A fence recognition method, system, and electronic device

Abstract

The present application belongs to the field of vehicle-mounted millimeter wave radar data processing, and particularly relates to a fence identification method, system and electronic device. The fence identification method comprises: obtaining a radar view of a vehicle radar, and the radar view comprising a plurality of point clouds; establishing a coordinate system for the radar view with the vehicle radar as the origin and the vehicle head direction as the positive direction of the longitudinal coordinate; dividing the coordinate system into a plurality of unit intervals along the transverse coordinate axis direction, and any unit interval being partially overlapped with its adjacent unit interval; screening out a candidate fence according to the number of point clouds falling into the unit interval; and identifying a target fence from the candidate fence through the transverse coordinates of the point clouds. The fence identification method of the present application can identify the target fence without complex traversal and repeated calculation, and has high reliability.

Description

Technical Field

[0001] This invention belongs to the field of vehicle-mounted millimeter-wave radar data processing, specifically relating to a fence recognition method, system, and electronic device. Background Technology

[0002] Fence identification has always been a crucial requirement, removing mirrored targets and preventing their prolonged presence that leads to frequent ID switching between the primary and mirrored targets. Fence identification relies on the fact that a fence is a series of dense points, with most fences being straight or slightly angled. This allows for fence extraction using methods that find straight lines. Classic methods for finding straight lines include the Hough transform and the Ransac method. However, these methods are computationally intensive and unsuitable for real-time line extraction.

[0003] In actual road testing, if the target vehicle moves along a fence, a false target often appears outside the fence. Since the mirrored target moves at the same speed and with a similar trajectory to the real target, traditional methods cannot be used to remove it. However, the mirrored target and its corresponding real target are symmetrical about the fence; that is, they have the same lateral distance, longitudinal distance, and longitudinal speed. Mirrored targets can be removed based on these three characteristics, but only if the fence is located. Summary of the Invention

[0004] In view of the shortcomings of the prior art described above, the purpose of this invention is to provide a fence recognition method that is simple to calculate, has a concise process, and is more practical.

[0005] To achieve the above and other related objectives, the present invention provides a fence recognition method, comprising: acquiring a radar view of a vehicle radar, wherein the radar view includes a plurality of point clouds; establishing a coordinate system for the radar view with the vehicle radar as the origin and the vehicle's heading as the positive ordinate; dividing the coordinate system into a plurality of unit intervals along the horizontal axis, wherein any one of the unit intervals partially overlaps with its adjacent unit intervals; selecting candidate fences based on the number of point clouds falling into the unit intervals; and identifying a target fence from the candidate fences using the horizontal coordinates of the point clouds.

[0006] According to a specific embodiment of the present invention, the step of selecting candidate fences based on the number of point clouds falling into the unit interval includes: determining whether the point cloud falls into the unit interval based on the horizontal coordinate of the point cloud, and dividing the point cloud into the corresponding unit intervals; counting the number of point clouds in each unit interval, and determining whether the unit interval is a candidate fence.

[0007] According to a specific embodiment of the present invention, the step of counting the number of point clouds in each unit interval and determining whether the unit interval is a candidate fence includes: if the number of point clouds in the unit interval exceeds a preset first threshold, and the number of point clouds in the unit interval exceeds the number of point clouds in the preceding unit interval and the following unit interval, then the unit interval is used as the candidate fence.

[0008] According to a specific embodiment of the present invention, the step of obtaining the candidate fence further includes: identifying the interval length of the unit interval that serves as the candidate fence; if the difference between the interval length and the fence width is less than a preset second threshold, then the unit interval is regarded as the candidate fence; if the difference between the interval length and the fence width is greater than the preset second threshold, then the point cloud in the unit interval is filtered, and the filtered point cloud is regarded as the candidate fence.

[0009] According to a specific embodiment of the present invention, the step of filtering the point cloud in the unit interval and considering the filtered point cloud as the candidate fence if the difference between the interval length and the fence width is greater than a preset second threshold includes: calculating a first mean value of the horizontal coordinate of the point cloud in the unit interval; and using the point cloud corresponding to the absolute value of the difference between the horizontal coordinate and the first mean value being less than a preset third threshold as the candidate fence.

[0010] According to a specific embodiment of the present invention, the step of identifying the target fence from the candidate fences by the abscissa of the point cloud includes: calculating a second mean of the abscissa of the point cloud in each candidate fence; for candidate fences located in the negative direction of the origin, taking the candidate fence corresponding to the largest second mean as the target fence; and / or, for candidate fences located in the positive direction of the origin, taking the candidate fence corresponding to the smallest second mean as the target fence.

