Radar calibration method and device, electronic equipment and storage medium

Abstract

The invention provides a radar calibration method and apparatus, an electronic device and a storage medium. The method comprises the steps of obtaining a first original point cloud collected by a first radar and a second original point cloud collected by a second radar on a target vehicle; determining a first non-ground point cloud included in the first original point cloud and a second non-ground point cloud included in the second original point cloud; based on a registration algorithm, performing registration processing on the first non-ground point cloud and the second non-ground point cloud to obtain a target external parameter between the first non-ground point cloud and the second non-ground point cloud; and calibrating the first radar and the second radar based on the target external parameters. According to the invention, radar calibration is carried out on the basis of the ground features of the target vehicle, professional personnel and professional equipment are not needed, the operation cost is reduced, and the calibration efficiency is improved.

Description

Technical Field

[0001] The present application relates to the field of intelligent driving technology, and more specifically, to a radar calibration method, device, electronic device and storage medium. Background Art

[0002] With the rapid development of science and technology, more and more control devices tend to be intelligent, especially intelligent driving. Intelligent driving refers to the technology that machines help people drive, and completely replace human driving in special circumstances. In the field of intelligent driving, radar calibration is a very important part.

[0003] The existing methods for radar calibration mainly include offline calibration, online calibration and adaptive calibration, etc. However, these calibration methods still have certain defects in terms of operation complexity, calibration cost and calibration accuracy. Summary of the invention

[0004] In view of this, the purpose of the present application is to provide a radar calibration method, device, electronic device and storage medium to overcome the problems in the prior art.

[0005] In a first aspect, an embodiment of the present application provides a radar calibration method, the method comprising:

[0006] Acquire a first original point cloud collected by a first radar and a second original point cloud collected by a second radar on a target vehicle;

[0007] Determine a first non-ground point cloud included in the first original point cloud and a second non-ground point cloud included in the second original point cloud;

[0008] Based on a registration algorithm, registering the first non-ground point cloud and the second non-ground point cloud to obtain a target extrinsic parameter between the first non-ground point cloud and the second non-ground point cloud;

[0009] The first radar and the second radar are calibrated based on the target extrinsic parameters.

[0010] In some technical solutions of the present application, the above-mentioned determining the first non-ground point cloud included in the first original point cloud includes:

[0011] Based on a first initial external parameter of the first radar, converting the first original point cloud into a first converted point cloud;

[0012] According to a preset height requirement, the first converted point cloud is screened to obtain a first candidate point cloud;

[0013] The first non-ground point cloud is obtained by performing ground correction on the first candidate point cloud.

[0014] In some technical solutions of the present application, the first conversion point cloud is screened according to the preset height requirement to obtain the first candidate point cloud, including:

[0015] Based on a first height threshold, the first converted point cloud is screened to obtain a first preliminary selected point cloud;

[0016] Based on the first height interval, the first preliminary point cloud is screened to obtain the first candidate point cloud within the first height interval.

[0017] In some technical solutions of the present application, the first height threshold is determined based on the first initial external parameter of the first radar;

[0018] The first height interval is determined based on height information of the first preliminary point cloud.

[0019] In some technical solutions of the present application, the first non-ground point cloud is obtained by performing ground correction on the first candidate point cloud, including:

[0020] Calculating an actual ground normal vector of the first candidate point cloud;

[0021] Determining an adjustment extrinsic parameter corresponding to the first candidate point cloud based on a difference between the actual ground normal vector and the ideal ground normal vector;

[0022] Iteratively adjusting the first candidate point cloud based on the adjustment external parameter until the ground height of the adjusted first candidate point cloud is less than a preset target height threshold;

[0023] Based on the second height interval, the first non-ground point cloud is determined from the adjusted first candidate point cloud.

[0024] In some technical solutions of the present application, the determining of the adjustment extrinsic parameter corresponding to the first candidate point cloud based on the difference between the actual ground normal vector and the ideal ground normal vector includes:

[0025] Determine a roll angle and a pitch angle based on the actual ground normal vector and the ideal ground normal vector;

[0026] Based on the flip angle and the pitch angle, adjusting the first candidate point cloud so that the angle between the adjusted first candidate point cloud and the ground is less than a preset angle threshold;

[0027] Calculating the height of the adjusted first candidate point cloud from the ground;

[0028] The roll angle, the pitch angle and the ground height are used as the adjustment external parameters.

