Target recognition method and device, driving assistance system and vehicle

Abstract

Embodiments of the present application disclose a target identification method and device, a driving assistance system and a vehicle. Position information, existence period and motion state information of each detection object detected by a radar on the vehicle are acquired, and a target object is determined according to the position information, existence period and speed information of each detection object. Thus, the embodiments determine the target object by comprehensively considering various factors such as position, speed and existence period, so as to improve target identification accuracy.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and more specifically, to a target recognition method, apparatus, driving assistance system, and vehicle. Background Technology

[0002] With the continuous development of intelligent transportation, radar technology is being increasingly widely used in advanced driver assistance systems (ADAS) for vehicles, playing an indispensable role in autonomous driving systems. In existing technologies, radar detects targets in the vehicle's environment by emitting radar signals and receiving radar feedback signals.

[0003] In ADAS systems, radar, especially millimeter-wave radar, is used for ranging and speed measurement, which has significant advantages over cameras. However, this also brings some limitations, such as the detection of mirrored targets, which typically occurs when the vehicle approaches a median strip. Because mirrored targets are false and there is no effective detection signal to update the tracking, the vehicle may drift into its own lane during subsequent target tracking, causing false braking. Summary of the Invention

[0004] In view of this, embodiments of the present invention provide a target recognition method, device, driving assistance system, and vehicle to determine target objects by comprehensively considering various factors such as position, speed, and period of existence, thereby improving the accuracy of target recognition.

[0005] In a first aspect, embodiments of the present invention provide a target recognition method, the method comprising:

[0006] The detection results of the radar on the vehicle are obtained, including the location information, existence period and motion state information of the detected object;

[0007] The target object is determined based on the location information, existence period, and motion state information of each of the detected objects.

[0008] In a second aspect, embodiments of the present invention provide a driving assistance system, the driving assistance system comprising:

[0009] Radar is used to detect environmental information and obtain detection results, which include the location information, existence period and motion state information of the detected object;

[0010] The target recognition device is used to acquire the detection results of the radar on the vehicle and determine the target object based on the position information, existence period and motion state information of each detected object.

[0011] Thirdly, embodiments of the present invention provide a vehicle, the vehicle comprising:

[0012] The vehicle itself; and

[0013] The driving assistance system as described in the second aspect of the embodiments of the present invention.

[0014] Fourthly, embodiments of the present invention provide a target recognition device, the target recognition device comprising:

[0015] The information acquisition unit is configured to acquire the detection results of the radar on the vehicle, the detection results including the location information, existence period and motion state information of the detected object;

[0016] The target recognition unit is configured to determine the target object based on the position information, existence period and motion state information of each of the detected objects.

[0017] Fifthly, embodiments of the present invention provide a computer-readable storage medium having stored thereon computer program instructions, which, when executed by a processor, implement the method described in the first aspect of the present invention.

[0018] This invention, in its embodiments, acquires the position information, existence period, and motion state information of each candidate object detected by radar on a vehicle, and determines the target object based on the position information, existence period, and speed information of each candidate object. Therefore, this embodiment can determine the target object by comprehensively considering various factors such as position, speed, and existence period, thereby improving the accuracy of target recognition. Attached Figure Description

[0019] The above and other objects, features and advantages of the present invention will become clearer from the following description of embodiments of the invention with reference to the accompanying drawings, in which:

[0020] Figure 1 This is a schematic diagram of the target detection results of related technologies;

[0021] Figure 2 This is a flowchart of a target recognition method according to an embodiment of the present invention;

[0022] Figure 3 This is a flowchart of the target object determination method according to an embodiment of the present invention;

[0023] Figure 4 This is a flowchart of another target recognition method according to an embodiment of the present invention;

[0024] Figure 5 This is a schematic diagram of the target recognition process according to an embodiment of the present invention;

[0025] Figure 6 This is a schematic diagram of a driving assistance system according to an embodiment of the present invention;

[0026] Figure 7This is a schematic diagram of a target recognition device according to an embodiment of the present invention. Detailed Implementation

[0027] The present invention is described below based on embodiments, but the invention is not limited to these embodiments. In the detailed description of the invention below, certain specific details are described in detail. Those skilled in the art will fully understand the invention even without these details. To avoid obscuring the essence of the invention, well-known methods, processes, flows, elements, and circuits are not described in detail.

[0028] Furthermore, those skilled in the art should understand that the accompanying drawings provided herein are for illustrative purposes only and are not necessarily drawn to scale.

[0029] Unless the context explicitly requires it, words such as "including" or "contains" in the instruction manual should be interpreted as including rather than exclusive or exhaustive; that is, meaning "including but not limited to".

[0030] In the description of this invention, it should be understood that the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance. Furthermore, in the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0031] In related technologies, the presence of other equipment (e.g., vehicles) at similar distances and speeds to the vehicle or other objects (e.g., median barriers) along a symmetry axis is typically detected to determine the mirror target. Figure 1 As shown, the center line a represents the median strip, and the dashed line b represents the lane lines. After comparing distance and speed information, two symmetrical sets of vehicles were obtained: one set includes vehicle 11 and vehicle 11', and the other set includes vehicle 12 and vehicle 12'. Vehicles 11 and 11' are symmetrical relative to the median strip, and their speeds and other information are basically the same, indicating that vehicle 11' is likely a mirror image of the real vehicle 11. However, vehicles 12 and 12' are clearly real vehicles. Therefore, if the mirror image is determined solely based on symmetry, relative distance, and speed information, real target vehicles may be mistakenly deleted. When a real target vehicle enters the lane, the target will be lost, potentially causing an accident. Therefore, this embodiment provides a target recognition method that comprehensively considers various factors such as position, speed, and duration of existence to determine the target object, thereby improving target recognition accuracy, avoiding the mistaken deletion of targets, and improving driving safety.

[0032] Figure 2 This is a flowchart of a target recognition method according to an embodiment of the present invention. Figure 2As shown, the target recognition method of this invention includes the following steps:

[0033] Step S110: Obtain the detection results from the radar on the vehicle. The detection results include the location information, existence period, and motion state information of the detected object.

[0034] In this embodiment, the vehicle is equipped with radar to detect the environment around the vehicle and obtain environmental information (i.e., detection results). This environmental information guides the vehicle's driving, improving driving safety. Optionally, the environmental information may include the location, periodicity, and motion status information of surrounding vehicles, median strips, fences, road signs, lane lines, or other obstacles.