[0011] According to a specific embodiment of the present invention, the method further includes: traversing all point clouds that are not the candidate fences, filtering out mirror point clouds and deleting them.

[0012] According to a specific embodiment of the present invention, the step of traversing all point clouds that are not the candidate fences, filtering out mirror point clouds and deleting them includes: detecting whether there are two symmetrical point clouds on both sides of the target fence; if there are two symmetrical point clouds on the left and right sides of the target fence, and the difference in the absolute value of the vertical coordinate of the two symmetrical point clouds is less than a fourth threshold, the difference in the absolute value of the horizontal coordinate is less than a fifth threshold, and the difference in the absolute value of the vertical velocity is less than a sixth threshold, then the point cloud that is farthest from the origin among the two symmetrical point clouds is taken as the mirror point cloud; and the mirror point cloud is deleted.

[0013] A fence recognition system includes: a data acquisition module for acquiring a radar view of a vehicle radar, wherein the radar view includes a plurality of point clouds; a coordinate system establishment module for establishing a coordinate system of the radar view with the vehicle radar as the origin and the vehicle's heading as the positive ordinate; a region division module for dividing the coordinate system into a plurality of unit intervals along the horizontal axis, wherein any one of the unit intervals partially overlaps with its adjacent unit intervals; a filtering module for filtering candidate fences based on the number of point clouds falling into the unit intervals; and a recognition module for recognizing the target fence from the candidate fences using the horizontal coordinate of the point clouds.

[0014] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of any of the methods described above.

[0015] The technical advantage of this invention lies in its ability to identify candidate fences by dividing the data into multiple unit intervals and counting the number of point clouds within each interval. Furthermore, the target fence is selected from the candidate fences based on the x-coordinates of the point clouds. This process only requires counting the number of point clouds, eliminating the need for complex traversal and repeated calculations. Moreover, the target fence identified through comparison and statistics is approximately the same as the actual fence location, making it more practical and reliable.

[0016] In addition, after obtaining the target fence, the present invention also extracts and deletes its mirror point cloud based on the target fence, which can prevent the mirror target from existing for a long time and causing frequent switching between the IDs of the main target and the mirror target. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating a specific embodiment of the fence recognition method provided by the present invention.

[0018] Figure 2 This is a schematic diagram of a specific embodiment of the fence recognition method provided by the present invention for dividing unit intervals in a road application;

[0019] Figure 3 This is a schematic diagram of a specific embodiment of the fence recognition method provided by the present invention for counting the number of point clouds in each unit interval in a road application;

[0020] Figure 4 This is a schematic diagram of another specific embodiment of the fence recognition method provided by the present invention for counting the number of point clouds in each unit interval in a road application;

[0021] Figure 5 This is a flowchart illustrating a specific embodiment of the fence recognition system provided by the present invention.

[0022] Figure 6This is a structural block diagram of a specific embodiment of the electronic device provided by the present invention. Detailed Implementation

[0023] The embodiments of the present invention will be described below with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are only for illustrating the present invention and not for limiting the scope of protection of the present invention.

[0024] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0025] In the following description, numerous details are explored to provide a more thorough explanation of embodiments of the invention. However, it will be apparent to those skilled in the art that embodiments of the invention may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring embodiments of the invention.

[0026] First, it should be noted that, in order to enable those skilled in the art to better understand the solution of this application, the technical solutions in the embodiments of this application will be clearly and completely described.

[0027] Fence recognition and extraction has always been a crucial requirement in road driving. It involves identifying fences on the road and extracting their mirror targets to prevent frequent ID switching between the primary and mirror targets due to the prolonged presence of mirror targets. In this embodiment, fence recognition and extraction are based on an onboard millimeter-wave radar. Typically, fences appear as a series of dense dots on the radar, with most fences being straight or slightly tilted. Straight line identification methods can be used to identify and extract fences. Classic straight line identification methods include Hough transform and Ransac method, but these are computationally intensive and unsuitable for real-time straight line extraction.

[0028] Therefore, in this embodiment, a statistical method is used to find the straight line, based on the vehicle traveling along the direction of the fence, which is perpendicular to the radar. The fence consists of a series of point clouds. By dividing several point clouds into regions according to their lateral distance, the number of point clouds in each region is counted. When the number of point clouds in a region reaches a preset threshold, the region is determined to be the area where the fence is located. Then, the position of the fence is obtained by calculating the lateral distance of the point clouds, and its mirror target is extracted and deleted.