[0029] In some technical solutions of the present application, the above method further includes:

[0030] Displaying a first initial external parameter of the first radar and / or a second initial external parameter of the second radar on a graphical user interface;

[0031] In response to a radar calibration operation, calibration parameters related to the calibration operation are dynamically displayed through the graphical user interface.

[0032] In a second aspect, an embodiment of the present application provides a radar calibration device, the device comprising:

[0033] An acquisition module, used to acquire a first original point cloud collected by a first radar and a second original point cloud collected by a second radar on a target vehicle;

[0034] a distinguishing module, configured to determine a first non-ground point cloud included in the first original point cloud and a second non-ground point cloud included in the second original point cloud;

[0035] A registration module, configured to perform registration processing on the first non-ground point cloud and the second non-ground point cloud based on a registration algorithm to obtain a target extrinsic parameter between the first non-ground point cloud and the second non-ground point cloud;

[0036] A calibration module is used to calibrate the first radar and the second radar based on the target external parameters.

[0037] In a third aspect, an embodiment of the present application provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned radar calibration method when executing the computer program.

[0038] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned radar calibration method are executed.

[0039] The technical solution provided by the embodiments of the present application may have the following beneficial effects:

[0040] The method of the present application includes obtaining a first original point cloud collected by a first radar and a second original point cloud collected by a second radar on a target vehicle; determining a first non-ground point cloud contained in the first original point cloud and a second non-ground point cloud contained in the second original point cloud; performing registration processing on the first non-ground point cloud and the second non-ground point cloud based on a registration algorithm to obtain target extrinsic parameters between the first non-ground point cloud and the second non-ground point cloud; and calibrating the first radar and the second radar based on the target extrinsic parameters. The present application performs radar calibration based on the ground features where the target vehicle is located, does not require professional personnel and professional equipment, reduces operating costs, and improves calibration efficiency.

[0041] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0043] Figure 1 A schematic diagram of a radar calibration method provided in an embodiment of the present application is shown;

[0044] Figure 2 A schematic diagram showing a specific implementation method provided by an embodiment of the present application is shown;

[0045] Figure 3 A schematic diagram of a radar calibration device provided in an embodiment of the present application is shown;

[0046] Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0047] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It should be understood that the drawings in the present application only serve the purpose of explanation and description and are not used to limit the scope of protection of the present application. In addition, it should be understood that the schematic drawings are not drawn in real proportion. The flowchart used in this application shows the operations implemented according to some embodiments of the present application. It should be understood that the operations of the flowchart can be implemented out of sequence, and the steps without logical context can be reversed in order or implemented simultaneously. In addition, those skilled in the art can add one or more other operations to the flowchart under the guidance of the content of the present application, or remove one or more operations from the flowchart.

[0048] In addition, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.

[0049] It should be noted that the term "comprising" will be used in the embodiments of the present application to indicate the existence of the features declared thereafter, but does not exclude the addition of other features.

[0050] The existing technology for radar calibration mainly includes the following:

[0051] Offline calibration: Use a pre-prepared calibration plate or calibration scene to manually measure the relative position and attitude with the Lidar (radar), and then use the calibration algorithm to calculate the accurate calibration parameters. This method requires a long preparation time and professional equipment, but the calibration results are more accurate.

[0052] Online calibration: This method performs calibration while the robot is running. The robot automatically calibrates by interacting with the environment. For example, during movement, the robot can monitor feature points or edge points on the ground and update calibration parameters through optimization algorithms. Online calibration is more flexible than offline calibration, but the calibration process may be affected by environmental changes.

[0053] Adaptive calibration: This method uses multiple lidars for calibration, which are calibrated with each other to improve the overall calibration accuracy. In adaptive calibration, each lidar can observe and compare their own measurements to determine the calibration parameters and apply them to the entire system. This method usually has high robustness and accuracy.

