between the vertical coordinates of two adjacent point traces within a preset identification range on the side of the vehicle in the vehicle coordinate system, and obtain the longitudinal position gradient between the two adjacent point traces; wherein, the determination module includes: a determination submodule, which is specifically used to determine that the two adjacent point traces are fence point traces if the lateral position gradient and / or longitudinal position gradient between the two adjacent point traces is less than a preset gradient threshold.

[0109] According to a specific implementation method of an embodiment of the present application, the calculation unit includes: an obtaining module, used to rotate the self-vehicle coordinate system around the origin of the self-vehicle coordinate system so that the longitudinal axis direction of the self-vehicle coordinate system is consistent with the longitudinal axis direction of the coordinate system of the first target, thereby obtaining the rotated self-vehicle coordinate system; a second calculation module, used to calculate the distance between the fence point trace and the first target in the rotated self-vehicle coordinate system.

[0110] According to a specific implementation method of an embodiment of the present application, the judgment unit includes: a judgment module, used to judge the number of fence points within a specified range in front and behind the first target based on the distance between the fence points and the first target.

[0111] According to a specific implementation method of an embodiment of the present application, the second determination unit includes: a statistical module, used to count whether the number of fence points within the specified range in the current frame is greater than a preset threshold. If the number of fence points within the specified range in the current frame is greater than the preset threshold, the first target is determined to be a false target; or count whether the number of fence points within the specified range in N consecutive frames is greater than the preset threshold. If the number of fence points within the specified range in the N consecutive frames is greater than the preset threshold, the first target is determined to be a false target; or if kf>m in N consecutive frames, the first target is determined to be a false target; wherein k is the number of frames in which the number of fence points within the specified range is greater than the preset threshold in the N consecutive frames; f is the number of frames in which the number of fence points within the specified range is 0 in the N consecutive frames; and m is the preset threshold.

[0112] In a third aspect, an embodiment of the present invention further provides an electronic device that can effectively enhance the recognition performance, simplify deployment and reduce the amount of computation in scenarios where the causes of false target formation are complex.

[0113] like Fig.11 As shown, the electronic device provided by an embodiment of the present invention may include: a processor 51 and a memory 52, the memory 52 is used to store executable program code; the processor 51 runs the program corresponding to the executable program code by reading the executable program code stored in the memory 52, so as to execute the false target identification method based on the fence area provided by any of the aforementioned embodiments.

[0114] The specific execution process of the above steps by the processor 52 and the steps further executed by the processor 52 by running the executable program code can be found in the description of the previous embodiment, which will not be repeated here.

[0115] In a fourth aspect, an embodiment of the present invention further provides a vehicle, on which is installed a false target identification device based on a fence area, and the false target identification device based on a fence area is any one of the false target identification devices based on a fence area provided in the aforementioned embodiments, and therefore can also achieve the corresponding technical effects, which have been described in detail in the previous text and will not be repeated here.

[0116] 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.

[0117] 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.

[0118] 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.

[0119] 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.

[0120] 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.

[0121] The above is only a specific embodiment 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 false target recognition method based on a fence area, characterized in that: include: Identify points within a preset identification range on the side of the vehicle to determine fence points within the identification range; wherein the points within the identification range are detected by a millimeter-wave radar on the vehicle; Calculating the distance between the fence point trace and the first target; Based on the distance between the fence point trace and the first target, determining the number of fence point traces within a specified range around the first target; Based on the number of fence points within the specified range, it is determined whether the first target is a false target.

2. The false target identification method based on the fence area according to claim 1 is characterized in that: The step of identifying points within a preset identification range of the side of the vehicle and determining fence points within the identification range includes: Calculate the relative position gradient between two adjacent points within the preset recognition range of the side of the vehicle; If the relative position gradient between the two adjacent point traces is less than a preset gradient threshold, the two adjacent point traces are determined to be fence point traces.

