Point cloud classification method and device, server and storage medium

Abstract

The invention provides a point cloud classification method and device, a server and a storage medium. The method comprises the following steps: acquiring initial point cloud data of a target road section and a target ground equation; based on a target ground equation, calculating a target distance from the initial point cloud data to the target ground equation; detecting whether the target distance is greater than a preset first distance threshold, and obtaining a first detection result; screening initial point cloud data of which the target distance is greater than a preset first distance threshold value from the first detection result as interference point cloud data; wherein the interference point cloud data comprises point cloud data corresponding to a fixed object in the target road section. Based on the above method, the interference point cloud data far away from the ground can be accurately screened and filtered by using the target ground equation, and the influence of the interference point cloud data on the tracking target point cloud data can be avoided subsequently.

Description

Technical Field

[0001] The present application relates to the technical field of point cloud data processing, and in particular to a point cloud classification method, device, server and storage medium. Background Art

[0002] With the development of sensor technology, millimeter-wave radars are widely used in urban transportation. In application scenarios such as urban intersections and highways, millimeter-wave radars are usually deployed at traffic lights and telephone poles for detection, collecting point cloud data of various observation objects. The point cloud data needs to be further classified and filtered, retaining the point cloud data of the vehicle for subsequent tracking, and filtering out the point cloud data of interference objects such as high-altitude crossbars, billboards, and tree crowns. The existing point cloud classification method is only applicable to the case where the millimeter-wave radar is installed horizontally and the ground is horizontal, so the applicability is poor. When the millimeter-wave radar is installed at an angle or the ground is inclined, the point cloud data of interference objects such as crossbars, billboards, and tree crowns cannot be accurately filtered out through the existing point cloud classification method. Therefore, when performing subsequent point cloud data processing, it will affect the tracking effect of the vehicle point cloud data, and thus affect the calculation accuracy of traffic information such as traffic volume, average speed, and vehicle headway. Summary of the invention

[0003] The present application provides a point cloud classification method, device, server and storage medium to solve the technical problem that the existing point cloud classification method cannot accurately filter out the point cloud data of interference objects, thereby affecting the tracking effect of the vehicle point cloud data.

[0004] In a first aspect, the present application provides a point cloud classification method, the method comprising: obtaining initial point cloud data and a target ground equation of a target road section; based on the target ground equation, calculating a target distance from the initial point cloud data to the target ground equation; detecting whether the target distance is greater than a preset first distance threshold, and obtaining a first detection result; from the first detection result, screening out the initial point cloud data whose target distance is greater than the preset first distance threshold as interference point cloud data; wherein the interference point cloud data includes point cloud data corresponding to fixed objects in the target road section.

[0005] In some embodiments, the initial point cloud data is obtained by scanning a target radar disposed above the target road section;

[0006] Correspondingly, after selecting the initial point cloud data whose target distance is greater than a preset first distance threshold from the first detection result as interference point cloud data, the method further includes:

[0007] A stop tracking instruction for the interference point cloud data is sent to the target radar.

[0008] In some embodiments, after detecting whether the target distance is greater than a preset first distance threshold and obtaining a first detection result, the method further includes:

[0009] From the first detection result, the initial point cloud data whose target distance is less than or equal to a preset first distance threshold is selected as the target point cloud data; wherein the target point cloud data includes the point cloud data corresponding to the vehicle in the target road section;

[0010] A tracking instruction for the target point cloud data is sent to the target radar.

[0011] In some embodiments, the target ground equation is pre-constructed as follows:

[0012] Acquire historical point cloud data of the target road section;

[0013] Based on the historical point cloud data, construct a plurality of first data sets; wherein the first data sets include three point cloud data randomly selected from the historical point cloud data;

[0014] respectively detecting whether the three point cloud data in the first data set are collinear, and obtaining a second detection result;

[0015] Filtering out the first data set whose second detection result is not collinear as the second data set;

[0016] Performing plane fitting using the second data set to determine a plurality of initial ground equations; wherein the initial ground equations correspond to the second data set in one-to-one correspondence;

[0017] Determining a fitting accuracy parameter for the initial ground equation according to the historical point cloud data;

[0018] An initial ground equation whose fitting accuracy parameter meets the requirements is selected as the target ground equation.

