Methods, devices, and storage media for identifying target objects based on vehicle-mounted radar

Abstract

This disclosure provides a method, apparatus, and storage medium for identifying target objects based on vehicle-mounted radar, which can accurately identify target objects to reduce the false alarm rate of vehicle driver assistance systems. Specifically, the method includes: receiving detection data from the vehicle-mounted radar within a preset spatial range; determining, based on the detection data, whether an object satisfying a first condition exists within the preset spatial range; when an object satisfying the first condition exists within the preset spatial range, further determining whether the object's position parameters satisfy a second condition; and when the object's position parameters satisfy the second condition, determining that the target object exists in the external space of the vehicle.

Description

Technical Field

[0001] This disclosure relates to the field of driver assistance systems, specifically to methods, devices, and storage media for identifying target objects based on vehicle-mounted radar. Background Technology

[0002] In the field of vehicle driver assistance, vehicle-mounted millimeter-wave radar transmits millimeter waves outward through an antenna and assesses the physical environment around the vehicle based on the collected echo data. This allows it to obtain environmental parameters such as the orientation and direction of motion of the measured object, the relative distance and relative speed between the vehicle and the measured object, and further track, identify, and classify the measured object based on the acquired environmental parameters, thereby reducing the false alarm rate and false alarm rate of the vehicle driver assistance system.

[0003] For example, guardrails, as a typical static category, exhibit a real-time change in their extension direction following the direction of the highway, and necessary openings exist along their installation path. In complex and ever-changing road conditions, considering the inherent limitations of radar and errors in assessing environmental parameters, the position and structural state of the object being measured cannot be accurately determined solely by the detection points and trajectory information of the object in the current frame. This means the target object cannot be accurately identified, thus affecting the false alarm rate of the driver assistance system. Summary of the Invention

[0004] This disclosure provides a method, apparatus, and storage medium for identifying target objects based on vehicle-mounted radar, which helps reduce the false alarm rate of vehicle driver assistance systems.

[0005] To achieve the above objectives, this disclosure provides a method for identifying target objects based on vehicle-mounted radar. The method includes: (S100): receiving detection data from the vehicle-mounted radar within a preset spatial range; (S200): determining, based on the detection data, whether an object satisfying a first condition exists within the preset spatial range; (S300): when an object satisfying the first condition exists within the preset spatial range, further determining whether the position parameters of the object satisfy a second condition; (S400): when the position parameters of the object satisfy the second condition, determining that the target object exists in the external space of the vehicle.

[0006] Optionally, step (S100) includes: dividing the preset spatial range into at least two regions, and configuring a buffer of preset capacity for each region to store the position parameters of objects detected in the corresponding region that satisfy the first condition.

[0007] On the other hand, this disclosure provides a device for identifying target objects based on vehicle-mounted radar. The device includes: a first processing unit, a second processing unit, a third processing unit, and a fourth processing unit. The first processing unit is used to receive detection data from the vehicle-mounted radar within a preset spatial range; the second processing unit is used to determine, based on the detection data, whether an object satisfying a first condition exists within the preset spatial range; the third processing unit is used to further determine, when an object satisfying the first condition exists within the preset spatial range, whether the object's position parameters satisfy a second condition; and the fourth processing unit is used to determine, when the object's position parameters satisfy the second condition, that the target object exists in the external space.

[0008] Optionally, the first processing unit is further configured to divide the preset spatial range into at least two regions, and to configure a buffer of preset capacity for each region to store the position parameters of objects detected in the corresponding region that satisfy the first condition.

[0009] For example, the first condition mentioned in the above scheme refers to the object's ground velocity being zero; and the position parameter is the lateral distance between the object and the vehicle-mounted radar; and the target object is a road guardrail.

[0010] In another aspect, this disclosure provides a computer-readable storage medium for storing processor-executable instructions, wherein the processor-executable instructions stored in the computer-readable storage medium, when executed, cause the processor to perform the method for identifying target objects based on vehicle radar as described above.

[0011] Compared to existing technologies, the method for identifying target objects based on vehicle-mounted radar provided in this disclosure helps reduce the false alarm rate of vehicle-assisted driving systems. The target object is characterized by being stationary relative to the ground and having a continuous structure along the road's extension direction, such as fixed road infrastructure like guardrails. Specifically, the method provided in this disclosure takes into account complex and varied road scenarios by dividing the detection range of the vehicle-mounted radar into several regions and configuring a buffer for each region to store the position parameters of multiple stationary objects within that region. Furthermore, based on the position parameters of the objects stored in the buffers of each region, and combined with a second condition, a comprehensive judgment is made as to whether an object matching the characteristics of a target object exists within the detection range of the vehicle-mounted radar. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of this disclosure, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 This is a schematic diagram illustrating the steps of a method for identifying target objects based on vehicle-mounted radar, as provided in an embodiment of this disclosure.

[0014] Figure 2 yes Figure 1 A schematic diagram of the sub-steps of S100.

[0015] Figure 3 This is a schematic diagram illustrating the area division of the detection range of a vehicle-mounted radar according to an embodiment of this disclosure.

[0016] Figure 4 This is a schematic diagram illustrating the area division of the detection range of a vehicle-mounted radar, as provided in another embodiment of this disclosure.

[0017] Figure 5 This is a schematic diagram of a driving environment provided in an embodiment of this disclosure.