3. The false target identification method based on the fence area according to claim 2 is characterized in that: The preset gradient threshold is determined according to the following formula: ； in, represents the distance coefficient, Indicates The radial distance of each fence point relative to the vehicle, Indicates a fixed distance threshold.

4. The false target identification method based on the fence area according to claim 2 is characterized in that: The calculation of the relative position gradient between two adjacent points within the preset recognition range of the side of the vehicle includes: Calculate the absolute value of the difference between the horizontal coordinates of two adjacent points within a preset recognition range of the side of the vehicle in the vehicle coordinate system to obtain a lateral position gradient between the two adjacent points; and / or Calculate the absolute value of the difference between the longitudinal coordinates of two adjacent point traces in the vehicle coordinate system within the preset recognition range of the side of the vehicle to obtain the longitudinal position gradient between the two adjacent point traces; Wherein, if the relative position gradient between the two adjacent point traces is less than a preset gradient threshold, determining that the two adjacent point traces are fence point traces includes: If the lateral position gradient and / or the longitudinal position gradient between the two adjacent point traces is less than a preset gradient threshold, the two adjacent point traces are determined to be fence point traces.

5. The false target identification method based on the fence area according to claim 1 is characterized in that: The calculating the distance between the fence point trace and the first target includes: Rotating the ego-vehicle coordinate system around the origin of the ego-vehicle coordinate system so that the direction of the longitudinal axis of the ego-vehicle coordinate system is consistent with the direction of the longitudinal axis of the coordinate system of the first target, thereby obtaining a rotated ego-vehicle coordinate system; In the rotated vehicle coordinate system, the distance between the fence point trace and the first target is calculated.

6. The false target identification method based on the fence area according to claim 1 is characterized in that: The determining, based on the distance between the fence point trace and the first target, the number of fence point traces within a specified range around the first target includes: Based on the distance between the fence point trace and the first target, the number of fence point traces within a specified range in front of and behind the first target is determined.

7. The false target identification method based on the fence area according to claim 1 is characterized in that: The determining whether the first target is a false target based on the number of fence points within the specified range includes: Counting the number of fence points within the specified range in the current frame to determine whether it is greater than a preset threshold, and if the number of fence points within the specified range in the current frame is greater than the preset threshold, determining that the first target is a false target; or Counting whether the number of fence points within the specified range in N consecutive frames is greater than a preset threshold, and if the number of fence points within the specified range in the N consecutive frames is greater than the preset threshold, determining that the first target is a false target; or If kf>m within N consecutive frames, it is determined that the first target is a false target; Wherein, k is the number of frames in which the number of fence points within the specified range is greater than a preset threshold within the continuous N frames; f is the number of frames in which the number of fence points within the specified range is 0 within the continuous N frames; m is the preset threshold.

8. The false target identification method based on the fence area according to claim 1 is characterized in that: After determining whether the first target is a false target based on the number of fence points within the specified range, the method further includes: Determining a point trace obscured by a second target detected by the millimeter-wave radar; The output of the point traces blocked by the second target is suppressed.

9. A false target identification device based on a fence area, characterized in that: include: A first determination unit is used to identify points within a preset identification range on the side of the vehicle, and determine fence points within the identification range; wherein the points within the identification range are detected by a millimeter wave radar on the vehicle; A calculation unit, used for calculating the distance between the fence point trace and the first target; A judging unit, configured to judge the number of fence points within a specified range around the first target based on the distance between the fence points and the first target; The second determination unit is used to determine whether the first target is a false target based on the number of fence points within the specified range.

10. An electronic device, characterized in that: The electronic device includes: a processor and a memory, the memory is used to store executable program code; the processor runs a program corresponding to the executable program code by reading the executable program code stored in the memory, and is used to execute the false target identification method based on the fence area described in any one of claims 1 to 8.

11. A vehicle, characterized in that: The vehicle is equipped with a false target recognition device based on a fence area, and the false target recognition device based on a fence area is the false target recognition device based on a fence area as claimed in claim 9.

Images