[0019] In some embodiments, determining a fitting accuracy parameter for the initial ground equation based on the historical point cloud data includes:

[0020] Calculating the vertical projection distances from the historical point cloud data to each initial ground equation respectively;

[0021] Detecting whether the vertical projection distance is less than a preset second distance threshold, and obtaining a second detection result;

[0022] Filtering out, from the second detection results, a vertical projection distance that is less than a preset second distance threshold as a third data set corresponding to the initial ground equation;

[0023] The number of data in each third data set is determined as a fitting accuracy parameter of each initial ground equation.

[0024] In some embodiments, respectively calculating the vertical projection distances of the historical point cloud data to each initial ground equation includes:

[0025] Inputting the horizontal coordinate data and the vertical coordinate data in the historical point cloud data into the initial ground equation to obtain the vertical projection data corresponding to the historical point cloud data;

[0026] The absolute value of the difference between the vertical coordinate data in the historical point cloud data and the vertical projection data is calculated as the vertical projection distance corresponding to the initial ground equation.

[0027] In some embodiments, selecting the initial ground equation whose fitting accuracy parameter meets the requirements as the target ground equation includes:

[0028] Arranging the fitting accuracy parameters in descending order to obtain a sorting result;

[0029] According to the ranking results, the fitting accuracy parameters with the highest ranking are screened out, and the initial ground equations corresponding to the fitting accuracy parameters with the highest ranking are determined as the target ground equations.

[0030] In a second aspect, the present application further proposes a point cloud classification device, the device comprising:

[0031] An acquisition module is used to obtain the initial point cloud data of the target road section and the target ground equation;

[0032] A calculation module, used for calculating a target distance from the initial point cloud data to the target ground equation based on the target ground equation;

[0033] A detection module, used to detect whether the target distance is greater than a preset first distance threshold, and obtain a first detection result;

[0034] A classification module is used to filter out the initial point cloud data whose target distance is greater than a preset first distance threshold from the first detection result as interference point cloud data; wherein the interference point cloud data includes point cloud data corresponding to fixed objects in the target section.

[0035] In a third aspect, the present application further proposes a server, comprising a processor and a memory for storing processor executable instructions, wherein the processor implements the steps of any of the above-mentioned point cloud classification methods when executing the instructions.

[0036] In a fourth aspect, the present application further proposes a storage medium on which computer instructions are stored, and when the instructions are executed by a processor, the steps of any of the above-mentioned point cloud classification methods are implemented.

[0037] The present application can achieve the following beneficial effects: the target ground equation is pre-constructed by plane fitting, and the target ground equation can accurately characterize the position information of the ground. Therefore, the interference point cloud data far away from the ground can be accurately screened and filtered out based on the target ground equation, and the interference point cloud data can be prevented from affecting the tracking target point cloud data subsequently. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying any creative work.

[0039] Figure 1 A schematic diagram of a flow chart of a point cloud classification method provided in an embodiment of the present application;

[0040] Figure 2 A schematic diagram of a flow chart of a point cloud classification method provided in an embodiment of the present application;

[0041] Figure 3 A schematic diagram of the structure of a point cloud classification device provided in an embodiment of the present application;

[0042] Figure 4 A hardware structure block diagram of a server provided in an embodiment of the present application. DETAILED DESCRIPTION

[0043] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0044] In the description of this application, it should be noted that, unless otherwise clearly specified and limited, the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article, unless otherwise specified, generally indicates that the associated objects before and after are in an "or" relationship.

[0045] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0046] Considering that the existing point cloud classification method usually classifies the point cloud data in combination with the installation height of the millimeter-wave radar, this method requires the installation height of the millimeter-wave radar to be known and the millimeter-wave radar to be installed horizontally, and the ground is regarded as a horizontal plane. Then, the ground position is calculated according to the installation height of the millimeter-wave radar, and then the height threshold is set to classify the point cloud data above the ground position and exceeding the height threshold as interference point cloud data for filtering. However, the existing point cloud classification method has poor applicability. When the installation of the millimeter-wave radar is tilted, or the ground is tilted, or the installation height of the millimeter-wave radar is unknown, the existing point cloud classification method cannot accurately filter out the point cloud data of interference objects such as crossbars, billboards, and tree crowns. Therefore, when performing subsequent point cloud data processing, it will affect the tracking effect of the vehicle point cloud data, and thus affect the calculation accuracy of traffic information such as traffic flow, average speed, and vehicle headway.