[0029] The methods described above generally involve an onboard millimeter-wave radar collecting data and transmitting it to an onboard terminal for execution, in order to achieve the final purpose and effect.

[0030] Example 1

[0031] Please see Figure 1 As shown, a fence recognition method includes:

[0032] Step S10: Obtain the radar view of the vehicle radar, wherein the radar view includes a number of point clouds.

[0033] Specifically, in applications, millimeter-wave radar is installed at the very front of the vehicle to collect data on road barriers. Roads typically have multiple lanes, each separated by barriers, so barriers are located on either side of the lane a vehicle is currently traveling in. By identifying the barriers on either side of the current lane, the distance between the vehicle and the barriers can be calculated, and other operations can be performed. Barriers are usually composed of several metal or plastic railings, and when a millimeter-wave radar scans one of these railings, it displays as a point cloud. Therefore, a barrier typically appears as a series of point clouds on a radar image. Simultaneously, when the millimeter-wave radar scans other vehicles and roadside obstacles, it also displays as a point cloud. Therefore, the radar image includes not only the point cloud of the barrier but also interference from other point clouds. This necessitates the identification of the point clouds in the radar image to pinpoint the target barrier. Furthermore, due to the refraction effect of millimeter-wave radar, the point cloud displayed for one barrier may contain a mirror image. By filtering out and deleting the mirror image of the barrier from the existing point cloud, interference and impact can be avoided.

[0034] Step S20: Establish a coordinate system for the radar view with the vehicle radar as the origin and the vehicle's front orientation as the positive vertical axis.

[0035] Step S30: Divide the coordinate system into several unit intervals along the horizontal axis, and any one of the unit intervals partially overlaps with its adjacent unit intervals.

[0036] Specifically, in application, the lateral distance between the point cloud and the radar is used to divide the area. A unit interval is defined at certain intervals according to actual needs, and each unit interval is set without gaps, so that the divided unit intervals can cover the area where the point cloud is located. If the horizontal coordinate of the point cloud lies between the horizontal coordinates of the left and right boundaries of a unit interval, then the point cloud is identified as a point within that unit interval. Furthermore, to avoid point clouds being distributed on the boundaries of unit intervals, any given unit interval partially overlaps with its adjacent unit interval, thus preventing unidentified points from being distributed on the boundaries. In this embodiment, the length of each unit interval can also be set differently depending on the actual situation. Each unit interval can be of the same length or different lengths, only requiring one unit interval to overlap with the preceding unit interval to cover the area where the point cloud is located and prevent point clouds distributed on the boundaries of unit intervals from failing to be successfully identified and extracted.

[0037] Step S40: Select candidate fences based on the number of point clouds falling into the unit interval.

[0038] Specifically, firstly, each point cloud is divided into different unit intervals based on its x-coordinate, and the number of point clouds in each unit interval is counted. Secondly, candidate fences are selected from the unit intervals, thus excluding non-fence intervals. Since there are multiple fences on the road, there are multiple candidate fences, while the actual target fence may only be one or two. Therefore, selection needs to be based on the actual situation. The specific steps for selecting candidate fences include: when the number of point clouds in a unit interval exceeds a preset first threshold, and the number of point clouds in that unit interval exceeds the number of point clouds in the preceding and following intervals, the unit interval is considered a candidate fence, and the preset threshold is set based on actual experience.

[0039] Furthermore, considering that in step S30, when dividing the coordinate system into several unit intervals along the horizontal axis, two situations may occur due to the different set lengths of the unit intervals. One is that the set length of the unit interval is close to the width of the fence on the actual road. In this case, the target fence can be directly identified from each candidate fence. That is, if the difference between the interval length and the fence width is less than the preset second threshold, the candidate fence that meets the condition is regarded as the target fence. The other is that the set length of the unit interval is much greater than the width of the fence on the actual road. In this case, there are interfering point clouds in the point cloud of the candidate fence, such as point clouds formed by vehicles on the road or point clouds formed by road obstacles. Therefore, it is necessary to filter the point cloud in the candidate fence to obtain the point cloud that truly constitutes the candidate fence. Specifically, firstly, the first mean of the horizontal coordinates of the point clouds in the candidate fences is calculated; and the point clouds corresponding to the absolute value of the difference between the horizontal coordinate and the first mean is less than a preset third threshold are taken as the candidate fences, that is, the point clouds close to the first mean are taken as the point clouds constituting the candidate fences, and then the process proceeds to step S50 to identify the target fence from each candidate fence.