[0018] Figure 6 This is a schematic diagram of a device for identifying target objects based on vehicle-mounted radar, provided in an embodiment of this disclosure. Detailed Implementation

[0019] The technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure. Furthermore, it should be understood that the specific embodiments described herein are only for illustration and explanation and are not intended to limit this disclosure. In this disclosure, unless otherwise stated, directional terms such as "upper" and "lower" generally refer to the upper and lower positions of the device in actual use or operation, specifically the drawing directions in the accompanying drawings; while "inner" and "outer" refer to the outline of the device.

[0020] On the one hand, to achieve the above objectives, this disclosure provides a method for identifying target objects based on vehicle-mounted radar, including as follows: Figure 1 The steps shown are as follows:

[0021] S100: Receives detection data from the vehicle-mounted radar within a preset spatial range.

[0022] In some embodiments of this disclosure, the vehicle-mounted radar may be at least one of millimeter-wave radar, lidar, and acoustic radar.

[0023] In some embodiments of this disclosure, objects that are stationary relative to the ground and have a continuous structure are used as target objects for identification by this method. It is understood that these objects are typically public facilities such as guardrails in the center of a highway or noise barriers along the roadside. After identifying an object that meets the characteristics of being stationary and having a continuous structure, this object is recorded as the target object, and the lateral distance between the target object and the vehicle-mounted radar is recorded, while also being categorized and stored. During subsequent vehicle driving, if the driver assistance system identifies an object with the same characteristics as the target object, it will not consider it an obstacle, thereby reducing the false alarm rate of the driver assistance system and ensuring the smoothness of vehicle driving.

[0024] Specifically, in some embodiments of this disclosure, S100 includes, for example... Figure 2 The sub-steps shown:

[0025] S110: Acquire one frame of detection data from the vehicle radar within a preset spatial range every preset time interval.

[0026] Specifically, the detection data for each frame includes a point trace that corresponds one-to-one with the object detected in that frame.

[0027] Typically, radars are installed at the four corners of the car body, and, for example, please refer to [reference needed]. Figure 3 The radar installed at the left rear corner of the vehicle can detect objects to the left of the vehicle's direction of travel, especially within the area to the left rear of the vehicle. Here, the preset spatial range refers to the area to the left of the vehicle's direction of travel, and the size of the detection range is mainly determined by the radar's functional characteristics.

[0028] The radar scans objects within a predetermined spatial range of the vehicle at fixed intervals. Electromagnetic waves (or sound waves, etc.) emitted by the radar are reflected off the surfaces of objects, and the radar receives the echo data to obtain a frame of detection data. A frame of detection data contains multiple reflection points of electromagnetic waves reflected from the surfaces of various objects. The reflection points corresponding to the same type of object can reflect the object's outline. Through clustering, the reflection points of the same type of object are grouped into a single point track to represent the object's position in the current frame. The points tracks in the current frame, together with the corresponding points tracks in previous frames, form tracks for different objects. That is, in one frame, one object corresponds to one point track, and the set of points corresponding to an object across multiple frames constitutes the track corresponding to that object.

[0029] Ideally, if there is a target object within the detection range for each frame, the driver assistance system will automatically classify objects that meet the same characteristics into the same category when processing the data in each frame, and will not interfere with the normal driving of the vehicle based on such objects.

[0030] However, considering the complex and varied scenarios on highways, such as gaps in guardrails or interruptions in mufflers, the system may not be able to accurately classify objects within the detection range for every frame. For example, it might mistake an interrupted guardrail for an obstacle and issue a warning to the driver via the vehicle's control center. This effectively increases the false alarm rate of the driver assistance system.

[0031] Therefore, in some embodiments of this disclosure, S100 further includes:

[0032] S120: Divide the preset space range into at least two areas.

[0033] S130: Configure a buffer of preset capacity for each region to store the position parameters of the objects detected in the corresponding region.

[0034] For example, please refer to Figure 3 Taking a radar installed at the left rear corner of the vehicle as an example. Specifically, a planar coordinate system is established with the location of the radar at the left rear corner as the origin, and the positive direction of the Y-axis is defined as the left side of the vehicle. Simultaneously, the detection range (preset spatial range) of the radar at the left rear corner is divided as follows: Figure 3 The three areas shown are: zone i Where i = [1,2,3], specifically, zone1 represents zone one, zone2 represents zone two, and zone3 represents zone three. It should be noted that the radar installed on the left rear of the vehicle mainly covers zone1 and zone2.

[0035] Furthermore, a buffer is configured for each region to store multiple frames of detection results for the same region. That is, for the same region, multiple frames of data are analyzed to store a preset number of lateral distances between stationary objects in the region and the vehicle-mounted radar, so as to accurately determine whether there are target objects that meet the conditions within the radar's detection range.

[0036] Specifically, with Figure 3 Taking the partitioning method shown as an example, the buffer corresponding to each region is: buffer i Where i = [1,2,3], specifically, buffer1, buffer2, and buffer3 represent the buffers corresponding to zone1, zone2, and zone3, respectively.

[0037] In some embodiments of this disclosure, S100 further includes:

[0038] S140: Initialize the buffer corresponding to each region.

[0039] Specifically, in the initial stage of identifying the target object, the buffer for each region needs to be initialized.