[0047] In view of the poor applicability and low accuracy of existing point cloud classification methods, and the inability to accurately divide interference point cloud data when the millimeter-wave radar is installed at an angle or the ground is tilted or the installation height of the millimeter-wave radar is unknown, the present application proposes a point cloud classification method, device, server and storage medium to overcome the problem of low point cloud data classification accuracy.

[0048] The following is an introduction to a point cloud classification method, device, server and storage medium provided by the present application in conjunction with the accompanying drawings.

[0049] In some embodiments, please refer to Figure 1 , the present application proposes a point cloud classification method, and in specific implementation, the method may include the following contents.

[0050] S101: Acquire initial point cloud data of a target road section and a target ground equation.

[0051] In some embodiments, the initial point cloud data is obtained by scanning a target radar deployed above the target road section.

[0052] In some embodiments, the target radar is specifically a millimeter wave radar.

[0053] In some embodiments, the format of the initial point cloud data is (x, y, z), where x represents horizontal coordinate data, y represents vertical coordinate data, and z represents vertical coordinate data.

[0054] In some embodiments, the target road section is scanned using a target radar to obtain initial point cloud data, which specifically includes ground point cloud data, vehicle point cloud data, and interference point cloud data (for example: crossbar point cloud data, billboard point cloud data, tree crown point cloud data, traffic light point cloud data, etc.).

[0055] In some embodiments, the target ground equation represents the position information of the ground in the target road segment. Figure 2 As shown, the target ground equation is pre-constructed as follows:

[0056] S201: Acquire historical point cloud data of the target road section.

[0057] In some embodiments, the historical point cloud data is static point cloud data collected within 5 minutes to 10 minutes in a historical time period; the static point cloud data includes point cloud data of objects whose positions do not change. Too much historical point cloud data is conducive to improving the fitting accuracy of the target ground equation, but it will increase the fitting calculation time. The present application sets the number of historical point cloud data to 20,000, which can not only ensure the fitting accuracy of the target ground equation, but also control the fitting calculation time to be shorter. The format of the historical point cloud data is also (x, y, z), where x represents the horizontal coordinate data, y represents the vertical coordinate data, and z represents the vertical coordinate data.

[0058] S202: constructing a plurality of first data sets based on the historical point cloud data; wherein the first data set includes three point cloud data randomly selected from the historical point cloud data.

[0059] In some embodiments, three point cloud data are randomly selected from the historical point cloud data to construct a first data set; the above steps are repeated to construct multiple first data sets; each first data set includes three randomly selected historical point cloud data.

[0060] S203: Detect whether the three point cloud data in the first data set are collinear respectively, and obtain a second detection result.

[0061] In some embodiments, since three collinear point cloud data cannot be plane-fitted, it is necessary to remove the collinear first data set and retain the non-collinear first data set as the second data set for subsequent plane fitting.

[0062] In some embodiments, taking the first data set M1 as an example, detecting whether three point cloud data in the first data set M1 are collinear to obtain the second detection result is exemplarily described.

[0063] Assume that the first data set M1 includes historical point cloud data A (x1, y1, z1), historical point cloud data B (x2, y2, z2), and historical point cloud data C (x3, y3, z3), where x1 represents the horizontal coordinate data of A, y1 represents the vertical coordinate data of A, z1 represents the vertical coordinate data of A, x2 represents the horizontal coordinate data of B, y2 represents the vertical coordinate data of B, z2 represents the vertical coordinate data of B, x3 represents the horizontal coordinate data of C, y3 represents the vertical coordinate data of C, and z3 represents the vertical coordinate data of C.