[0054] The existing calibration method has the following defects:

[0055] Complexity: Multi-lidar calibration usually requires high technical requirements and professional equipment, and the calibration process is relatively complex. For non-professionals, additional training and experience may be required to perform calibration correctly.

[0056] Time consumption: Whether it is offline calibration or online calibration, it takes a certain amount of time to complete the calibration process. Especially when multiple Lidars are used in a large-scale system, the calibration process may take a long time, affecting the real-time performance of the actual application.

[0057] Impact of environmental changes: Online calibration methods usually have high requirements on the environment. For scenes with large environmental changes, such as changes in lighting and weather, the calibration accuracy may be affected to a certain extent.

[0058] High cost: Multi-lidar calibration solutions usually require additional hardware equipment and professional tools for calibration, which increases the cost of the overall system.

[0059] Calibration error: Even with advanced calibration methods, there may still be some errors in the calibration results due to measurement errors, algorithm errors, etc. These errors may have a certain impact on the overall positioning and perception capabilities of the system.

[0060] Based on this, the embodiments of the present application provide a radar calibration method, device, electronic device and storage medium, which are described below through embodiments.

[0061] Figure 1 A schematic flow chart of a radar calibration method provided in an embodiment of the present application is shown, wherein the method includes steps S101-S104; specifically:

[0062] S101, obtaining a first original point cloud collected by a first radar and a second original point cloud collected by a second radar on a target vehicle;

[0063] S102, determining a first non-ground point cloud included in the first original point cloud and a second non-ground point cloud included in the second original point cloud;

[0064] S103, performing registration processing on the first non-ground point cloud and the second non-ground point cloud based on a registration algorithm to obtain a target extrinsic parameter between the first non-ground point cloud and the second non-ground point cloud;

[0065] S104: Calibrate the first radar and the second radar based on the target external parameters.

[0066] This application performs radar calibration based on the ground features where the target vehicle is located. It does not require professional personnel and professional equipment, reduces operating costs, and improves calibration efficiency.

[0067] Some embodiments of the present application are described in detail below. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0068] S101, obtaining a first original point cloud collected by a first radar and a second original point cloud collected by a second radar on a target vehicle.

[0069] The target vehicle in the embodiment of the present application is an intelligent driving vehicle, and multiple radars are installed on the intelligent driving vehicle. During the driving process, the target vehicle collects information through radar, and realizes the intelligent driving function after processing the information collected by the radar. In order to ensure the accuracy of the information collected by the radar, the radar needs to be calibrated. Radar calibration, that is, radar external parameter calibration, refers to the relative position relationship between radars, including translation and rotation.

[0070] When calibrating the radar installed on the target vehicle, it is generally calibrated according to the preset calibration order. For example, according to the position between the radars, the radars are numbered ABCD, etc. In order of numbering, radar A and radar B, radar B and radar C, radar C and radar D, etc. are calibrated in sequence until all the radars on the target vehicle are calibrated. In other words, when calibrating the radar of the target vehicle, it is calibrated in groups of two. In this group of radars, any radar is used as a reference radar, and the other radar is used as a radar to be aligned. The radar to be aligned is aligned to make it consistent with the external parameters of the reference radar. In order to distinguish the radars in a group, the embodiment of the present application refers to any radar in the group as the first radar, and the other radar is called the second radar. When the first radar here is the reference radar, the second radar is the radar to be aligned; when the first radar here is the radar to be aligned, the second radar is the reference radar.

[0071] After determining the first radar and the second radar of a certain group on the target vehicle, the embodiment of the present application obtains the first original point cloud collected by the first radar and the second original point cloud collected by the second radar. Before the first radar collects the first original point cloud and the second radar collects the second original point cloud, the embodiment of the present application needs to stop the target vehicle in a pre-set target area. The target area here has certain requirements on the flatness of the ground, that is, the embodiment of the present application selects a relatively flat place in advance as the target area. When the target vehicle stops in the target area, the first radar and the second radar are started to work respectively to obtain the first original point cloud and the second original point cloud in their respective detection areas.