[0040] Step S50: Identify the target fence from the candidate fences using the horizontal coordinates of the point cloud.

[0041] Specifically, since there are multiple fences along the road, the target fences are obtained according to actual needs. In this embodiment, the target fences are the fences on both sides of the current lane; therefore, the two fences closest to the radar are the target fences. However, there may also be cases where the current driving lane only has a fence on one side; therefore, only one side of the fence can be obtained.

[0042] First, the second mean of the x-coordinates of the point cloud in each of the candidate fences is calculated. Second, for candidate fences located in the negative direction from the origin, the candidate fence corresponding to the largest second mean is selected as the target fence, i.e., the left fence of the current driving lane. Simultaneously and / or, for candidate fences located in the positive direction from the origin, the candidate fence corresponding to the smallest second mean is selected as the target fence, i.e., the right fence of the current driving lane. The final target fences obtained are the fences on both sides of this lane.

[0043] Step S60: Traverse all point clouds that are not the candidate fences, filter out the mirror point clouds and delete them.

[0044] After acquiring the target fence, to avoid frequent switching between the main target and the mirror target ID due to the prolonged existence of the mirror target, the mirror point cloud of the target fence is extracted and deleted to achieve the final goal. Specifically, it is detected whether there are two similar point clouds on both sides of the target fence. Since the point cloud in the interval where the target fence is located is considered to constitute the point cloud of the target fence, the point clouds on both sides of the target fence are located in other intervals. If there are two point clouds on both sides of the target fence, and the difference between the ordinates of the two point clouds is less than a first threshold, the difference between the absolute values ​​of the abscissas is less than a second threshold, and the difference between the longitudinal velocities is less than a third threshold, then these two point clouds are considered as mirror point clouds of the target fence. The millimeter-wave radar acquires the longitudinal velocity of the point clouds while simultaneously acquiring their coordinates.

[0045] By identifying and deleting the mirror point cloud of the target fence, frequent switching between the mirror target and the main target ID can be avoided.

[0046] It should be noted that the steps of the various methods described above are only for clarity. In practice, they can be combined into one step or some steps can be split into multiple steps. As long as they contain the same logical relationship, they are all within the scope of protection of this patent. Adding insignificant modifications or introducing insignificant designs to the algorithm or process, but without changing the core design of the algorithm and process, are also within the scope of protection of this patent.

[0047] Example 2

[0048] Please see Figures 2-4 As shown, this application also provides a specific embodiment of a fence recognition method, such as... Figure 2 As shown, the radar is divided into unit intervals from left to right, namely D1 to D19. The length of each unit interval approximates the width of a fence on an actual road; that is, the horizontal coordinate area of ​​D1 is -1 meter to 1 meter, the horizontal coordinate area of ​​D2 is 0 meter to 2 meters, the horizontal coordinate area of ​​D3 is -2 meters to 0 meters, and so on. Here, the division of unit intervals in this embodiment is explained accordingly. To avoid overlapping intervals that would make the intervals in the diagram difficult to distinguish, the coordinate system is divided by two rows of unit intervals. This can be approximated as partially overlapping intervals, covering the entire coordinate system and preventing point cloud distribution on interval boundaries. Simultaneously, a preset threshold of 20 is set for the number of point clouds. The number of point clouds in each region from D1 to D19 is counted according to the above method. Figure 3 , 4As shown. According to statistics, the number of point clouds in D10 and D17 is greater than the preset threshold. Since the number of fences on both the left and right sides is less than 2, the fences in the D10 and D17 interval are the target fences. Further, based on the first lateral distance between the point clouds of D10 and D17, as shown in Tables 1 and 2 below, the second lateral distance of the fence is calculated to be D10 = 5.1 and D17 = -7.9.

[0049] Table 1. First lateral distance of point cloud within interval D10

[0050] 4.8 5.6 4.7 5.2 5.0 5.5 4.9 5.3 4.3 5.8 5.3 5.6 5.5 4.7 5.2 4.4 4.7 4.9 5.2 5.8 4.2 5.2 4.3 5.6

[0051] Table 1. First lateral distance of point cloud within interval D17

[0052] -7.6 -8.9 -8.2 -7.4 -7.9 -7.6 -8.0 -7.6 -7.8 -7.6 -7.2 -8.6 -7.6 -7.2 -8.6 -8.3 -8.6 -7.6 -7.7 -7.8 -7.9

[0053] Example 3

[0054] Please see Figure 5 As shown in the illustration, this application also provides a fence recognition system, including:

[0055] The data acquisition module 10 is used to acquire the radar view of the vehicle radar, and the radar view includes several point clouds.