[0040] For example, the buffers corresponding to each region are configured to have the same capacity. Here, the buffer capacity refers to the number of elements it can store. In some embodiments of this disclosure, the buffer... i The elements stored in are zones i The position parameters of the detected stationary object, namely the lateral distance between the stationary object and the vehicle-mounted radar, which is also the coordinate value of the stationary object in the Y-axis direction (y). i (where i = [1,2,3]).

[0041] For ease of description, let buffer be an example. i The capacity is 4, meaning that each region's corresponding cache can store 4 elements (y). i It should be noted that in other embodiments of this disclosure, the capacity of the buffers in each region may not be exactly the same to accommodate different partitioning methods or differences in radar detection accuracy in different regions. For example, buffer 1 of zone 1 can store four position parameters (y1), which represent the lateral distance between a detected stationary object in zone 1 and the vehicle-mounted radar. Similarly, buffer 2 of zone 2 can store four y2 values; and buffer 3 can store four y3 values.

[0042] To distinguish buffers i The four positional parameters (y) i ), will buffer i The index of each positional parameter in the buffer is denoted as currentId, and currentId = [1, 2, 3, 4]. Further, the buffer... i The occupied space (y has already been stored) i The quantity is expressed in size. i This indicates that, and for a buffer with a capacity of 4... i size i = [0,1,2,3,4]. For example, if the position parameter (y1) has not yet been stored in buffer1, then the size of buffer1 is 0 (or denoted as size). i =0); if buffer1 already contains two positional parameters (y1), then size1 = 2.

[0043] During the initialization phase, the buffer is initialized. iThe parameters mentioned above. Specifically, clearing the buffer. i The location parameter is stored in [the memory]. Correspondingly, size [is used for this]. i =0 indicates the current buffer i The position parameters (y) of the object are not stored in the data. i At the same time, let the buffer i In the buffer, currentId = 1 indicates that it will soon be added to the buffer. i Write y at index 1 i .

[0044] S200: Determine whether there is an object within the preset space range that meets the first condition based on the detection data.

[0045] Specifically, for each frame of detection data, based on the trajectory corresponding to each point within the frame, it is determined whether the object corresponding to each point in the frame satisfies the first condition. The position parameters of the objects that satisfy the first condition are then stored in the buffer corresponding to the corresponding region.

[0046] In some embodiments of this disclosure, the first condition refers to the ground velocity of the object corresponding to the dot being zero, i.e., the detected object is stationary relative to the ground. The position parameter refers to the distance between the object and the vehicle-mounted radar.

[0047] For example, the ground velocity of an object within the detection range can be determined based on the object's trajectory. Specifically, the trajectory data of the objects corresponding to each point in the current frame are updated based on the data of each point in the current frame. Based on the updated trajectory data, the ground velocity of the object corresponding to each trajectory is determined. Objects with a ground velocity of zero are identified as stationary objects.

[0048] Furthermore, the zone to which the stationary object's dots belong in the current frame is further confirmed. i This allows the position parameters of the stationary object to be stored in the buffer of the corresponding region. i The position parameters include the lateral distance between the object and the vehicle-mounted radar, i.e., the coordinate value of the object in the Y-axis direction (y...). i ).

[0049] In some embodiments of this disclosure, for the same frame of probe data, if the buffer i If the capacity has been exhausted, then start from index 1 (currentId). i =1) location overwrite storage.

[0050] S300: When there is an object within the preset space that meets the first condition, further determine whether the position parameters of the object meet the second condition.

[0051] Specifically, the location parameters stored in the buffer corresponding to each region are sorted according to a preset rule, so as to extract at least one location parameter from the buffer corresponding to each region. Then, it is determined whether the location parameters extracted from the buffer corresponding to each region satisfy the second condition.

[0052] The second condition refers to the following: within a preset number of frames, the position parameters of preferred objects located in different regions that meet the first condition satisfy a first numerical relationship, a second numerical relationship, or a third numerical relationship.

[0053] For example, please continue reading Figure 3 Taking a radar installed at the left rear corner of a vehicle as an example, within the detection range corresponding to the current frame, the position parameters of objects that meet the first condition are stored in the buffer of the corresponding area. i In the context of this, the preset rule is as follows: The position parameters of objects stored in the buffer corresponding to each region are sorted according to their numerical values ​​from smallest to largest, and the index of the position parameter with the smallest numerical value is set to 1 (currentId = 1). The preferred object position parameters can be obtained from each buffer. i The second smallest positional parameter (currentId) in the table. i =2), but not limited to this. The buffer i The position parameters of the preferred object in the buffer are extracted and determined. i Extracted position parameters (y i Does it meet the second condition?

[0054] In some embodiments of this disclosure, the second condition includes: y2 is a valid value, and y1, y2, and y3 should satisfy a first numerical relationship. In this case, the first numerical relationship is: |y1-y2| < threshold and / or |y3-y2| < threshold.

[0055] For the radar installed at the left rear corner of the vehicle, considering that it mainly covers two areas, zone1 and zone2, the first numerical relationship is: |y1-y2|<threshold.

[0056] For the radar installed at the left front corner of the vehicle, considering that it mainly covers two areas, zone3 and zone2, the first numerical relationship is: |y3-y2|<threshold.