[0072] S102: Determine a first non-ground point cloud included in the first original point cloud and a second non-ground point cloud included in the second original point cloud.

[0073] After obtaining the first original point cloud collected by the first radar and the second original point cloud collected by the second radar, the embodiment of the present application needs to distinguish the first original point cloud and the second original point cloud, and determine the first ground point cloud and the first non-ground point cloud contained in the first original point cloud, and the second ground point cloud and the second non-ground point cloud contained in the second original point cloud.

[0074] The process of distinguishing the first original point cloud: the first original point cloud is collected by the first radar, that is, the first original point cloud is collected in the coordinate system where the first radar is located. After obtaining the first original point cloud, the embodiment of the present application needs to convert it from the first radar coordinate system to the ground coordinate system. The specific conversion method is: according to the first initial external parameter of the first radar, the coordinate system of the first original point cloud is converted to obtain the first converted point cloud in the ground coordinate system. The first initial external parameter here is determined during the production process of the target vehicle, which represents the relative position relationship between the radar and the ground after the production of the target vehicle. For example, based on the initial estimated external parameter (first initial external parameter) T1_guess of the first original point cloud, P1_base-ground = T1_guess*P1_lidar1, the first original point cloud is transferred to the ground coordinate system P1_base-ground. Among them, P is the abbreviation of point cloud pointcloud, and T is the abbreviation of transform matrix transform.

[0075] After obtaining the first conversion point cloud, in order to improve the efficiency of segmentation, the embodiment of the present application first screens the first conversion point cloud, and after excluding the highly abnormal points in the first conversion point cloud, a first candidate point cloud is obtained. Here, the basis for excluding the highly abnormal points is whether the preset height requirement is met. If a point cloud meets the height requirement, the point cloud is retained, and if a point cloud does not meet the height requirement, the point cloud is excluded.

[0076] The height requirement here includes two sub-requirements, namely the first height threshold requirement and the first height interval requirement. The specific screening process is: based on the first height threshold, the first conversion point cloud is screened to obtain the first preliminary point cloud; based on the first height interval, the first preliminary point cloud is screened to obtain the first candidate point cloud located within the first height interval. The embodiment of the present application considers that point clouds greater than or equal to the first height threshold are abnormal point clouds, so it is necessary to exclude point clouds exceeding the first height threshold from the first conversion point cloud to obtain the first preliminary point cloud with a height less than the first height threshold. After obtaining the first preliminary point cloud, the first preliminary point cloud needs to be screened again. The screening of the first preliminary point cloud is based on the first height interval, and the point clouds in the first preliminary point cloud that are not in the first height interval are excluded, and the point clouds in the first height interval are used as the first candidate point cloud.

[0077] It should be noted that the first height threshold in the embodiment of the present application is determined based on the first initial external parameter of the first radar. After obtaining the first initial external parameter of the first radar, the first initial external parameter is multiplied by a preset external parameter multiplier to obtain the first height threshold. The first height interval in the embodiment of the present application is determined based on the height information of each first preliminary point cloud. After obtaining the first preliminary point cloud, the embodiment of the present application calculates the average height of each first preliminary point cloud, and then adds and subtracts a preset height from the average height to obtain the first height interval here. In the specific implementation, in order to filter out points with too high a height and reduce the calculation time, P1_base-ground is sorted by height z, and points (first preliminary point cloud) less than 1.5 times (preset external parameter multiplier) in the first initial external parameter T1_guess are obtained. Calculate the z average value, and use the z average value +-10cm (preset height) as the first candidate point cloud.