[0035] Alternatively, the vehicle may be equipped with four radars located at the upper left, lower left, upper right, and lower right corners, respectively, to obtain information about the surrounding environment of the vehicle by reflecting information after the radars are activated.

[0036] In one optional implementation, the radar type involved in this embodiment may include, but is not limited to, millimeter-wave radar. Millimeter-wave radar is a detection radar that operates in the millimeter-wave band. Compared with other infrared, laser, and other radars, millimeter-wave radar has higher penetration and higher spatial resolution, can identify multiple small targets, and has all-weather, all-time characteristics.

[0037] Alternatively, this embodiment may perform target detection using one-dimensional or multi-dimensional Fast Fourier Transform (FFT) processing, Constant False Alarm Rate (CFAR) detection processing, angle measurement processing, target tracking processing, or filtering processing. It should be understood that this embodiment does not limit the processing method used for radar-based target detection; it only requires that the target be detected based on radar reflection signals.

[0038] In one optional implementation, the position information of the detected object can be its relative position to the vehicle. For example, a coordinate system can be established with the vehicle as the center, and the coordinates of each detected object in this coordinate system can be determined based on the radar's reflected signals. Further optionally, a coordinate system can be established with the vehicle as the center point, the vehicle's driving direction (i.e., the direction the vehicle is facing) as the x-axis, and the vehicle's lateral direction (i.e., the direction perpendicular to the vehicle's facing) as the y-axis. Optionally, the position information of the detected object can be its relative position to the vehicle, or it can include the distance between the detected object and the vehicle in the x-axis direction and the distance in the y-axis direction; this embodiment does not limit this.

[0039] Optionally, in this embodiment, the position information of the detected object may further include information on the change in the relative position of the candidate position with respect to the vehicle. In actual driving scenarios, due to the instability of the mirror target after its establishment, there may be jumps, which can affect the detection of the relative position between the mirror target and the vehicle. Therefore, the relative position change information can be used as a criterion for determining whether the detected object is a mirror target. Optionally, this embodiment may use the variance of the distance change between the detected object and the vehicle on the x-axis and / or the variance of the distance change on the y-axis to characterize the above-mentioned relative position change information. It should be understood that this embodiment is not limited to variance, and may also use parameters that can characterize relative position changes, such as standard deviation.

[0040] In this embodiment, the existence period is used to characterize the duration of the detected object's existence. In real-world scenarios, the formation of a mirror target requires certain conditions, such as radar coverage, reflecting objects, and the angle of the actual object; therefore, the existence period of a mirror target is usually short. Thus, this embodiment uses the existence period of the detected object as a criterion for judging mirror targets, which can improve both the accuracy and efficiency of target recognition.

[0041] In one optional implementation, the motion state information can characterize whether the detected object is in motion or stationary, and may further include the speed of the detected object when it is in motion. It should be understood that if the detected object is stationary during travel, it generally does not pose a danger and the occurrence is brief. Therefore, this embodiment primarily uses a mirror image of a moving target, whose speed is essentially the same as the corresponding actual target. Optionally, in this embodiment, the speed of the detected object relative to the vehicle is used as the speed information.

[0042] Step S120: Determine the target object based on the position information, existence period, and motion state information of each detected object. The target object is also the mirror image of the aforementioned real object.

[0043] Figure 3 This is a flowchart of a target object determination method according to an embodiment of the present invention. In one optional implementation, such as... Figure 3 As shown, the target determination method in this embodiment includes the following steps:

[0044] Step S121: Determine at least one candidate object based on the location information, existence period and motion state information of each detected object in the detection results.

[0045] In one optional implementation, since the existence of a period, the relative position to the vehicle, and the motion state of the detected object can all affect whether a corresponding mirror target exists, this embodiment can perform preliminary screening of each detected object based on the existence of a period, the relative position to the vehicle, and the motion state of the detected object. That is, it can screen out candidate objects that may have mirror targets, and then detect objects that are symmetrical to these candidate objects, thereby determining the mirror target among the detected objects with symmetrical relationships. Thus, this embodiment can improve the accuracy and efficiency of mirror target recognition.

[0046] Optionally, since objects near reflective surfaces such as median strips are more likely to produce mirror targets, this embodiment can first determine the first distance between each object and reflective objects such as median strips, and then filter out objects whose first distance is less than a first threshold (e.g., 5m) as initial screening objects. These initial screening objects are then further filtered based on their existence period, relative position to the vehicle, and motion state to obtain candidate objects. This can further improve target recognition efficiency.

[0047] In one optional implementation, this embodiment determines the motion state reference value, position reference value, and existence period reference value of each preliminary screening object based on its motion state information, position information, and existence period. A weighted sum of these reference values ​​is calculated based on a first weighted reassembly to obtain a candidate reference value for each preliminary screening object. If a candidate reference value is greater than or equal to a first reference threshold, the corresponding preliminary screening object is identified as a candidate object. Optionally, the reference threshold can be determined based on testing in a specific application scenario. For example, the first reference threshold can be 0.7.

[0048] In this embodiment, the candidate reference value Metric1 is calculated as follows:

[0049] Metric1=a*x1+b*x2+c*x3 (1)

[0050] Where x1, x2, and x3 represent the motion state reference value, position reference value, and existence period reference value, respectively, and a, b, and c are the weights of the motion state reference value, position reference value, and existence period reference value, respectively. Optionally, a + b + c = 1.

[0051] Optionally, in this embodiment, the weighted recombination [a,b,c] can be predetermined according to the specific application scenario. Further, optionally, the weighted recombination [a,b,c] can be predetermined using exhaustive methods or similar techniques according to the specific application scenario. It should be understood that this embodiment does not limit the method for determining the weighted recombination [a,b,c].

[0052] Optionally, the motion state reference value x1 of the initial screening object is determined based on whether the initial screening object is moving. Further, optionally, if the initial screening object is in motion, the motion state reference value x1 can be 1; if the initial screening object is stationary, the motion state reference value x1 can be 0 (or a negative value). It should be understood that the reference values ​​regarding whether the initial screening object is in motion or stationary in this embodiment are merely exemplary. Other values ​​that tend to be selected when in motion and screened out when stationary, thus affecting the screening results, can also be applied to this embodiment.