[0057] For the radar installed on the middle left side of the vehicle, considering that it mainly covers three areas: zone1, zone2 and zone3, the first numerical relationship is: the values ​​of |y1-y2| and |y3-y2| are both less than the threshold.

[0058] In this context, a valid value can be understood as a value that conforms to the current detection environment. For example, if the target object is a guardrail in the center of a highway, and the extracted position parameter is significantly larger than the road width, or clearly does not conform to a reasonable distance range between the vehicle-mounted radar and the guardrail (e.g., too small or too large), then the position parameter is an invalid value. If the position parameter conforms to a reasonable lateral distance range between the guardrail and the vehicle-mounted radar, then the position parameter can be considered a valid value. For example, if the objects corresponding to position parameters y1, y2, and y3 are highway guardrails distributed in zones 1, 2, and 3, then the threshold can be understood as the installation tolerance of the highway guardrail, specifically, the installation tolerance along the highway extension direction.

[0059] In some embodiments of this disclosure, the second condition further includes: when y2 is not a valid value and / or y1, y2, and y3 do not satisfy the first numerical relationship, y2 satisfies a second numerical relationship with a preset value. In this case, the second numerical relationship is: |preset value - y2| < threshold.

[0060] Specifically, when y2 is not a valid value and / or y1, y2, and y3 do not satisfy |y1-y2|<threshold and / or |y3-y2|<threshold, a preset value is obtained, and the relationship between y2 (the position parameter extracted from buffer2) and the preset value is determined. If the absolute value of the difference between the two does not exceed a threshold, then y2 is considered a valid value, and the object corresponding to y2 can be considered the target object, and the lateral distance between the target object and the vehicle radar is y2. It should be noted that the preset value is determined based on the current environment of the vehicle. For example, if there is a guardrail on one side of the vehicle radar, the preset value will be set to be slightly greater than (or slightly less than, or equal to) the distance between the vehicle radar and the guardrail. Therefore, the preset value can also be understood as the expected distance between the guardrail (target object) and the vehicle radar. If it is determined that y2 is close to the preset value, then y2 can be considered a valid value, and the object corresponding to it is the target object, i.e., the guardrail.

[0061] In some embodiments of this disclosure, the second condition further includes: when y2 is not a valid value and / or y1, y2, and y3 do not satisfy the first numerical relationship, and y2 does not satisfy the second numerical relationship, at least one of y1, y2, and y3 satisfies a third numerical relationship with the preset value. In this case, the third numerical relationship is: y - preset value < threshold, where y is at least one of y1, y2, and y3.

[0062] Specifically, if y2 is not a valid value and / or y1, y2, and y3 do not satisfy |y1-y2|<threshold and / or |y3-y2|<threshold, and |preset value-y2|≥threshold, then if it is determined that at least one of y1, y2, and y3 satisfies the numerical relationship "y-preset value<threshold" with the preset value, it indicates that at least one of the objects corresponding to y1, y2, and y3 is located on the side of the expected guardrail closer to the vehicle. Therefore, a target object may exist within the detection range of the vehicle-mounted radar, and this target object refers to the expected guardrail whose distance from the vehicle-mounted radar is approximately the preset value.

[0063] S400: When the position parameters of the object meet the second condition, it is determined that there is a target object in the space outside the vehicle.

[0064] Specifically, the object corresponding to the position parameter that satisfies the second condition is determined to be the target object.

[0065] For example, if the extracted y2, y1, and / or y3 satisfy the second condition, then the objects corresponding to y2, y1, and / or y3 are considered to be the same type of object, and this object is the target object. Meanwhile, the lateral distance between the target object and the vehicle-mounted radar can be determined based on the position parameter (y2) of the preferred object in buffer2.

[0066] In some embodiments of this disclosure, after the target object is determined, in order to ensure that the identified target object is the correct one, it is also necessary to verify the position parameters of the identified target object based on a verification condition.

[0067] The verification condition is that the position parameters of the target object and the position parameters of objects within the preset spatial range that do not meet the first condition satisfy a fourth numerical relationship.

[0068] Specifically, in Figure 3 When the angle corresponding to Region 2 shown is relatively small, the objects in this region are relatively close to the vehicle radar, and the detection wave emitted by the radar is essentially perpendicular to the radar. The detection wave has strong reflectivity on the surface of the objects in Region 2, which may mask the echo data of some objects in Region 1 and Region 3.

[0069] Typically, to address this issue, the angular range corresponding to Region 1 and Region 3 is appropriately increased. However, as the detection space of Region 1 and Region 3 expands, some objects positioned above the road will enter the detection range of the vehicle-mounted radar, such as road signs, billboards, overpasses, or the tops of tunnels. The position parameters of these objects are then stored in buffer 1 and / or buffer 3. Simultaneously, the position parameters of manhole covers, roadblocks, and objects mistakenly identified as stationary vehicles on the road surface are also stored in buffer 2.

[0070] Due to the existence of chance, if the aforementioned objects are selected as preferred objects, and their respective position parameters (y2', y1', and / or y3') happen to satisfy the second condition mentioned above, namely: y2' is a valid value, and |y1'-y2'| and / or |y2'-y3'| are less than a threshold, then the system will mistakenly identify the target object and record the lateral distance between the "target object" and the vehicle radar as y2'. However, the "target object" at this time does not meet the characteristics of being stationary and having a continuous structure. In fact, the objects corresponding to y2', y1', and / or y3' may neither form a continuous structure nor be stationary; it's just that at a certain moment, the line formed by these objects in space happens to be approximately parallel to the direction of road extension. In this case, it can be considered that the vehicle's driver assistance system has misidentified the target object.