[0064] Determine the distance m1 between A and B using the following formula:

[0065]

[0066] Determine the distance m2 between A and C using the following formula:

[0067]

[0068] Determine the distance m3 between B and C using the following formula:

[0069]

[0070] Determine the semi-perimeter p of the triangle ABC formed by A, B, and C using the following formula:

[0071]

[0072] Determine the area S of triangle ABC using the following formula:

[0073]

[0074] After determining the area S of triangle ABC, detect whether the area S of triangle ABC is less than or equal to the first area threshold; when it is determined that the area S of triangle ABC is less than or equal to the first area threshold, determine that the second detection result of the first data set M1 is "collinear"; when it is determined that the area S of triangle ABC is greater than the first area threshold, determine that the second detection result of the first data set M1 is "non-collinear"; wherein, the first area threshold can be set to 0.0001 (unit: square meter), and this application does not limit the value of the first area threshold.

[0075] Based on the method described in the above embodiment, it is possible to detect in turn whether the three point cloud data in each first data set are collinear, and obtain a second detection result.

[0076] S204: Filter out the first data set whose second detection results are not collinear as the second data set.

[0077] In some embodiments, the collinear first data sets are eliminated, and the non-collinear first data sets are retained as the second data sets.

[0078] S205: Performing plane fitting using the second data set to determine a plurality of initial ground equations; wherein the initial ground equations correspond to the second data set in a one-to-one manner.

[0079] In some embodiments, the initial ground equation is set to ax+by+cz+d=0, where a represents the lateral fitting parameter, b represents the longitudinal fitting parameter, c represents the vertical fitting parameter, and d represents the constant fitting parameter. Plane fitting is performed using the three point cloud data in each second data set to obtain multiple groups of {a, b, c, d}, and then multiple initial ground equations can be determined; each second data set corresponds to an initial ground equation.

[0080] Based on the method described in the above embodiment, only three point cloud data are needed for each plane fitting. Compared with the method of using a large amount of point cloud data for plane fitting based on the least squares method in the prior art, the amount of data processing can be greatly reduced and the calculation speed can be improved.

[0081] S206: Determine a fitting accuracy parameter for the initial ground equation according to the historical point cloud data.

[0082] In some embodiments, determining a fitting accuracy parameter for the initial ground equation based on the historical point cloud data specifically includes:

[0083] S1: Calculate the vertical projection distance from the historical point cloud data to each initial ground equation respectively;

[0084] S2: Detect whether the vertical projection distance is less than a preset second distance threshold, and obtain a second detection result;

[0085] S3: Filtering out the vertical projection distances less than a preset second distance threshold from the second detection results as a third data set corresponding to the initial ground equation;

[0086] S4: Determine the number of data in each third data set as a fitting accuracy parameter of each initial ground equation.

[0087] In some embodiments, respectively calculating the vertical projection distances of the historical point cloud data to each initial ground equation includes:

[0088] S1: inputting the horizontal coordinate data and the vertical coordinate data in the historical point cloud data into the initial ground equation to obtain the vertical projection data corresponding to the historical point cloud data;

[0089] S2: Calculate the absolute value of the difference between the vertical coordinate data in the historical point cloud data and the vertical projection data as the vertical projection distance corresponding to the initial ground equation.

[0090] In some embodiments, the vertical projection distance z_diff_abs may be determined by the following formula:

[0091] z_diff_abs=|z-z_pro| (6)

[0092] Among them, z represents the vertical coordinate data, and z_pro represents the vertical projection data.

[0093] In some embodiments, the horizontal coordinate data, the vertical coordinate data, and the vertical projection data satisfy the initial ground equation.

[0094] In some embodiments, the method of calculating the vertical projection distance is exemplified by taking the historical point cloud data D (x4, y4, z4) as an example. First, the horizontal coordinate data x4 and the vertical coordinate data y4 in the historical point cloud data D are input into the initial ground equation ax+by+cz+d=0, and the vertical projection data z_pro that satisfies the initial ground equation is solved, that is, x4, y4, z_pro satisfy a·x4+b·y4+c·z_pro+d=0; then, the vertical projection distance z_diff_abs is determined according to the following formula:

[0095] z_diff_abs=|z4-z_pro| (7)

[0096] Among them, z4 represents the vertical coordinate data corresponding to the historical point cloud data D.