[0053] Optionally, the position reference value x2 of the initially screened object is determined based on the relative position of the initially screened object and the vehicle. Optionally, in this embodiment, it is determined based on the lateral distance (perpendicular to the direction of travel) between the initially screened object and the vehicle. Further optionally, in response to the lateral distance between the initially screened object and the vehicle being less than or equal to a first distance threshold, the position reference value x2 of the initially screened object can be 1; in response to the lateral distance between the initially screened object and the vehicle being greater than the first distance threshold, the position reference value x2 of the initially screened object can be 0 (or a negative value). In other optional implementations, the optimal distance can be determined through testing. When the lateral distance between the initially screened object and the vehicle is the optimal distance, mirror targets are most likely to be generated, or the generated mirror targets are more likely to affect the driver's driving. Therefore, when the lateral distance between the initially screened object and the vehicle is the optimal distance, the position reference value is set to the maximum value; the further the lateral distance between the initially screened object and the vehicle is from the optimal distance, the smaller the corresponding position reference value. In other words, in this embodiment, a position reference value can be preset based on a first distance threshold, or a function between the position reference value and the lateral distance can be predetermined, and the corresponding position reference value can be determined based on this function and the lateral distance between the initially screened object and the vehicle.

[0054] Optionally, the existence period reference value x3 of the initially screened object is determined based on the duration of the existence period of the initially screened object and a duration threshold. Further, optionally, in this embodiment, if the duration of the existence period of the initially screened object is greater than or equal to the duration threshold, the existence period reference value of the initially screened object can be 1; if the duration of the existence period of the initially screened object is less than the duration threshold, the existence period reference value of the initially screened object can be 0 (or a negative value). It should be understood that the existence period reference value in this embodiment is merely exemplary; other assignments that tend to select objects when the existence period is greater than the duration threshold and tend to screen them when the existence period is less than the duration threshold, thus affecting the screening results, can all be applied to this embodiment. In other optional implementations, a function for the existence period reference value and the existence period can be pre-created to determine the existence period reference value of the initially screened object based on this function and the existence period of the initially screened object. Optionally, the existence period reference value can increase as the duration of the existence period increases.

[0055] For example, suppose the weighted recombination [a,b,c] is [0.35,0.3,0.35]. If radar detection determines that the initially screened object is currently in motion, its lateral distance from the vehicle is greater than a first distance threshold, and the duration of its existence period is greater than a duration threshold, then the motion state reference value x1 = 1, the position reference value x2 = 1, and the existence period reference value x3 = 1 for the initially screened object. The candidate reference value Metric1 for the initially screened object is 0.35*1 + 0.3*1 + 0.35*1 = 1 > 0.7. Therefore, the initially screened object can be identified as a candidate object.

[0056] Step S122: Detect the detection object that is symmetrical to each candidate object to determine the target object.

[0057] In one optional implementation, for each candidate object, this embodiment determines the velocity difference, position difference, and position change difference between the candidate object and each detected object. In response to the velocity difference, position difference, and position change difference satisfying predetermined conditions, it determines that the corresponding detected object is symmetrical to the candidate object. That is, in this embodiment, a global search is performed on the acquired candidate object among all detected objects (or detected objects in all motion states) to obtain the symmetrical detected object. If no symmetrical detected object is detected, then the candidate object does not have a mirror target and can be deleted from the candidate object list.

[0058] In one optional implementation, determining the symmetry between the corresponding detection object and the candidate object in response to the speed difference, position difference, and position change difference satisfying a predetermined condition can specifically be performed as follows: calculating the speed difference reference value, position difference reference value, and position change difference reference value corresponding to the speed difference, position difference, and position change difference of each candidate object, respectively; calculating the weighted value of the speed difference reference value, position difference reference value, and position change difference reference value of each candidate object based on a second weighting reassembly; obtaining the target reference value of each candidate object; and determining the symmetry between the corresponding detection object and the candidate object in response to the target reference value being greater than or equal to a second reference threshold. That is, in this embodiment, if the speed difference, position difference, and / or position change difference of a detection object and a candidate object are substantially similar, then the detection object and the candidate object are likely symmetrical. Optionally, the second reference threshold can be determined based on specific application scenario testing. For example, the reference threshold can be 0.7.

[0059] In this embodiment, the target reference value Metric2 is calculated as follows:

[0060] Metric1=d*Δdv+e*Δdx+f*Δvardy (2)

[0061] Where Δdv, Δdx, and Δvardy represent the reference values ​​for velocity difference, position difference, and position change difference, respectively, and d, e, and f are the weights of these reference values, respectively. Optionally, d + e + f = 1.

[0062] Optionally, in this embodiment, the weighted recombination [d,e,f] can be predetermined according to the specific application scenario. Further, optionally, the weighted recombination [d,e,f] can be predetermined using exhaustive methods or similar techniques according to the specific application scenario. It should be understood that this embodiment does not limit the method for determining the weighted recombination [d,e,f].

[0063] Optionally, the speed difference reference value Δdv of the candidate object is used to characterize the speed difference between the candidate object and the detected object. Optionally, in this embodiment, the speeds of the candidate object and the detected object relative to the vehicle are obtained by radar detection, and the speed difference between the candidate object and the detected object is calculated based on the relative speeds. Further optionally, since the speeds of two mirrored objects have a certain symmetry, that is, their speed difference is not greater than a speed threshold. The speed threshold can be a value close to 0, which can be determined based on speed difference tests of two mirrored objects in a specific scenario, and this embodiment does not limit it.

[0064] Therefore, in one optional implementation, if the speed difference between the candidate object and the detected object is less than or equal to a speed threshold, the speed difference reference value Δdv between the candidate object and the detected object can be 1; if the speed difference between the candidate object and the detected object is greater than the speed threshold, the speed difference reference value Δdv between the candidate object and the detected object can be 0 (or negative). It should be understood that the assignment of the speed difference reference value in this embodiment is merely exemplary. Other assignments that tend to be symmetrical when the speed difference is less than or equal to the speed threshold and asymmetrical when the speed difference is greater than the speed threshold, thus affecting the screening results, can all be applied to this embodiment. In other optional implementations, a function for the speed difference reference value Δdv and the speed difference can be pre-created to determine the speed difference reference value Δdv between the candidate object and the detected object based on this function and the speed difference between the candidate object and the detected object. Optionally, the speed difference reference value Δdv can decrease as the speed difference increases, etc.