[0056] The coordinate system establishment module 20 is used to establish a coordinate system for the radar view with the vehicle radar as the origin and the vehicle's front orientation as the positive direction of the vertical coordinate.

[0057] The region division module 30 is used to divide the coordinate system into several unit intervals along the horizontal axis, and any one of the unit intervals partially overlaps with the previous unit interval.

[0058] The filtering module 40 is used to filter out candidate fences based on the number of points in the point cloud that fall into the unit interval.

[0059] The filtering module further includes:

[0060] The point cloud statistics unit determines whether the point cloud falls within the unit interval based on the horizontal coordinate of the point cloud, and divides the point cloud into the corresponding unit intervals.

[0061] The fence determination unit counts the number of point clouds in each unit interval and determines whether the unit interval is a candidate fence.

[0062] The identification module 50 is used to identify the target fence from the candidate fences by using the horizontal coordinate of the point cloud.

[0063] The identification module further includes:

[0064] The calculation unit calculates the second mean of the x-coordinates of the point cloud in each of the candidate fences.

[0065] The fence identification unit, for candidate fences located in the negative direction of the origin, selects the candidate fence corresponding to the largest second mean as the target fence; and / or, for candidate fences located in the positive direction of the origin, selects the candidate fence corresponding to the smallest second mean as the target fence.

[0066] The system includes a candidate fence identification module and a mirror point cloud processing module. The candidate fence identification module determines the location of the candidate fence based on the set length of the unit interval. The mirror point cloud processing module traverses all point clouds that are not candidate fences, filters out mirror point clouds, and deletes them.

[0067] It should be noted that the fence recognition system provided in the above embodiments and the fence recognition method provided in Embodiment 1 belong to the same concept. The specific ways in which each module and unit performs its operation have been described in detail in the method embodiments and will not be repeated here. In practical applications, the fence recognition method provided in Embodiment 1 can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. This is not a limitation here.

[0068] Example 4

[0069] Please see Figure 6 As shown, embodiments of this application also provide an electronic device, including a memory 2, a processor 1, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of any of the methods described above.

[0070] The memory includes at least one type of readable storage medium, such as flash memory, portable hard drive, multimedia card, card-type memory (e.g., SD or DX memory), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory can be an internal storage unit of an electronic device, such as a portable hard drive. In other embodiments, the memory can be an external storage device of the electronic device, such as a plug-in portable hard drive, Smart Media Card (SMC), Secure Digital (SD) card, Flash Card, etc. Furthermore, the memory can include both internal and external storage units of the electronic device. The memory can be used not only to store application software and various types of data installed on the electronic device, but also to temporarily store data that has been output or will be output.

[0071] In some embodiments, a processor may be composed of integrated circuits, such as a single packaged integrated circuit or multiple integrated circuits packaged with the same or different functions. This includes combinations of one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips. The processor is the control unit of the electronic device, connecting various components of the device via various interfaces and lines. It executes programs or modules stored in the memory and calls data stored in the memory to perform various functions and process data within the electronic device.

[0072] The processor executes the operating system of the electronic device and various installed applications. The processor executes the applications to implement the steps in the above method embodiments.

[0073] For example, the computer program may be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the electronic device.

[0074] The integrated unit, implemented as a software functional module, can be stored in a computer-readable storage medium. This software functional module, stored in a storage medium, includes several instructions to cause a computer device (which may be a personal computer, computer equipment, or network device, etc.) or processor to execute some functions of the lithium battery cold solder joint detection method of the various embodiments of the present invention.

[0075] In summary, the technical advantage of this invention lies in its ability to identify candidate fences by dividing the data into multiple unit intervals and counting the number of point clouds within each interval. Furthermore, the target fence is selected from the candidate fences based on the x-coordinates of the point clouds. This only requires counting the number of point clouds, eliminating the need for complex traversal and repeated calculations to identify the target fence. Moreover, the target fence identified through comparison and statistics is approximately the same as the actual fence position, making it more practical and reliable.

[0076] In addition, after obtaining the target fence, the present invention also extracts and deletes its mirror point cloud based on the target fence, which can prevent the mirror target from existing for a long time and causing frequent switching between the IDs of the main target and the mirror target.