[0071] Therefore, in order to further improve the accuracy of target object recognition in the vehicle assisted driving system, in some embodiments of this disclosure, the detection space of the vehicle radar may further include a zone 4, and a corresponding buffer 4 is used to store the position parameters (y4') of the dynamic objects detected in zone 4. Furthermore, the position parameters of the selected dynamic objects in buffer 4 are used to verify whether the target object identified in S300 is correct.

[0072] Specifically, taking the radar installed in the middle of the left side of the vehicle as an example, the detection range of the vehicle-mounted radar is as follows: Figure 4 The areas are divided as shown. Each area (zone) i It has a corresponding buffer. i Furthermore, the angles (α) of the four regions iThe ranges are as follows: 5° < α1 < 70°, 80° < α2 < 100°, 115° < α3 < 175°, 60° < α4 < 120°; where i = [1, 2, 3, 4]. Further, buffer1, buffer2, and buffer3 are used to store the position parameters of static objects detected in zone1, zone2, and zone3, respectively; buffer4 is used to store the position parameters of dynamic objects detected in zone four.

[0073] The lateral distance (y2') between the target object determined in S300 and the vehicle-mounted radar, and the position parameters (y4') of the preferred object extracted from buffer4 are obtained. It is then determined whether y2' and y4' satisfy a verification condition. Specifically, if y2' is a valid value, the verification condition is that y4' and y2' should satisfy the fourth numerical relationship of "|y4'-y2'|≥threshold".

[0074] To help readers better understand the methods provided in this disclosure, please refer to [link / reference needed]. Figure 5 The driving environment is shown in the diagram. Specifically, there is a vehicle 1 traveling side-by-side with at least one vehicle 2 on one side of the road guardrail, and road signs are spaced apart in the space above the road. Vehicle 1 has an onboard radar mounted on the center left side of its body; for vehicle 1, the correct target object is the guardrail.

[0075] Specifically, y4' represents the lateral distance between the vehicle-mounted radar of vehicle one and vehicle two. If the target object's y2' and y4' determined in S300 satisfy the verification condition, i.e., |y4'-y2'| ≥ the threshold, then the identified target object is considered correct. If |y4'-y2'| < the threshold, it means that the position of the identified target object is close to or even overlaps with the position of vehicle two traveling normally in the lane. That is, the object actually corresponding to y2' may be vehicle two, or it may be an object such as a manhole cover or road barrier placed on the road surface in the same lane, and not part of the guardrail. In this case, the identified target object is incorrect.

[0076] It should be noted that when y2' is a valid value and |y4'-y2'| ≥ the threshold, there are two possibilities. One possibility is that the object corresponding to y2' is located between vehicle two and vehicle one. However, considering the inherent blind zone of the radar and the actual distance between vehicle one and vehicle two in a real-world scenario, an object between the two vehicles is almost impossible to detect. The other possibility is that the object corresponding to y2' is located on the side of vehicle two furthest from vehicle one. This could be a guardrail, an object near the guardrail, or an object between the guardrail and vehicle two. However, considering the actual distance between the guardrail and vehicle two in a real-world scenario, an object between vehicle two and the guardrail can be considered equivalent to an object near the guardrail. Furthermore, considering the installation error of the guardrail and the necessary changes in the guardrail laying path angle, an object near the guardrail could also be the guardrail itself. Therefore, when y2' is a valid value and |y4'-y2'| is greater than or equal to the threshold, it can be considered that the correct target object has been identified, i.e., the guardrail has been identified.

[0077] If the position parameters of the identified target object do not meet the verification conditions, i.e., the identified target object is incorrect, then the position parameter (y2') of the incorrect target object in buffer2 is cleared. Specifically, the position parameter with the largest value in buffer2 is overwritten at the position of the position parameter corresponding to the incorrect target object, and the position parameter with the largest value (larger position index) is cleared while ensuring the continuity of the position parameters stored in buffer2, in order to reduce the duplication rate of position parameters in buffer2. At the same time, the position parameters of dynamic objects stored in buffer4 are processed in the same way to ensure that the verification conditions can be updated in a timely manner with buffer2. If necessary, y1' and / or y3' that together with y2' constitute the incorrect target object in buffer1 and / or buffer3 can also be cleared in the same way.

[0078] In some embodiments of this disclosure, after completing the above verification steps, in order to update the buffer in a timely manner... i The location parameters stored in the buffer can be aged and cleaned according to a preset period, where i = [1, 2, 3, 4].

[0079] Specifically, the period can be a frame period, and a fixed number of frames are contained within one period. Each time a number of frames corresponding to the frame period are acquired, an aging-up cleanup is performed on the buffers corresponding to each region to update the buffers. i The position parameters are stored in the buffer. In the next frame period, the position parameters of objects within the detection range that meet the first condition will continue to be stored in the buffer. i middle.