[0097] In some embodiments, for each initial plane equation, the vertical projection distances corresponding to different historical point cloud data are calculated respectively. It is detected whether the vertical projection distance is less than a preset second distance threshold, that is, whether z_diff_abs﹤thr is satisfied, where thr represents the preset second distance threshold, and thr can be 1 (unit: meter), and a second detection result is obtained; from the second detection result, the vertical projection distances less than the preset second distance threshold are screened out as the third data set corresponding to the initial ground equation; each initial ground equation corresponds to a third data set; the number of data in each third data set is determined as the fitting accuracy parameter of each initial ground equation.

[0098] In some embodiments, most of the historical point cloud data are distributed on the ground in the target road section, and most of the historical point cloud data are very close to the ground. Therefore, the number of data in the third data set can represent both the number of point clouds close to the initial ground equation corresponding to the third data set and the degree of proximity of the initial ground equation to the ground. The more data in the third data set, the larger the fitting accuracy parameter, indicating that the initial ground equation is closer to and fits the ground, and its fitting effect on the ground is better; the fewer data in the third data set, the smaller the fitting accuracy parameter, indicating that the initial ground equation is farther away from the ground, and its fitting effect on the ground is worse.

[0099] S207: Filter out the initial ground equation whose fitting accuracy parameters meet the requirements as the target ground equation.

[0100] In some embodiments, selecting the initial ground equation whose fitting accuracy parameter meets the requirements as the target ground equation specifically includes:

[0101] S1: Arrange the fitting accuracy parameters in descending order to obtain a sorting result;

[0102] S2: According to the ranking results, the fitting accuracy parameters with the highest ranking are screened out, and the initial ground equations corresponding to the fitting accuracy parameters with the highest ranking are determined as the target ground equations.

[0103] In some embodiments, the first-ranked fitting accuracy parameter may be screened out from the sorting results, and an initial ground equation corresponding to the first-ranked fitting accuracy parameter may be determined as the target ground equation.

[0104] In some embodiments, the target ground equation has the best fitting effect on the ground, can reflect the inclination of the ground, and can accurately characterize the position information of the ground.

[0105] In some embodiments, the target ground equation is of the form a_best·x+b_best·y+c_best·z+d_best=0.

[0106] S102: Based on the target ground equation, calculate the target distance from the initial point cloud data to the target ground equation.

[0107] In some embodiments, the target distance from the initial point cloud data to the target ground equation may be calculated according to the following formula:

[0108]

[0109] Among them, D goal Indicates the target distance, a_best indicates the lateral fitting parameter in the target ground equation, b_best indicates the longitudinal fitting parameter in the target ground equation, c_best indicates the vertical fitting parameter in the target ground equation, and d_best indicates the constant fitting parameter in the target ground equation.

[0110] Based on the method described in the above embodiment, the target distance from each initial point cloud data to the target ground equation is calculated respectively.

[0111] S103: Detect whether the target distance is greater than a preset first distance threshold, and obtain a first detection result.

[0112] In some embodiments, the first distance threshold may be set to 7 (unit: meter). The present application does not limit the value of the first distance threshold.

[0113] S104: Filter out the initial point cloud data whose target distance is greater than a preset first distance threshold from the first detection result as interference point cloud data; wherein the interference point cloud data includes point cloud data corresponding to fixed objects in the target section.

[0114] In some embodiments, initial point cloud data with a target distance greater than a preset first distance threshold, including point cloud data that is far away from the ground and has a fixed position, such as crossbar point cloud data, billboard point cloud data, tree crown point cloud data, traffic light point cloud data, telephone pole top point cloud data, noise point point cloud data, etc., are used as interference point cloud data.

[0115] In some embodiments, after screening out the initial point cloud data whose target distance is greater than a preset first distance threshold from the first detection result as interference point cloud data, the method further includes: sending a stop tracking instruction for the interference point cloud data to the target radar; the target radar is deployed above the target road section to collect initial point cloud data.