[0065] Optionally, the position difference reference value Δdx of the candidate object is determined based on the relative positions of the candidate object, the detected object, and the vehicle, respectively. Optionally, this embodiment obtains the distances of the candidate object and the detected object relative to the vehicle in the driving direction through radar detection, and then calculates the distance difference between them based on these relative distances to determine the position difference. Further optionally, since the positions of two mirrored objects have a certain symmetry, that is, their distance difference in the driving direction is not greater than a distance threshold. The distance threshold can be a value close to 0, which can be determined based on distance difference tests of two mirrored objects in a specific scenario in the driving direction; this embodiment does not impose this limitation.

[0066] Therefore, in one optional implementation, if the distance difference between the candidate object and the detected object in the driving direction is less than or equal to a distance threshold, the position difference reference value Δdx between the candidate object and the detected object can be 1; if the distance difference between the candidate object and the detected object is greater than the distance threshold, the distance difference reference value Δdx between the candidate object and the detected object can be 0 (or negative). It should be understood that the assignment of the distance difference reference value in this embodiment is merely exemplary. Other assignments that tend to be symmetrical when the distance difference is less than or equal to the distance threshold and asymmetrical when the distance difference is greater than the distance threshold, thus affecting the screening results, can all be applied to this embodiment. In other optional implementations, a function relating the distance difference reference value Δdx and the distance difference can be pre-created to determine the distance difference reference value Δdx between the candidate object and the detected object based on this function and the distance difference between the candidate object and the detected object. Optionally, the distance difference reference value Δdx can decrease as the distance difference increases.

[0067] Optionally, the reference value Δvardy for the position change difference of the candidate object is determined based on the relative position changes of the candidate object, the detected object, and the vehicle, respectively. Optionally, this embodiment obtains the lateral distance change parameters of the candidate object and the detected object relative to the vehicle through radar detection, for example, represented by the lateral distance change variance. Then, based on these distance change parameters, the difference in lateral distance change between them is calculated to determine the position change difference. Further, optionally, since there is instability for a period of time after target establishment in the radar detection system, i.e., there may be jumps, the similarity of the lateral position changes of two objects can be used as a basis for whether they are symmetrical. Since the position changes of two mirrored objects have a certain symmetry, i.e., their lateral distance change difference is no greater than the transformation difference threshold. The transformation difference threshold can be a value close to 0, which can be determined based on tests of the lateral position change difference between two mirrored objects in a specific scenario; this embodiment does not impose this limitation.

[0068] In one optional implementation, taking the variance of lateral distance variation as an example to represent positional change, if the difference between the variance of lateral distance variation between the candidate object and a detected object is less than or equal to the transformation difference threshold, then the reference value Δvardy for the positional change difference between the candidate object and the detected object can be 1; if the difference between the variance of lateral distance variation between the candidate object and a detected object is greater than the transformation difference threshold, then the reference value Δvardy for the positional change difference between the candidate object and the detected object can be 0 (or negative). It should be understood that the assignment of the distance difference reference value in this embodiment is merely exemplary. Other assignments that tend to be symmetrical when the difference of lateral distance variation variance is less than or equal to the transformation difference threshold, and tend to be asymmetrical when the difference of lateral distance variation variance is greater than the transformation difference threshold, thus affecting the screening results, can all be applied to this embodiment. In other optional implementations, a function for the difference between the positional change difference reference value Δvardy and the variance of lateral distance variation can also be pre-created, and the reference value Δvardy for the positional change difference between the candidate object and the detected object can be determined based on this function and the difference of the variance of lateral distance variation between the candidate object and the detected object. Optionally, the reference value for the difference in positional variation, Δvardy, can decrease as the difference in variance of the lateral distance variation increases.

[0069] For example, suppose the weighted recombination [d,e,f] is [0.4,0.4,0.3]. If the detection determines that the speed difference between a candidate object and a detected object is less than a speed threshold, the distance difference in the driving direction is less than a distance threshold, and the difference in the variance of the lateral distance change in the vehicle's lateral direction is less than a transformation difference threshold, that is, the speed difference reference value Δdv = 1, the position difference reference value Δdx = 1, and the position change difference reference value Δvardy = 1 between the candidate object and the detected object, then the target reference value Metric2 between the candidate object and the detected object is 0.4*1 + 0.4*1 + 0.3*1 = 1 > 0.7. Therefore, it can be determined that the candidate object and the detected object are symmetrical, that is, the detected object is the mirror image of the candidate object.

[0070] In one alternative implementation, since the candidate object is determined based on its existence period, its relative position to the vehicle, and the motion state of the detected object, it is more inclined to be an actual existing target. Therefore, in this embodiment, the detected object that is symmetrical to the candidate object can be determined as a mirror target, that is, a virtual target constructed by multipath reflection.

[0071] In another optional implementation, to further reduce recognition errors, this embodiment can further identify objects with symmetrical relationships based on reference objects such as lane lines and medians, thereby ultimately determining the virtual mirror target. Further, in this embodiment, step 220 can also be executed as follows: determining candidate objects and detection objects symmetrical to the candidate objects as a target object group, and determining the target object from the target object group based on reference objects. Reference objects include lane lines and / or medians. Optionally, in this embodiment, the positional relationship between the objects in the target object group and the median is determined based on the position of the reference objects and the positions of objects in the target object group, and objects located outside the median in the target object group are determined as target objects. It should be understood that in this embodiment, being inside or outside the median is determined relative to the vehicle itself. Therefore, this embodiment can further reduce errors such as... Figure 1 The vehicles 12 and 12' that are symmetrical to each other may be misjudged as having mirror targets, thereby further improving the recognition accuracy.

[0072] In an optional implementation, since the target object determined in this embodiment is a virtual object generated by multipath reflection, in order to avoid the driving danger caused by the target object interfering with the driver, the method of this embodiment further includes: deleting each target object from each detected object determined by radar detection. Thus, the driver can avoid seeing each mirror target and improve driving safety.

[0073] This embodiment can initially screen each detection object based on its existence period, relative position to the vehicle, and motion state, i.e., filter out candidate objects that may have mirror images. Then, based on speed differences, position differences, and position change differences, a global search is performed on all detection objects to identify those that exist for each candidate object. Detection objects symmetrical to these candidate objects are then detected, thereby identifying the mirror image target among the symmetrically related detection objects. Therefore, this embodiment can improve the accuracy and efficiency of mirror image target recognition, thereby improving driving safety.