[0080] For example, to maintain the buffer i The location parameters stored in the buffer are contiguous, and each time only the buffer is targeted. i The location parameters with the smallest and largest values ​​are aged and cleared respectively. It should be noted that at this time, the buffer... i The location parameters stored in the buffer have been sorted according to preset rules. Generally, the location parameter with the smallest value is in the buffer. i The index in is 1. Specifically, first determine the buffer. i Whether it is empty, i.e., size i Is it greater than 0? (In size) i If the value is greater than 0, clear the buffer. i The position parameter with the largest value in the middle is then determined, and the buffer is checked again at this point. i Is it empty? Furthermore, in size... i If the value is greater than 0, the position parameter with the largest value at that time is overwritten and stored at the position of the position parameter with the smallest value, and then another position parameter with the largest value (with a larger position index) is cleared to reduce the buffer size. i The repetition rate of the position parameters in the middle. This can be understood as the process of clearing the original buffer during a cleaning cycle (or aging cycle). i At least one position parameter with a large value in the buffer i If the buffer is not empty, the second-largest position parameter is used to age the first-smallest position parameter to ensure the cleaned buffer... i The continuity of position parameters.

[0081] This disclosure provides a method for identifying target objects based on vehicle-mounted radar. The target objects are characterized by being stationary relative to the ground and having a continuous structure along the road's extension direction, such as fixed road infrastructure like guardrails. Compared to existing technologies, this method considers complex and varied road scenarios by dividing the detection range of the vehicle-mounted radar into several regions and configuring a buffer for each region to store the position parameters of multiple stationary objects within that region. Furthermore, based on the position parameters of the objects stored in the buffers of each region, and combining the second condition and verification condition, it is determined whether a target object satisfying the aforementioned characteristics exists within the detection range of the vehicle-mounted radar. The method provided by this disclosure can accurately identify target objects. Specifically, for road public facilities such as guardrails, even if the target object has structural defects within a small area during vehicle operation, such as openings in the guardrail, the method provided by this disclosure can determine that the target object is continuously continuous along the road's extension direction. Therefore, the vehicle's driver assistance system will not treat guardrails with fixed openings detected by the vehicle-mounted radar as obstacles and issue a warning to the driver through the vehicle's control center. The method provided in this disclosure helps to reduce the false alarm rate of vehicle driver assistance systems.

[0082] On the other hand, to achieve the above objectives, this disclosure provides a device for identifying target objects based on vehicle-mounted radar, such as... Figure 6 As shown, the device 6 includes: a first processing unit 61, a first processing unit 62, a first processing unit 63, and a first processing unit 64.

[0083] Wherein: the first processing unit 61 is used to receive the detection data of the vehicle radar within a preset spatial range; the second processing unit 62 is used to determine whether there is an object satisfying a first condition within the preset spatial range based on the detection data; the third processing unit 63 is used to further determine whether the position parameters of the object satisfy a second condition when there is an object satisfying the first condition within the preset spatial range; the fourth processing unit 64 is used to determine that the target object exists in the external space when the position parameters of the object satisfy the second condition.

[0084] In some embodiments of this disclosure, the first condition is that the object's velocity relative to the ground is zero; the second condition is that the position parameters of preferred objects that satisfy the first condition and are located in different regions within a preset number of frames satisfy a first numerical relationship, or a second numerical relationship, or a third numerical relationship. The target object is characterized by being stationary relative to the ground and having a continuous structure in the direction of road extension, such as a guardrail or other fixed road infrastructure.

[0085] In some embodiments of this disclosure, the first processing unit 61 acquires a frame of detection data of the vehicle-mounted radar within the preset spatial range every preset time interval, wherein the detection data corresponding to each frame includes a dot pattern that corresponds one-to-one with the object detected in the current frame.

[0086] In some embodiments of this disclosure, for each frame of detection data, the second processing unit 62 determines whether the object corresponding to each point in the frame satisfies the first condition based on the trajectory corresponding to each point in the frame.

[0087] In some embodiments of this disclosure, the first processing unit 61 is further configured to divide the preset spatial range into at least two regions; and to configure a buffer of preset capacity for each region to store the position parameters of the detected objects associated with the corresponding region; and to initialize the buffer corresponding to each region.

[0088] In some embodiments of this disclosure, if an object corresponding to at least one dot in the current frame satisfies the first condition, the second processing unit 62 is further configured to store the position parameters of the object that satisfies the first condition into a buffer corresponding to the corresponding region.

[0089] In some embodiments of this disclosure, the third processing unit 63 is used to sort the position parameters stored in the buffer corresponding to each region according to a preset rule and extract at least one position parameter from the buffer corresponding to each region; and to determine whether the position parameter extracted from the buffer corresponding to each region satisfies the second condition.

[0090] In some embodiments of this disclosure, if the position parameters extracted from the buffer corresponding to each of the regions satisfy the second condition, the fourth processing unit 64 is used to determine that the object corresponding to the position parameters that satisfy the second condition is the target object.

[0091] In some embodiments of this disclosure, after the target object is determined, in order to ensure that the identified target object is the correct one, the fourth processing unit 64 is further configured to verify the position parameters of the identified target object based on a verification condition. The verification condition is that the position parameters of the target object and the position parameters of objects within the preset spatial range that do not satisfy the first condition satisfy a fourth numerical relationship.

[0092] In some embodiments of this disclosure, in order to update the location parameters stored in the cache corresponding to each region in a timely manner, the location parameters in the cache corresponding to each region can be cleaned and aged according to a preset period.