[0078] After obtaining the first candidate point cloud, it is necessary to identify the first ground point cloud and the first non-ground point cloud in the first candidate point cloud. For the first candidate point cloud, the embodiment of the present application adopts a method of ground correction of the first candidate point cloud. Specifically, the embodiment of the present application first calculates the actual ground normal vector of the first candidate point cloud, for example, by solving the actual ground normal vector of the first candidate point cloud through SVD. Then, with the purpose of approaching the ideal ground normal vector, the actual ground normal vector of the first candidate point cloud is adjusted. The ideal ground plane is the one with a normal vector of (0, 0, 1), and the adjustment here is based on the adjustment extrinsic parameters of the first candidate point cloud. The adjustment extrinsic parameters of the first candidate point cloud include roll angle, pitch angle and ground height. Among them, roll and pitch are obtained by converting the difference between the actual ground normal vector of the first candidate point cloud and the ideal ground normal vector, and the ground height is the average height of the first candidate point cloud. Specifically, based on the actual ground normal vector and the ideal ground normal vector, a flip angle and a pitch angle are determined; based on the flip angle and the pitch angle, the first candidate point cloud is adjusted so that the angle between the adjusted first candidate point cloud and the ground is less than a preset angle threshold; the ground height of the adjusted first candidate point cloud from the ground is calculated; and the flip angle, the pitch angle and the ground height are used as the adjustment external parameters.

[0079] When adjusting the actual ground normal vector of the first candidate point cloud, continuous iteration is required until the ground height of the adjusted point cloud is less than the preset target height threshold. The target height threshold here is set based on the actual height of the ground. The ground height of the adjusted point cloud is less than the preset target height threshold, which indicates that the actual ground normal vector of the first candidate point cloud coincides with or is close to the ideal ground normal vector (the difference is small and can be ignored), which can also be understood as the first candidate point cloud coinciding with the actual ground. At this time, the adjusted first candidate point cloud is screened again through the second height interval, and the first non-ground point cloud is obtained from the first candidate point cloud. The embodiment of the present application considers that the points in the second height interval of the first candidate point cloud are the first ground point cloud, and the first non-ground point cloud in the first candidate point cloud can be obtained after excluding the first ground point cloud in the first candidate point cloud. The second high zone interval here is determined based on the adjustment external parameters of the first radar. That is, the second height interval can be obtained by adding and subtracting a preset height from the adjustment external parameters of the first radar.

[0080] like Figure 2 As shown, in the specific implementation, SVD (Singular Value Decomposition, a mathematical technique for decomposing a matrix into three independent matrices) is used to solve the ground normal vector normal1_svd and height h1_svd of the first candidate point cloud. The normal vector normal1_svd solved by svd and the ideal ground normal vector normal_ground (0, 0, 1) are used to calculate the difference, and the roll1 and pitch1 converted by the difference. The ideal ground plane is the normal vector (0, 0, 1). The vector difference is converted into a rotation matrix, and then the rotation matrix is ​​converted into roll and pitch. After calculating the adjusted variables of roll and pitch, the variables are used to adjust the ground point cloud so that the ground point cloud conforms to the real ground. The average height H1_ground-everage of the converted ground point cloud is calculated to obtain the lidar height ground height H1_fix = -H1_ground-everage. The adjusted roll1, pitch1 and H1_fix are used as external parameters T1_plane-fit for reflux to adjust the input point cloud. Until the H1_fix adjustment value is less than 0.5cm. (Preprocess the point cloud with the adjustment external parameter T1_plane-fit and transform it to make the point cloud more parallel to the ground. Each iteration will optimize the new adjustment external parameter T1_plane-fit). Based on the adjusted T1_plane-fit, remove the ground point cloud of +-15cm to obtain the first non-ground point cloud P1_outofground.

[0081] The process of distinguishing the second original point cloud: the second original point cloud is collected by the second radar, that is, the second original point cloud is collected in the coordinate system where the second radar is located. After obtaining the second original point cloud, the embodiment of the present application needs to convert it from the second radar coordinate system to the ground coordinate system. The specific conversion method is: according to the second initial external parameter of the second radar, the coordinate system of the second original point cloud is converted to obtain the second converted point cloud in the ground coordinate system. The second initial external parameter here is determined during the production process of the target vehicle, which represents the relative position relationship between the radar and the ground after the target vehicle is produced. For example, based on the initial estimated external parameter (second initial external parameter) T2_guess of the second original point cloud, P2_base-ground = T2_guess*P2_lidar2, the second original point cloud is transferred to the ground coordinate system P2_base-ground. Among them, P is the abbreviation of point cloud pointcloud, and T is the abbreviation of transform matrix transform.