[0116] In some embodiments, after detecting whether the target distance is greater than a preset first distance threshold and obtaining a first detection result, the method further includes:

[0117] S1: Filtering out the initial point cloud data whose target distance is less than or equal to a preset first distance threshold from the first detection result as target point cloud data; wherein the target point cloud data includes point cloud data corresponding to the vehicle in the target road section;

[0118] S2: Sending a tracking instruction for the target point cloud data to the target radar.

[0119] In some embodiments, the target point cloud data also includes point cloud data corresponding to pedestrians in the target road section.

[0120] Based on the above embodiments, a target ground equation representing ground position information can be accurately constructed without knowing the installation height of the target radar. The ground can be either a horizontal plane or an inclined plane, and the initial point cloud data is divided into interference point cloud data and target point cloud data based on the distance from the initial point cloud data to the target ground equation. Furthermore, a tracking prohibition instruction is issued for the interference point cloud data, and a tracking instruction is issued for the target point cloud data. This can effectively reduce the data processing volume of real-time tracking of point cloud data, eliminate the influence of interference point clouds, and improve the accuracy of vehicle and pedestrian tracking, which has important engineering value for traffic radar applications.

[0121] In a specific scenario example, the present application may use the following steps and / or pseudo code to implement the above point cloud classification method:

[0122] 1. Obtain 20,000 historical point cloud data, recorded as pointdata dataset;

[0123] 2. Set the initial ground equation to ax+by+cz+d=0; set the first area threshold to S0;

[0124] Initialize the lateral fitting parameter a_best=0 in the target ground equation; initialize the longitudinal fitting parameter b_best=0 in the target ground equation; initialize the vertical fitting parameter c_best=0 in the target ground equation; initialize the constant term fitting parameter d_best=0 in the target ground equation; initialize the maximum value num_best of the fitting accuracy parameter=0;

[0125] For i=1:10000: (i represents the number of the first data set)

[0126] Initialize the fitting accuracy parameter num corresponding to the current initial plane equation to 0;

[0127] Randomly select three point cloud data from pointdata; and calculate the area S of the triangle formed by these three point cloud data;

[0128] If(S≤S0)

[0129] Continue;

[0130] Else(S>S0)

[0131] Solve the initial plane equation corresponding to the current round;

[0132] End

[0133] For j = 1:20000 (j represents the number of historical point cloud data)

[0134] Substitute x and y in the current historical point cloud data (x, y, z) into the initial plane equation of the current round to solve the vertical projection data z_pro corresponding to the current historical point cloud data;

[0135] Calculate z_diff_abs=|z-z_pro|;

[0136]

[0137] 3. After determining the target ground equation, respectively calculate the target distances from multiple initial point cloud data to the target ground equation, and further determine the initial point cloud data whose target distance is greater than a preset first distance threshold as the interference point cloud data; filter out the interference point cloud data, and issue a stop tracking instruction.

[0138] Based on the above point cloud classification method, one or more embodiments of the present application also provide a point cloud classification device. Figure 3 It is a structural diagram of an embodiment of a point cloud classification device provided in the present application. The point cloud classification device provided in the present application may include: an acquisition module 301, a calculation module 302, a detection module 303, and a classification module 304.

[0139] An acquisition module 301 is used to acquire initial point cloud data of a target road section and a target ground equation;

[0140] A calculation module 302 is used to calculate a target distance from the initial point cloud data to the target ground equation based on the target ground equation;

[0141] A detection module 303 is used to detect whether the target distance is greater than a preset first distance threshold and obtain a first detection result;

[0142] The classification module 304 is used to filter out the initial point cloud data whose target distance is greater than a preset first distance threshold from the first detection result as interference point cloud data; wherein the interference point cloud data includes point cloud data corresponding to fixed objects in the target section.

[0143] In some embodiments, the above-mentioned point cloud classification device is also used to send a stop tracking instruction for the interference point cloud data to the target radar.

[0144] In some embodiments, the above-mentioned point cloud classification device is also used to filter out the initial point cloud data whose target distance is less than or equal to a preset first distance threshold from the first detection result as target point cloud data; wherein the target point cloud data includes point cloud data corresponding to the vehicle in the target section; and send a tracking instruction for the target point cloud data to the target radar.