[0093] Other aspects of the apparatus for identifying target objects based on vehicle-mounted radar provided in this disclosure are the same as or similar to the method for identifying target objects based on vehicle-mounted radar described above (for example, the expression and meaning of the first numerical relationship, the second numerical relationship, the third numerical relationship, and the fourth numerical relationship are the same as or similar to those described above in conjunction with the embodiments of the method for identifying target objects based on vehicle-mounted radar), and will not be repeated here.

[0094] This disclosure also provides a computer-readable storage medium for storing processor-executable instructions, wherein the processor-executable instructions stored in the computer-readable storage medium, when executed, can cause the processor to perform any of the methods for identifying target objects based on vehicle radar as described above.

[0095] Those skilled in the art should understand that all or part of the processes in the above-described embodiments of the method for identifying target objects based on vehicle-mounted radar can be implemented by hardware and / or software related to computer program instructions. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can perform the processes in the above-described method embodiments. Any references to memory, storage, databases, or other media used in the embodiments provided in this disclosure can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), enhanced SDRAM (ESDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0096] This disclosure provides a method, apparatus, and storage medium for identifying target objects based on vehicle-mounted radar. The target object is characterized by being stationary relative to the ground and having a continuous structure along the road's extension direction, such as fixed road infrastructure like guardrails. This disclosure enables accurate identification of target objects, helping to reduce the false alarm rate of vehicle driver assistance systems. Specifically, for road public facilities like guardrails, even if the target object has structural defects within a small area during vehicle operation, such as openings in the guardrail, the method and apparatus provided in this disclosure can determine that the target object is entirely continuous along the road's extension direction. Therefore, the vehicle's driver assistance system will not treat guardrails with fixed openings detected by the vehicle-mounted radar as obstacles and issue unnecessary warnings to the driver through the vehicle's control center.

[0097] The method, apparatus, and storage medium for identifying target objects based on vehicle-mounted radar provided in the embodiments of this disclosure have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this disclosure. The descriptions of the above embodiments are only for the purpose of helping to understand the method and its core ideas. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this disclosure. Therefore, the content of this specification should not be construed as a limitation of this disclosure.

Claims

1. A method for identifying a target object based on a vehicle-mounted radar, characterized in that, The method comprises: (S100): receiving detection data of the vehicle-mounted radar in a preset spatial range; (S200): determining whether an object meeting a first condition exists in the preset spatial range according to the detection data, wherein the first condition is that the object has a ground speed of zero; (S300): when the object meeting the first condition exists in the preset spatial range, further determining whether a position parameter of the object meets a second condition, wherein the second condition is that position parameters of objects meeting the first condition and detected in different regions in a preset number of frames meet a first numerical relationship, or a second numerical relationship, or a third numerical relationship; the different regions are obtained by dividing the preset spatial range, and the different regions include a region one, a region two, and a region three; (S400): when the position parameter of the object meets the second condition, determining that the target object exists in an external space of the vehicle; After determining that the object corresponding to the position parameter meeting the second condition is the target object, the (S400) further comprises verifying the position parameter of the target object based on a verification condition; The second condition comprises: y2 is a valid value within a reasonable distance range of the stationary object and the vehicle-mounted radar in a current detection environment, and y1, y2, and y3 satisfy a first numerical relationship, the first numerical relationship being: |y1-y2|<threshold value, and / or |y3-y2|<threshold value; a lateral distance y between the stationary object and the vehicle-mounted radar i , i=[1, 2, 3], y1 is a lateral distance between the stationary object and the vehicle-mounted radar in the region one, y2 is a lateral distance between the stationary object and the vehicle-mounted radar in the region two, and y3 is a lateral distance between the stationary object and the vehicle-mounted radar in the region three. In a case where y2 is not a valid value and / or y1, y2, y3 do not meet the first numerical relationship, a second numerical relationship is met between y2 and a preset value, and the second numerical relationship is | preset value-y2 | < threshold value; In a case where y2 is not a valid value and / or y1, y2, y3 do not meet the first numerical relationship, and y2 does not meet the second numerical relationship, at least one of y1, y2, and y3 meets a third numerical relationship with the preset value, and the third numerical relationship is y-preset value < threshold value, wherein y is at least one of y1, y2, and y3.

2. The method of claim 1, wherein, The (S100) comprises: acquiring detection data of the vehicle-mounted radar in the preset spatial range every preset time, wherein each frame of detection data comprises a track corresponding to an object detected in the frame.

3. The method of claim 2, wherein the target object is identified based on a radar mounted on the vehicle. The (S200) comprises: For each frame of detection data, determining whether an object corresponding to each track in the frame meets the first condition based on the track.

4. The method of claim 1, wherein the target object is identified based on a radar mounted on the vehicle. The (S100) comprises: configuring a buffer with a preset capacity for each region to store a position parameter of an object detected in the corresponding region.

5. The method of claim 4, wherein the target object is identified based on a radar mounted on the vehicle. In a case where at least one object corresponding to at least one track in the current frame meets the first condition, the (S200) comprises: storing the position parameter of the object meeting the first condition in the buffer corresponding to the corresponding region.

6. The method of claim 5, wherein the target object is identified based on a radar mounted on the vehicle. The (S300) comprises: sorting the position parameters stored in the buffer corresponding to each region according to a preset rule and extracting at least one position parameter from the buffer corresponding to each region; determining whether the position parameter extracted from the buffer corresponding to each region meets the second condition.