[0082] After obtaining the second conversion point cloud, in order to improve the division efficiency, the embodiment of the present application first screens the second conversion point cloud, and after excluding the highly abnormal points in the second conversion point cloud, a second candidate point cloud is obtained. Here, the basis for excluding highly abnormal points is whether the preset height requirement is met. If a point cloud meets the height requirement, the point cloud is retained, and if a point cloud does not meet the height requirement, the point cloud is excluded.

[0083] The height requirement here includes two sub-requirements, namely the second height threshold requirement and the third height interval requirement. The specific screening process is: based on the second height threshold, the second converted point cloud is screened to obtain the second preliminary point cloud; based on the third height interval, the second preliminary point cloud is screened to obtain the second candidate point cloud located within the third height interval. The embodiment of the present application considers that point clouds greater than or equal to the second height threshold are abnormal point clouds, so it is necessary to exclude point clouds exceeding the second height threshold from the second converted point cloud to obtain a second preliminary point cloud with a height less than the second height threshold. After obtaining the second preliminary point cloud, the second preliminary point cloud needs to be screened again. The screening of the second preliminary point cloud is based on the third height interval, and the point clouds in the second preliminary point cloud that are not in the third height interval are excluded, and the point clouds in the third height interval are used as the second candidate point cloud.

[0084] It should be noted that the second height threshold in the embodiment of the present application is determined based on the second initial external parameter of the second radar. After obtaining the second initial external parameter of the second radar, the second initial external parameter is multiplied by a preset external parameter multiplier to obtain the second height threshold. The third height interval in the embodiment of the present application is determined based on the height information of each second preliminary point cloud. After obtaining the second preliminary point cloud, the embodiment of the present application calculates the average height of each second preliminary point cloud, and then adds and subtracts a preset height from the average height to obtain the third height interval here. In specific implementation, in order to filter out points with too high height and reduce calculation time, P2_base-ground is sorted by height z, and points (second preliminary point cloud) less than 1.5 times (preset external parameter multiplier) in the second initial external parameter external parameter T2_guess are obtained. Calculate the z average value, and use the z average value +-10cm (preset height) as the second candidate point cloud.

[0085] After obtaining the second candidate point cloud, it is necessary to identify the second ground point cloud and the second non-ground point cloud in the second candidate point cloud. For the second candidate point cloud, the embodiment of the present application adopts a method of ground correction of the second candidate point cloud. Specifically, the embodiment of the present application first calculates the actual ground normal vector of the second candidate point cloud, for example, by solving the actual ground normal vector of the second candidate point cloud through SVD. Then, with the purpose of approaching the ideal ground normal vector, the actual ground normal vector of the second candidate point cloud is adjusted. The ideal ground plane is the one with a normal vector of (0, 0, 1), and the adjustment here is based on the adjustment extrinsic parameters of the second candidate point cloud. The adjustment extrinsic parameters of the second candidate point cloud include roll angle, pitch angle and ground height. Among them, roll and pitch are obtained by converting the difference between the actual ground normal vector of the second candidate point cloud and the ideal ground normal vector, and the ground height is the average height of the second candidate point cloud. Specifically, based on the actual ground normal vector and the ideal ground normal vector, a flip angle and a pitch angle are determined; based on the flip angle and the pitch angle, the second candidate point cloud is adjusted so that the angle between the adjusted second candidate point cloud and the ground is less than a preset angle threshold; the ground height of the adjusted second candidate point cloud from the ground is calculated; and the flip angle, the pitch angle and the ground height are used as the adjustment external parameters.

[0086] When adjusting the actual ground normal vector of the second candidate point cloud, continuous iteration is required until the ground height of the adjusted point cloud is less than the preset target height threshold. The target height threshold here is set based on the actual height of the ground. The ground height of the adjusted point cloud is less than the preset target height threshold, which indicates that the actual ground normal vector of the second candidate point cloud coincides with or is close to the ideal ground normal vector (the difference is small and can be ignored), which can also be understood as the second candidate point cloud coinciding with the actual ground. At this time, the adjusted second candidate point cloud is screened again by the second height interval, and the second non-ground point cloud is obtained from the second candidate point cloud. The embodiment of the present application considers that the points in the second height interval in the second candidate point cloud are the second ground point cloud, and the second non-ground point cloud in the second candidate point cloud can be obtained after excluding the second ground point cloud in the second candidate point cloud. The second high zone interval here is determined based on the adjustment external parameters of the second radar. That is, the second height interval can be obtained by adding and subtracting two preset heights respectively by the adjustment external parameters of the second radar.