[0145] In some embodiments, the above-mentioned point cloud classification device also includes a construction module, which is used to pre-construct a target ground equation in the following manner: obtain historical point cloud data of the target road section; construct multiple first data sets based on the historical point cloud data; wherein the first data set includes three point cloud data randomly selected from the historical point cloud data; respectively detect whether the three point cloud data in the first data set are collinear to obtain a second detection result; filter out the first data set whose second detection result is not collinear as the second data set; use the second data set to perform plane fitting respectively to determine multiple initial ground equations; wherein the initial ground equation and the second data set correspond one to one; determine the fitting accuracy parameters for the initial ground equation based on the historical point cloud data; filter out the initial ground equation whose fitting accuracy parameters meet the requirements as the target ground equation.

[0146] It should be noted that the above-mentioned device may also include other implementation modes according to the description of the method embodiment. The specific implementation mode may refer to the description of the relevant method embodiment, which will not be described one by one here.

[0147] The present application also provides a server, comprising at least one processor and a memory for storing processor executable instructions, wherein the processor can implement any one of the above method embodiments when executing the instructions. For example, when the instructions are executed by the processor, the following steps are implemented: obtaining the initial point cloud data and the target ground equation of the target road section; based on the target ground equation, calculating the target distance from the initial point cloud data to the target ground equation; detecting whether the target distance is greater than a preset first distance threshold, and obtaining a first detection result; from the first detection result, filtering out the initial point cloud data whose target distance is greater than the preset first distance threshold as interference point cloud data; wherein the interference point cloud data includes point cloud data corresponding to fixed objects in the target road section.

[0148] It should be noted that the above server may also include other implementations according to the description of the method or device embodiment. The specific implementation methods can refer to the description of the relevant method embodiments, which will not be described one by one here.

[0149] Specifically, Figure 4 The hardware structure block diagram of an embodiment of a server provided in this application. Figure 4 As shown in the figure, the server 10 may include one or more (only one is shown in the figure) processors 100 (the processor 100 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 200 for storing data, a transmission module 300 for communication functions, and an internal bus 400 for data interaction. It can be understood by those skilled in the art that Figure 4 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 4 More or fewer components shown, for example, may also include other processing hardware, such as a database multi-level cache, a GPU, a volatile memory, or a processor having the same Figure 4 Different configurations are shown.

[0150] The memory 200 can be used to store software programs and modules of application software, such as program instructions / modules corresponding to a point cloud classification method in an embodiment of the present application. The processor 100 executes various functional applications and data processing by running the software programs and modules stored in the memory 200. The memory 200 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory.

[0151] The memory 200 may further include a memory remotely arranged relative to the processor 100, and these remote memories may be connected to the computer terminal via a network. Examples of the above network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0152] The transmission module 300 is used to receive or send data via a network. The above-mentioned network specific embodiment may include a wireless network provided by a communication provider of the computer terminal. In one embodiment, the transmission module 300 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one embodiment, the transmission module 300 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0153] In addition, the logic instructions in the above-mentioned memory can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art, and the computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk.

[0154] The present application also provides a storage medium, which is a computer-readable storage medium, and the storage medium stores computer program instructions, which, when executed, implement the following: obtaining initial point cloud data and a target ground equation of a target road section; calculating a target distance from the initial point cloud data to the target ground equation based on the target ground equation; detecting whether the target distance is greater than a preset first distance threshold to obtain a first detection result; and screening out, from the first detection result, initial point cloud data whose target distance is greater than the preset first distance threshold as interference point cloud data; wherein the interference point cloud data includes point cloud data corresponding to fixed objects in the target road section.

[0155] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0156] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0157] In summary, although the present application has been disclosed as above with preferred embodiments, the above preferred embodiments are not intended to limit the present application. Ordinary technicians in this field can make various changes and modifications without departing from the spirit and scope of the present application. Therefore, the scope of protection of the present application shall be based on the scope defined in the claims.