[0074] Figure 4 This is a flowchart of another target recognition method according to an embodiment of the present invention. Figure 5 This is a schematic diagram of the target recognition process according to an embodiment of the present invention. Figure 4 As shown, the target recognition method of this invention includes the following steps:

[0075] Step S210: Obtain the detection results from the radar on the vehicle. These results include the position information, existence period, and motion state information of multiple detected objects. For example... Figure 5 As shown, this vehicle 5 is equipped with multiple radars, which detect the surrounding environment information. The detection results include the median strip A, multiple lane lines B, vehicle 51, vehicle 51', vehicle 52, vehicle 52', vehicle 53, and vehicle 54, etc.

[0076] Step S220 involves determining the first reference information of the detected object. The first reference information includes motion state reference values, position reference values, and existence period reference values, specifically including steps S221-S223. It should be understood that in this embodiment, steps S221-S223 do not have a specific execution order; they can be executed synchronously or sequentially.

[0077] Alternatively, in an optional implementation, this embodiment can perform the candidate object determination step (steps S220-250) on each detection object individually, or it can perform the candidate object determination step on some or more detection objects simultaneously, or on all detection objects in parallel. This embodiment does not limit itself to this; the description here takes performing the candidate object determination step on each detection object individually as an example.

[0078] In one alternative implementation, objects near reflective surfaces such as median strips are more likely to produce mirror targets. Therefore, before executing step S220, this embodiment can first determine the first distance between each object and reflective objects such as median strips, and filter out objects with a first distance less than a first threshold (e.g., 5m) as initial screening objects. Then, based on the existence period, the relative position to the vehicle, and the motion state of the object, the initial screening objects are further filtered to obtain candidate objects. This can further improve target recognition efficiency.

[0079] Specifically, in step S221, a motion state reference value for the object to be detected is determined. In this embodiment, the relative speed of the object to be detected relative to the vehicle 5 is determined by the reflected wave of the radar. Based on this relative speed and the driving speed of the vehicle 5, it can be determined whether the object to be detected is in motion. Further optionally, if the object to be screened is in motion, the motion state reference value x1 of the object to be screened can be 1; if the object to be screened is stationary, the motion state reference value x1 of the object to be screened can be 0 (or a negative value). This embodiment describes the determination of the motion state reference value using the above assignment method as an example. It should be understood that other assignments that tend to be selected when in motion and tend to be screened out when stationary, thus affecting the screening results, can also be applied to this embodiment.

[0080] In step S222, a position reference value for the detected object is determined. In this embodiment, this is determined based on the position of the detected object relative to the vehicle 5 in the lateral direction (e.g., ...). Figure 5 The distance along the y-axis is determined. Further optionally, in response to the distance between the detected object and the vehicle 5 in the y-axis direction being less than or equal to a first distance threshold, the position reference value x2 of the detected object can be 1; in response to the distance between the detected object and the vehicle 5 in the y-axis direction being greater than the first distance threshold, the position reference value x2 of the detected object can be 0 (or a negative value). This embodiment describes the determination of the position reference value using the above assignment method as an example. It should be understood that other optional implementation methods, such as determining the optimal distance through testing, may result in the most likely generation of mirror targets or the generation of mirror targets being more likely to affect the driver's driving when the distance between the detected object and the vehicle 5 in the y-axis direction is the optimal distance. Therefore, when the distance between the detected object and the vehicle 5 in the y-axis direction is the optimal distance, the position reference value is set to the maximum value; the further the distance between the detected object and the vehicle 5 in the y-axis direction is from the optimal distance, the smaller the corresponding position reference value. In other words, in this embodiment, a position reference value can be preset based on a first distance threshold, or a function between the position reference value and the lateral distance can be predetermined, and the corresponding position reference value can be determined based on this function and the distance between the detected object and the vehicle 5 in the y-axis direction.

[0081] In step S223, a reference value for the existence period of the detected object is determined. Further optionally, in this embodiment, if the duration of the existence period of the detected object is greater than or equal to a duration threshold, the reference value for the existence period of the detected object can be 1; if the duration of the existence period of the detected object is less than the duration threshold, the reference value for the existence period of the detected object can be 0 (or a negative value). This embodiment describes the determination of the existence period reference value using the above assignment method as an example. It should be understood that other assignments that tend to select objects when the existence period is greater than the duration threshold and tend to filter them when the existence period is less than the duration threshold, thus affecting the filtering results, can also be applied to this embodiment. In other optional implementations, a function for the existence period reference value and the existence period can be pre-created to determine the existence period reference value of the detected object based on this function and the existence period of the detected object. Optionally, the existence period reference value can increase as the duration of the existence period increases.

[0082] Step S230: Determine candidate reference values ​​based on the motion state reference value, position reference value, and existence period reference value of the detected object.

[0083] Based on the first weighted recombination, the weighted sum of the motion state reference value, position reference value, and existence period reference value of the detected object is calculated to obtain the candidate reference value of the detected object. It should be understood that the specific calculation method is as described in the above formula (1), and will not be repeated here.

[0084] Step S240: Determine whether the candidate reference value corresponding to the current detection object is greater than the first reference threshold. If it is greater than or equal to the first reference threshold, then execute step S250. If it is less than the first reference threshold, then execute step S210, that is, obtain the next detection object and execute steps S210-240.

[0085] Step S250: In response to a candidate reference value being greater than or equal to a first reference threshold, the corresponding detection object is determined as a candidate object, and the currently searched detection object is determined. In this embodiment, a global search is performed on the acquired candidate object among all detection objects detected by the radar (or detection objects in all motion states) to obtain the detection object symmetrical to the candidate object. If no detection object symmetrical to the candidate object is detected, then the candidate object does not have a mirror target and can be deleted from the candidate object list. In this embodiment, the example of searching each search object one by one to determine the detection object symmetrical to the current candidate object is described. It should be understood that multiple detection objects can also be executed in parallel to improve recognition efficiency, and this embodiment does not limit this.

[0086] Step S260 involves determining the second reference information for the candidate object. The second reference information includes a velocity difference reference value, a position difference reference value, and a position change difference reference value. This specifically includes steps S261-S263. It should be understood that in this embodiment, steps S261-S263 do not have a specific execution order; they can be executed synchronously or sequentially.