7. The method of claim 6, wherein the target object is identified based on a radar mounted on the vehicle. In a case where the position parameter extracted from the buffer corresponding to each region meets the second condition, the (S400) comprises: determining that an object corresponding to a position parameter satisfying the second condition is the target object; cleaning the position parameter in the buffer corresponding to each of the regions according to a preset period.

8. The method of claim 1, wherein the target object is identified based on a radar mounted on the vehicle. The verification condition is that the position parameter of the target object and the position parameter of an object in the preset space range that does not satisfy the first condition satisfy a fourth numerical relationship.

9. The method of claim 1, wherein the target object is identified based on a radar mounted on the vehicle. The position parameter is a lateral distance between the object and the vehicle-mounted radar.

10. The method of claim 1, wherein the target object is identified based on a radar mounted on the vehicle. The target object is a road guardrail.

11. A device for identifying target objects based on vehicle-mounted radar, characterized in that, The device comprises: a first processing unit configured to receive detection data of the vehicle-mounted radar in a preset space range; a second processing unit configured to determine, according to the detection data, whether there is an object satisfying a first condition in the preset space range, wherein the first condition is that the object has a ground speed of zero; a third processing unit configured to, when there is an object satisfying the first condition in the preset space range, further determine whether a position parameter of the object satisfies a second condition, wherein the second condition is that position parameters of objects satisfying the first condition and located in different regions detected in a preset number of frames satisfy a first numerical relationship, or a second numerical relationship, or a third numerical relationship; the different regions are obtained by dividing the preset space range, and the different regions include a region one, a region two and a region three; a fourth processing unit configured to determine that there is the target object in the outside space when the position parameter of the object satisfies the second condition; The fourth processing unit is further configured to verify the position parameter of the target object based on a verification condition. The second condition comprises: y2 is a valid value within a reasonable distance range of the stationary object and the vehicle-mounted radar in a current detection environment, and y1, y2, and y3 satisfy a first numerical relationship, the first numerical relationship being: |y1-y2|<threshold value, and / or |y3-y2|<threshold value; a lateral distance y between the stationary object and the vehicle-mounted radar i , i=[1, 2, 3], y1 is a lateral distance between the stationary object and the vehicle-mounted radar in the region one, y2 is a lateral distance between the stationary object and the vehicle-mounted radar in the region two, and y3 is a lateral distance between the stationary object and the vehicle-mounted radar in the region three. In a case where y2 is not a valid value and / or y1, y2 and y3 do not satisfy the first numerical relationship, a second numerical relationship is satisfied between y2 and a preset value, and the second numerical relationship is |preset value-y2|<threshold value. In a case where y2 is not a valid value and / or y1, y2 and y3 do not satisfy the first numerical relationship, and y2 does not satisfy the second numerical relationship, at least one of y1, y2 and y3 satisfies a third numerical relationship with the preset value, and the third numerical relationship is y-preset value<threshold value, wherein y is at least one of y1, y2 and y3.

12. The apparatus for identifying a target object based on a vehicle-mounted radar according to claim 11, wherein, The first processing unit acquires detection data of the vehicle-mounted radar in the preset space range every preset time, wherein the detection data of each frame includes a track corresponding to each object detected in the frame.

13. The apparatus for identifying a target object based on a vehicle-mounted radar according to claim 12, wherein, For the detection data of each frame, the second processing unit determines, based on a track corresponding to each track in the frame, whether an object corresponding to each track in the frame satisfies the first condition.

14. The apparatus for identifying a target object based on a vehicle-mounted radar according to claim 11, wherein, The first processing unit is further configured to configure a buffer with a preset capacity for each of the regions, and the buffer is used to store a position parameter of an object detected in the corresponding region.

15. The apparatus for identifying a target object based on a vehicle-mounted radar according to claim 14, wherein, In a case where an object corresponding to at least one track in the current frame satisfies the first condition, the second processing unit is further configured to store the position parameter of the object satisfying the first condition in the buffer corresponding to the corresponding region.

16. The device for identifying a target object based on a vehicle-mounted radar according to claim 15, characterized in that, The third processing unit is configured to sort the position parameters stored in the buffer corresponding to each of the regions according to a preset rule, and extract at least one position parameter from each of the buffers corresponding to the regions, respectively. And determine whether the position parameters extracted from the buffer corresponding to each of the regions satisfy the second condition.

17. The apparatus for identifying a target object based on a vehicle-mounted radar according to claim 16, wherein, In a case where the position parameters extracted from the buffer corresponding to each of the regions satisfy the second condition, the fourth processing unit is configured to determine that the object corresponding to the position parameters satisfying the second condition is the target object, and clean the position parameters in the buffer corresponding to each of the regions according to a preset period.

18. The apparatus for identifying a target object based on a vehicle-mounted radar according to claim 11, wherein, The verification condition is that the position parameters of the target object and the position parameters of the objects in the preset space range that do not satisfy the first condition satisfy a fourth numerical relationship. 19.A computer readable storage medium storing processor-executable instructions, the processor-executable instructions stored in the computer readable storage medium, when executed, cause a processor to implement the method of identifying a target object based on a vehicle-mounted radar according to any one of claims 1-10.

Images