[0087] like Figure 2 As shown, in the specific implementation, SVD (Singular Value Decomposition, a mathematical technique for decomposing a matrix into three independent matrices) is used to solve the ground normal vector normal2_svd and height h2_svd of the second candidate point cloud. The normal vector normal2_svd solved by svd and the ideal ground normal vector normal_ground (0, 0, 1) are used to calculate the difference, and the roll2 and pitch2 converted by the difference. The ideal ground plane is the normal vector (0, 0, 1). The vector difference is converted into a rotation matrix, and then the rotation matrix is ​​converted into roll and pitch. After calculating the adjusted variables of roll and pitch, the variables are used to adjust the ground point cloud so that the ground point cloud conforms to the real ground. The average height H2_ground-everage of the converted ground point cloud is calculated to obtain the lidar height ground height H2_fix = -H2_ground-everage. The adjusted roll2, pitch2 and H2_fix are used as external parameters T2_plane-fit for reflux to adjust the input point cloud. Until the H2_fix adjustment value is less than 0.5cm. (Preprocess the point cloud with the adjustment external parameter T2_plane-fit and transform it to make the point cloud more parallel to the ground. Each iteration will optimize the new adjustment external parameter T2_plane-fit). Based on the adjusted T2_plane-fit, remove the ground point cloud of +-15cm to obtain the second non-ground point cloud P2_outofground.

[0088] S103: Perform registration processing on the first non-ground point cloud and the second non-ground point cloud based on a registration algorithm to obtain target extrinsic parameters between the first non-ground point cloud and the second non-ground point cloud.

[0089] S104: Calibrate the first radar and the second radar based on the target external parameters.

[0090] After obtaining the first non-ground point cloud and the second non-ground point cloud, the embodiment of the present application performs registration processing on them through a preset registration algorithm to obtain the target extrinsic parameter between the first non-ground point cloud and the second non-ground point cloud. The specific registration algorithm here can be NICP, and the first non-ground point cloud and the second non-ground point cloud are processed by NICP to obtain the target extrinsic parameter T_final. Then, the first radar and the second radar are calibrated based on the target extrinsic parameter T_final.

[0091] As an optional embodiment, in order to ensure the accuracy of the data, the embodiment of the present application also provides a graphical user interface with the isomorphic terminal device, and after obtaining the first initial external parameter of the first radar and / or the second initial external parameter of the second radar, it is also displayed on the graphical user interface. The user can view the first initial external parameter and / or the second initial external parameter through the graphical user interface. If the user thinks that the first initial external parameter and / or the second initial external parameter are obviously abnormal, the first initial external parameter and / or the second initial external parameter can be adjusted through the adjustment control designed on the touch graphical user interface. After the user adjusts the first initial external parameter and / or the second initial external parameter, the adjusted first initial external parameter and / or the second initial external parameter are dynamically displayed on the graphical user interface. When performing radar calibration, the first initial external parameter and / or the second initial external parameter adjusted by the user are used.

[0092] Figure 3 A schematic diagram of the structure of a radar calibration device provided in an embodiment of the present application is shown, and the device includes:

[0093] An acquisition module, used to acquire a first original point cloud collected by a first radar and a second original point cloud collected by a second radar on a target vehicle;

[0094] a distinguishing module, configured to determine a first non-ground point cloud included in the first original point cloud and a second non-ground point cloud included in the second original point cloud;

[0095] A registration module, configured to perform registration processing on the first non-ground point cloud and the second non-ground point cloud based on a registration algorithm to obtain a target extrinsic parameter between the first non-ground point cloud and the second non-ground point cloud;

[0096] A calibration module is used to calibrate the first radar and the second radar based on the target external parameters.