[0077] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. A fence recognition method, characterized in that, include: Obtain a radar view of the vehicle radar, wherein the radar view includes a plurality of point clouds; A coordinate system is established with the vehicle radar as the origin and the vehicle's front orientation as the positive vertical axis relative to the radar view. The coordinate system is divided into several unit intervals along the horizontal axis, and any one of the unit intervals partially overlaps with its adjacent unit intervals. The process involves selecting candidate fences based on the number of point clouds falling within the unit intervals, and includes the following steps: determining whether a point cloud falls within a unit interval based on its horizontal coordinate, and dividing the point cloud into corresponding unit intervals; counting the number of point clouds in each unit interval and determining whether the unit interval is a candidate fence; wherein, the interval length of the unit interval used as a candidate fence is identified; if the difference between the interval length and the fence width is less than a preset second threshold, the unit interval is considered a candidate fence; if the difference between the interval length and the fence width is greater than the preset second threshold, the point clouds in the unit interval are filtered, and the filtered point clouds are considered candidate fences. The target fence is identified from the candidate fences using the x-coordinate of the point cloud.

2. The fence recognition method according to claim 1, characterized in that, The step of counting the number of point clouds in each unit interval and determining whether the unit interval is a candidate fence includes: If the number of point clouds in the unit interval exceeds a preset first threshold, and the number of point clouds in the unit interval exceeds the number of point clouds in the preceding and following unit intervals, then the unit interval is used as the candidate fence.

3. The fence recognition method according to claim 1, characterized in that, The step of filtering the point cloud in the unit interval if the difference between the interval length and the fence width is greater than a preset second threshold, and considering the filtered point cloud as the candidate fence, includes: Calculate the first mean of the x-coordinates of the point cloud in the unit interval; The point cloud corresponding to the absolute value of the difference between the horizontal coordinate and the first mean is less than a preset third threshold is used as the candidate fence.

4. The fence recognition method according to claim 1, characterized in that, The step of identifying the target fence from the candidate fences using the x-coordinate of the point cloud includes: Calculate the second mean of the x-coordinate of the point cloud in each of the candidate fences; For candidate fences located in the negative direction of the origin, the candidate fence corresponding to the largest second mean is taken as the target fence. And / or, for candidate fences located in the positive direction of the origin, the candidate fence corresponding to the smallest second mean is taken as the target fence.

5. The fence recognition method according to claim 1, characterized in that, Also includes: Iterate through all point clouds that are not the candidate fences, filter out the mirror point clouds and delete them.

6. The fence recognition method according to claim 5, characterized in that, The step of traversing all point clouds that are not candidate fences, filtering out mirror point clouds and deleting them includes: Detect whether there are two symmetrical point clouds on both sides of the target fence: If there are two symmetrical point clouds on the left and right sides of the target fence, and the difference between the absolute values ​​of the vertical coordinates of the two symmetrical point clouds is less than the fourth threshold, the difference between the absolute values ​​of the horizontal coordinates is less than the fifth threshold, and the difference between the absolute values ​​of the vertical velocity is less than the sixth threshold, then the point cloud that is farthest from the origin among the two symmetrical point clouds is taken as the mirror point cloud. Delete the mirrored point cloud.

7. A fence recognition system, characterized in that, include: The data acquisition module is used to acquire the radar view of the vehicle radar, and the radar view includes several point clouds; The coordinate system establishment module is used to establish a coordinate system with the vehicle radar as the origin and the vehicle's front orientation as the positive direction of the vertical coordinate for the radar view. The region division module is used to divide the coordinate system into several unit intervals along the horizontal axis, and any one of the unit intervals partially overlaps with its adjacent unit intervals. A filtering module is used to filter candidate fences based on the number of point clouds falling into the unit interval, and the steps include: determining whether the point cloud falls into the unit interval based on the horizontal coordinate of the point cloud, and dividing the point cloud into the corresponding unit intervals; counting the number of point clouds in each unit interval, and determining whether the unit interval is a candidate fence; wherein, the interval length of the unit interval that is a candidate fence is identified; if the difference between the interval length and the fence width is less than a preset second threshold, the unit interval is regarded as a candidate fence; if the difference between the interval length and the fence width is greater than the preset second threshold, the point clouds in the unit interval are filtered, and the filtered point clouds are regarded as candidate fences. The identification module is used to identify the target fence from the candidate fences by using the horizontal coordinate of the point cloud.

8. An electronic device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of the method according to any one of claims 1 to 6.

Images