[0087] Specifically, in step S261, a speed difference reference value for the candidate object is determined. Optionally, in this embodiment, the speeds of the candidate object and the detected object relative to the vehicle are obtained through radar detection, and the speed difference between the candidate object and the detected object is calculated based on the relative speeds. Further optionally, since the speeds of two mirrored objects have a certain symmetry, that is, their speed difference is not greater than a speed threshold. The speed threshold can be a value close to 0, which can be determined based on speed difference tests of two mirrored objects in a specific scenario, and this embodiment does not limit this. In an optional implementation, if the speed difference between the candidate object and a detected object is less than or equal to the speed threshold, the speed difference reference value Δdv between the candidate object and the detected object can be 1; if the speed difference between the candidate object and a detected object is greater than the speed threshold, the speed difference reference value Δdv between the candidate object and the detected object can be 0 (or a negative value). It should be understood that this embodiment uses the above-described assignment method as an example. Other assignments that can make the speed difference tend to be symmetrical when it is less than or equal to the speed threshold, and tend to be asymmetrical when the speed difference is greater than the speed threshold, thus affecting the screening results, can also be applied to this embodiment. For example, creating a speed difference reference value Δdv and a function of speed difference, etc., will not be elaborated here.

[0088] In step S262, a reference value for the position difference between the candidate object and the detected object is determined. Optionally, in this embodiment, the distances between the candidate object and the detected object relative to the vehicle in the driving direction are obtained by radar detection, and then the distance difference between them is calculated based on the relative distances to determine the position difference. Further optionally, since the positions of two mirrored objects have a certain symmetry, that is, their distance difference in the driving direction is not greater than a distance threshold. The distance threshold can be a value close to 0, which can be determined based on the distance difference test of two mirrored objects in the driving direction in a specific scenario, and this embodiment does not limit this. In an optional implementation, if the distance difference between the candidate object and a detected object in the driving direction is less than or equal to the distance threshold, the reference value Δdx for the position difference between the candidate object and the detected object can be 1; if the distance difference between the candidate object and the detected object is greater than the distance threshold, the reference value Δdx for the distance difference between the candidate object and the detected object can be 0 (or a negative value). It should be understood that this embodiment uses the above-described assignment method as an example. Other assignments that can make the result tend to be symmetrical when the distance difference is less than or equal to the distance threshold, and asymmetrical when the distance difference is greater than the distance threshold, thus affecting the filtering results, can also be applied to this embodiment. For example, pre-creating a distance difference reference value Δdx and a function for the distance difference, etc., will not be elaborated here.

[0089] In step S263, a reference value for the position change difference of the candidate object is determined. Optionally, in this embodiment, the distance change parameters of the candidate object and the detected object relative to the vehicle in the lateral direction are obtained by radar detection, for example, represented by the variance of the lateral distance change. Then, the difference in lateral distance change between them is calculated based on the distance change parameters to determine the position change difference. Further optionally, since there is instability for a period of time after the target is established in the radar detection system, that is, there may be jumps, the similarity of the position change of the two objects in the lateral direction of the vehicle can be used as a basis for whether they are symmetrical. Since the position changes of two mirrored objects have a certain symmetry, that is, the difference in their lateral distance change is not greater than the transformation difference threshold. The transformation difference threshold can be a value close to 0, which can be determined based on the test of the difference in lateral position change of two mirrored objects in a specific scenario. This embodiment does not limit this. In one optional implementation, taking the variance of lateral distance variation as an example to represent positional change, if the difference between the variance of lateral distance variation of the candidate object and a detected object is less than or equal to the transformation difference threshold, then the reference value Δvardy for the positional change difference between the candidate object and the detected object can be 1; if the difference between the variance of lateral distance variation of the candidate object and a detected object is greater than the transformation difference threshold, then the reference value Δvardy for the positional change difference between the candidate object and the detected object can be 0 (or negative). It should be understood that this embodiment uses the above assignment method as an example. Other assignments that can make the difference tend to be symmetrical when the difference of lateral distance variation variance is less than or equal to the transformation difference threshold, and asymmetrical when the difference of lateral distance variation variance is greater than the transformation difference threshold, thus affecting the screening results, can all be applied to this embodiment. For example, creating a function for the difference between the reference value Δvardy for positional change difference and the variance of lateral distance variation, etc., will not be elaborated here.

[0090] Step S270: Determine the target reference value based on the velocity difference reference value, position difference reference value, and position change difference reference value of the candidate object.

[0091] Optionally, the target reference value of the candidate object can be obtained by calculating the weighted sum of the velocity difference reference value, position difference reference value, and position change difference reference value of the candidate object based on the second weighting. It should be understood that the specific calculation method is described in formula (2) above, and will not be repeated here.

[0092] Step S280: Determine whether the target reference value corresponding to the current candidate object is greater than the second reference threshold. If it is greater than or equal to the second reference threshold, proceed to step S290. If it is less than the second reference threshold, proceed to step S250, that is, obtain the next search target and proceed to steps S250-280.

[0093] Step S290: In response to the target reference value being greater than or equal to the second reference threshold, it is determined that the current candidate object and the detected object have a symmetrical relationship, and the target object is determined from the group of objects with a symmetrical relationship based on the reference object. The reference object includes lane lines and / or median strips.

[0094] In this embodiment, the candidate object and the detection object symmetrical to the candidate object are determined as the target object group. Based on the position of the reference object and the position of the object in the target object group, the positional relationship between the object in the target object group and the isolation zone is determined, and the object located outside the isolation zone in the target object group is determined as the target object.

[0095] Step S2A0: Delete the target object from the detection results. Since the target object determined in this embodiment is a virtual object generated by multipath reflection, in order to avoid the driving danger caused by the target object interfering with the driver, the method of this embodiment further includes: deleting each target object from each detected object determined by radar detection. This can prevent the driver from seeing each mirror target and improve driving safety.

[0096] by Figure 5 The road conditions shown are provided as examples; please understand that... Figure 5 This is merely an example and does not represent a real-world scenario. For example... Figure 5 As shown, vehicle 5 uses radar to detect the environment and obtain detection results, which include the following detected objects: median strip A, multiple lane lines B, vehicle 51, vehicle 51', vehicle 52, vehicle 52', vehicle 53, and vehicle 54. Among them, median strip A, multiple lane lines B, and vehicle 54 are stationary, while vehicle 51, vehicle 51', vehicle 52, vehicle 52', and vehicle 53 are in motion.

[0097] In this embodiment, after performing steps S220-250 on the detection objects in the detection results, vehicle 51, vehicle 52, vehicle 52', and vehicle 53 are determined as candidate objects.