Claims

1. A point cloud classification method, characterized in that: include: Obtain the initial point cloud data and target ground equation of the target road section; Based on the target ground equation, calculating a target distance from the initial point cloud data to the target ground equation; Detecting whether the target distance is greater than a preset first distance threshold, and obtaining a first detection result; From the first detection result, the initial point cloud data whose target distance is greater than a preset first distance threshold is screened out as interference point cloud data; wherein the interference point cloud data includes point cloud data corresponding to fixed objects in the target section.

2. The method according to claim 1, characterized in that The initial point cloud data is obtained by scanning a target radar arranged above the target road section; Correspondingly, after selecting the initial point cloud data whose target distance is greater than a preset first distance threshold from the first detection result as interference point cloud data, the method further includes: A stop tracking instruction for the interference point cloud data is sent to the target radar.

3. The method according to claim 2, characterized in that After detecting whether the target distance is greater than a preset first distance threshold and obtaining a first detection result, the method further includes: From the first detection result, the initial point cloud data whose target distance is less than or equal to a preset first distance threshold is selected as the target point cloud data; wherein the target point cloud data includes the point cloud data corresponding to the vehicle in the target road section; A tracking instruction for the target point cloud data is sent to the target radar.

4. The method according to claim 1, characterized in that: The target ground equation is pre-constructed as follows: Acquire historical point cloud data of the target road section; Based on the historical point cloud data, construct a plurality of first data sets; wherein the first data sets include three point cloud data randomly selected from the historical point cloud data; respectively detecting whether the three point cloud data in the first data set are collinear, and obtaining a second detection result; Filtering out the first data set whose second detection result is not collinear as the second data set; Performing plane fitting using the second data set to determine a plurality of initial ground equations; wherein the initial ground equations correspond to the second data set in one-to-one correspondence; Determining a fitting accuracy parameter for the initial ground equation according to the historical point cloud data; An initial ground equation whose fitting accuracy parameter meets the requirements is selected as the target ground equation.

5. The method according to claim 4, characterized in that Determining fitting accuracy parameters for the initial ground equation according to the historical point cloud data includes: Calculating the vertical projection distances from the historical point cloud data to each initial ground equation respectively; Detecting whether the vertical projection distance is less than a preset second distance threshold, and obtaining a second detection result; Filtering out, from the second detection results, a vertical projection distance that is less than a preset second distance threshold as a third data set corresponding to the initial ground equation; The number of data in each third data set is determined as a fitting accuracy parameter of each initial ground equation.

6. The method according to claim 5, characterized in that Calculating the vertical projection distances of the historical point cloud data to each initial ground equation respectively, including: Inputting the horizontal coordinate data and the vertical coordinate data in the historical point cloud data into the initial ground equation to obtain the vertical projection data corresponding to the historical point cloud data; The absolute value of the difference between the vertical coordinate data in the historical point cloud data and the vertical projection data is calculated as the vertical projection distance corresponding to the initial ground equation.

7. The method according to claim 4, characterized in that Screening out the initial ground equation whose fitting accuracy parameters meet the requirements as the target ground equation includes: Arranging the fitting accuracy parameters in descending order to obtain a sorting result; According to the ranking results, the fitting accuracy parameters with the highest ranking are screened out, and the initial ground equations corresponding to the fitting accuracy parameters with the highest ranking are determined as the target ground equations.

8. A point cloud classification device, characterized in that: include: An acquisition module is used to obtain the initial point cloud data of the target road section and the target ground equation; A calculation module, used for calculating a target distance from the initial point cloud data to the target ground equation based on the target ground equation; A detection module, used to detect whether the target distance is greater than a preset first distance threshold, and obtain a first detection result; A classification module is used to filter out the initial point cloud data whose target distance is greater than a preset first distance threshold from the first detection result as interference point cloud data; wherein the interference point cloud data includes point cloud data corresponding to fixed objects in the target section.

9. A server, characterized in that: The invention comprises a processor and a memory for storing processor-executable instructions, wherein the processor implements the steps of the method according to any one of claims 1 to 7 when executing the instructions.

10. A storage medium, characterized in that: Computer instructions are stored thereon, and when the instructions are executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images