[0098] Next, a global search is performed on candidate vehicles 51, 52, 52', and 53 across all detected objects or objects in all motion states (including vehicles 51, 51', 52, 52', and 53) to determine symmetrical detected objects. This involves executing steps S250-290 to confirm that vehicles 51 and 51', and vehicles 52 and 52', have a symmetrical relationship. Finally, based on lane line B and / or median strip A, it is determined that vehicle 51' is located outside median strip A relative to vehicle 5. Therefore, vehicle 51' is identified as the target object. In other words, vehicle 51' is a virtual mirror image of vehicle 51 generated by multipath reflection, while vehicles 52 and 52', even if symmetrical, are actually existing vehicles. Thus, vehicle 51' can be removed from the detection results to avoid interfering with the driver or central control system, improving driving safety.

[0099] This embodiment can initially screen each detection object based on its existence period, relative position to the vehicle, and motion state, i.e., filter out candidate objects that may have mirror images. Then, based on speed differences, position differences, and position change differences, a global search is performed on all detection objects to identify those that exist for each candidate object. Detection objects symmetrical to these candidate objects are then detected, thereby identifying the mirror image target among the symmetrically related detection objects. Therefore, this embodiment can improve the accuracy and efficiency of mirror image target recognition, thereby improving driving safety.

[0100] Figure 6 This is a schematic diagram of a driving assistance system according to an embodiment of the present invention. Figure 6 As shown, the driving assistance system 6 of this embodiment includes at least one radar 61 and a target recognition device 62. The radar 61's coverage area includes at least the driver's blind spot; therefore, the radar 61 can be installed in the center of the rear of the vehicle. If multiple radars 61 are present, they can be installed at the left and right rear corners of the vehicle. Alternatively, radars 61 can be installed at the left front, right front, left rear, and right rear corners of the vehicle to cover the area around the vehicle.

[0101] Radar 61 is used to detect environmental information and acquire detection results, including the location information, existence period, and motion state information of the detected objects. Target recognition device 62 is used to acquire the detection results from the radar on the vehicle and determine the target object based on the location information, existence period, and motion state information of each detected object.

[0102] This embodiment acquires the position information, existence period, and motion state information of each detected object by radar on the vehicle, and determines the target object based on the position information, existence period, and speed information of each detected object. Therefore, this embodiment determines the target object by comprehensively considering various factors such as position, speed, and existence period, thereby improving the accuracy of target recognition.

[0103] In another embodiment of the present invention, a vehicle including a driving assistance system 6 is provided. The vehicle can acquire position information, existence period, and motion state information of each detected object detected by radar on the vehicle, and determine a target object based on the position information, existence period, and speed information of each detected object. Therefore, this embodiment determines the target object by comprehensively considering various factors such as position, speed, and existence period, thereby improving target recognition accuracy, avoiding interference from mirror targets, and improving driving safety.

[0104] Figure 7 This is a schematic diagram of a target recognition device according to an embodiment of the present invention. The target recognition device 7 of this embodiment includes an information acquisition unit 71 and a target recognition unit 72.

[0105] The information acquisition unit 71 is configured to acquire the detection results of the radar on the vehicle, the detection results including the position information, existence period, and motion state information of the detected object. Optionally, the position information of the detected object includes the relative position of the detected object relative to the vehicle and / or the change information of the relative position. The target recognition unit 72 is configured to determine the target object based on the position information, existence period, and motion state information of each detected object.

[0106] In one alternative implementation, the target recognition unit 72 includes a candidate object determination subunit and a detection subunit.

[0107] A candidate object determination subunit is configured to determine at least one candidate object based on the detection results. A detection subunit is configured to detect objects symmetrical to each of the candidate objects to determine the target object.

[0108] In one alternative implementation, the candidate object determination subunit includes a distance determination module, a preliminary screening module, and a candidate object determination module.

[0109] The distance determination module is configured to determine a first distance between each of the detected objects and the reflecting object. The initial screening module is configured to acquire initial screening objects in response to the first distance being less than a first threshold. The candidate object determination module is configured to acquire each of the candidate objects based on the motion state information, position information, and existence period of each of the initial screening objects.

[0110] In one alternative implementation, the candidate object determination module includes a first reference value determination submodule, a first processing submodule, and a candidate object determination submodule.

[0111] The first reference value determination submodule is configured to determine the motion state reference value, position reference value, and existence period reference value of each of the preliminary screening objects based on their motion state information, position information, and existence period. The first processing submodule is configured to calculate a weighted sum of the motion state reference value, position reference value, and existence period reference value of each of the preliminary screening objects based on a first weighted recombination, thereby obtaining a candidate reference value for each of the preliminary screening objects. The candidate object determination submodule is configured to determine the corresponding preliminary screening object as a candidate object in response to the candidate reference value being greater than or equal to a first reference threshold.

[0112] Optionally, the motion state reference value is determined based on whether the corresponding initial screening object is moving, the position reference value is determined based on the relative position of the corresponding initial screening object and the vehicle, and the existence period reference value is determined based on the duration of the existence period and a duration threshold.

[0113] In one optional implementation, the detection subunit includes a difference determination module and a detection module. The difference determination module is configured to determine the velocity difference, position difference, and position change difference between the candidate object and each of the detection objects. The detection module is configured to determine that the corresponding detection object is symmetrical to the candidate object in response to the velocity difference, position difference, and position change difference satisfying predetermined conditions.

[0114] In one alternative implementation, the detection module includes a second reference value determination submodule, a second processing submodule, and a detection submodule.

[0115] The second reference value determination submodule is configured to calculate the velocity difference reference value, position difference reference value, and position change difference reference value corresponding to the velocity difference, position difference, and position change difference, respectively. The second processing submodule is configured to calculate the weighted values ​​of the velocity difference reference value, position difference reference value, and position change difference reference value based on a second weighting reassembly, to obtain the target reference value for each candidate object. The detection submodule is configured to determine that the corresponding detection object is symmetrical to the candidate object in response to the target reference value being greater than or equal to a second reference threshold.

[0116] In an optional implementation, the detection subunit is further configured to identify the candidate object and a detection object symmetrical to the candidate object as a target object group, and to identify the target object from the target object group based on a reference object, wherein the reference object includes lane lines and / or median strips. Further optionally, the detection subunit is further configured to determine the positional relationship between the objects in the target object group and the median strip based on the position of the reference object and the position of the objects in the target object group, and to identify objects in the target object group located outside the median strip as target objects.

[0117] In an alternative implementation, the target recognition device 7 further includes a deletion unit. The deletion unit is configured to delete each of the detected targets from each of the detected objects.

[0118] This embodiment acquires the position information, existence period, and motion state information of each detected object by radar on the vehicle, and determines the target object based on the position information, existence period, and speed information of each detected object. Therefore, this embodiment determines the target object by comprehensively considering various factors such as position, speed, and existence period, thereby improving the accuracy of target recognition.

[0119] Another embodiment of the present invention relates to a non-volatile storage medium for storing a computer-readable program for use by a computer to execute some or all of the above-described method embodiments.

[0120] That is, those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. This program is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods described in the embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0121] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. For those skilled in the art, the present invention can be modified and varied in various ways. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of the present invention should be included within the scope of protection of the present invention.

Claims

1. A target recognition method, characterized in that, The method includes: The detection results of the radar on the vehicle are obtained, including the location information, existence period and motion state information of the detected object; The target object is determined based on the location information, existence period, and motion state information of each of the detected objects; Determining the target object based on the location information, existence period, and motion state information of each of the detected objects includes: Based on the motion state information, position information, and existence period of each preliminary screening object, the motion state reference value, position reference value, and existence period reference value of each preliminary screening object are determined; the preliminary screening objects are obtained by screening the detection objects using the first distance between each detection object and the reflecting object; Based on the first weighted recombination, the weighted sum of the motion state reference value, position reference value and existence period reference value of each of the preliminary screening objects is calculated to obtain the candidate reference value of each of the preliminary screening objects; In response to the candidate reference value being greater than or equal to the first reference threshold, the corresponding preliminary screening object is determined as a candidate object; Detecting a detection object that is symmetrical to each of the candidate objects to determine the target object.

2. The method according to claim 1, characterized in that, The location information of the detected object includes the relative position of the detected object with respect to the vehicle and / or information on changes in the relative position.

3. The method according to claim 1, characterized in that, Based on the detection results, at least one candidate object is identified, including: Determine the first distance between each of the detection objects and the reflecting object; In response to the first distance being less than the first threshold, the initial screening object is obtained; Based on the motion state information, position information, and existence period of each of the preliminary screening objects, each of the candidate objects is obtained.

4. The method according to claim 3, characterized in that, The motion state reference value is determined based on whether the corresponding initial screening object is moving, the position reference value is determined based on the relative position of the corresponding initial screening object and the vehicle, and the existence period reference value is determined based on the duration of the existence period and a duration threshold.

5. The method according to claim 1, characterized in that, The detection of objects symmetrical to each of the aforementioned candidate objects includes: Determine the velocity difference, position difference, and position change difference between the candidate object and each of the detected objects; In response to the speed difference, position difference, and position change difference satisfying predetermined conditions, the corresponding detection object is determined to be symmetrical with the candidate object.

6. The method according to claim 5, characterized in that, In response to the speed difference, position difference, and position change difference satisfying predetermined conditions, determining that the corresponding detection object is symmetrical to the candidate object includes: Calculate the reference values ​​for velocity difference, position difference, and position change difference, respectively, corresponding to the velocity difference, position difference, and position change difference. Based on the second weighted recombination, the weighted values ​​of the velocity difference reference value, position difference reference value, and position change difference reference value are calculated to obtain the target reference value for each candidate object; In response to the target reference value being greater than or equal to the second reference threshold, it is determined that the corresponding detection object is symmetrical with the candidate object.

7. The method according to any one of claims 3-6, characterized in that, Detecting a detection object symmetrical to each of the candidate objects to determine the target object includes: The candidate objects and the detection objects that are symmetrical to the candidate objects are identified as the target object group; The target object is determined from the target object group based on a reference object, the reference object including lane lines and / or median strips.

8. The method according to claim 7, characterized in that, Determining the target object from the target object group based on the reference object includes: Based on the position of the reference object and the position of the object in the target object group, determine the positional relationship between the object in the target object group and the isolation zone; The objects located outside the isolation zone in the target object group are identified as target objects.

9. The method according to claim 7, characterized in that, The method further includes: Delete each of the target objects from each of the detection objects.

10. A driving assistance system, characterized in that, The driving assistance system includes: Radar is used to detect environmental information and obtain detection results, which include the location information, existence period and motion state information of the detected object; A target recognition device is used to acquire the detection results of radar on a vehicle and determine the target object based on the position information, existence period and motion state information of each detected object; Determining the target object based on the location information, existence period, and motion state information of each of the detected objects includes: Based on the motion state information, position information, and existence period of each preliminary screening object, the motion state reference value, position reference value, and existence period reference value of each preliminary screening object are determined; the preliminary screening objects are obtained by screening the detection objects using the first distance between each detection object and the reflecting object; Based on the first weighted recombination, the weighted sum of the motion state reference value, position reference value and existence period reference value of each of the preliminary screening objects is calculated to obtain the candidate reference value of each of the preliminary screening objects; In response to the candidate reference value being greater than or equal to the first reference threshold, the corresponding preliminary screening object is determined as a candidate object; Detecting a detection object that is symmetrical to each of the candidate objects to determine the target object.

11. A vehicle, characterized in that, The vehicles include: The vehicle itself; and The driving assistance system as described in claim 10.

12. A target recognition device, characterized in that, The target identification device includes: The information acquisition unit is configured to acquire the detection results of the radar on the vehicle, the detection results including the location information, existence period and motion state information of the detected object; The target recognition unit is configured to determine the target object based on the position information, existence period and motion state information of each of the detected objects; Determining the target object based on the location information, existence period, and motion state information of each of the detected objects includes: Based on the motion state information, position information, and existence period of each preliminary screening object, the motion state reference value, position reference value, and existence period reference value of each preliminary screening object are determined; the preliminary screening objects are obtained by screening the detection objects using the first distance between each detection object and the reflecting object; Based on the first weighted recombination, the weighted sum of the motion state reference value, position reference value and existence period reference value of each of the preliminary screening objects is calculated to obtain the candidate reference value of each of the preliminary screening objects; In response to the candidate reference value being greater than or equal to the first reference threshold, the corresponding preliminary screening object is determined as a candidate object; Detecting a detection object that is symmetrical to each of the candidate objects to determine the target object.

13. A computer-readable storage medium storing computer program instructions thereon, characterized in that, The computer program instructions, when executed by a processor, implement the method as described in any one of claims 1-9.

Images