[0097] Determining a first non-ground point cloud included in the first original point cloud includes:

[0098] Based on a first initial external parameter of the first radar, converting the first original point cloud into a first converted point cloud;

[0099] According to a preset height requirement, the first converted point cloud is screened to obtain a first candidate point cloud;

[0100] The first non-ground point cloud is obtained by performing ground correction on the first candidate point cloud.

[0101] The step of screening the first conversion point cloud according to the preset height requirement to obtain a first candidate point cloud includes:

[0102] Based on a first height threshold, the first converted point cloud is screened to obtain a first preliminary selected point cloud;

[0103] Based on the first height interval, the first preliminary point cloud is screened to obtain the first candidate point cloud within the first height interval.

[0104] The first height threshold is determined based on the first initial external parameter of the first radar;

[0105] The first height interval is determined based on height information of the first preliminary point cloud.

[0106] The step of obtaining the first non-ground point cloud by performing ground correction on the first candidate point cloud includes:

[0107] Calculating an actual ground normal vector of the first candidate point cloud;

[0108] Determining an adjustment extrinsic parameter corresponding to the first candidate point cloud based on a difference between the actual ground normal vector and the ideal ground normal vector;

[0109] Iteratively adjusting the first candidate point cloud based on the adjustment external parameter until the ground height of the adjusted first candidate point cloud is less than a preset target height threshold;

[0110] Based on the second height interval, the first non-ground point cloud is determined from the adjusted first candidate point cloud.

[0111] The determining, based on the difference between the actual ground normal vector and the ideal ground normal vector, an adjustment extrinsic parameter corresponding to the first candidate point cloud includes:

[0112] Determine a roll angle and a pitch angle based on the actual ground normal vector and the ideal ground normal vector;

[0113] Based on the flip angle and the pitch angle, adjusting the first candidate point cloud so that the angle between the adjusted first candidate point cloud and the ground is less than a preset angle threshold;

[0114] Calculating the height of the adjusted first candidate point cloud from the ground;

[0115] The roll angle, the pitch angle and the ground height are used as the adjustment external parameters.

[0116] The device further comprises a response module, configured to display a first initial external parameter of the first radar and / or a second initial external parameter of the second radar on a graphical user interface;

[0117] In response to a radar calibration operation, calibration parameters related to the calibration operation are dynamically displayed through the graphical user interface.

[0118] like Figure 4 As shown, an embodiment of the present application provides an electronic device for executing the radar calibration method in the present application, the device includes a memory, a processor, a bus, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the radar calibration method when executing the computer program.

[0119] Specifically, the above-mentioned memory and processor may be general-purpose memory and processor, which are not specifically limited here. When the processor runs the computer program stored in the memory, the above-mentioned radar calibration method can be executed.

[0120] Corresponding to the radar calibration method in the present application, an embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned radar calibration method are executed.

[0121] Specifically, the storage medium can be a general storage medium, such as a mobile disk, a hard disk, etc. When the computer program on the storage medium is run, the above-mentioned radar calibration method can be executed.

[0122] In the embodiments provided in the present application, it should be understood that the disclosed systems and methods can be implemented in other ways. The system embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, indirect coupling or communication connection of systems or units, which can be electrical, mechanical or other forms.

[0123] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0124] In addition, each functional unit in the embodiments provided in the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0125] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. 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. The computer software product is stored in a storage medium and includes several instructions for 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 methods 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 magnetic disk or an optical disk.

[0126] It should be noted that similar numbers and letters represent similar items in the following figures. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. In addition, the terms "first", "second", "third", etc. are only used to distinguish the description and are not to be understood as indicating or implying relative importance.

[0127] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The protection scope of the present application is not limited thereto. Although the present application is described in detail with reference to the above-mentioned embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-mentioned embodiments within the technical scope disclosed in the present application, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application. They should all be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A radar calibration method, characterized in that: The method comprises: Acquire a first original point cloud collected by a first radar and a second original point cloud collected by a second radar on a target vehicle; Determine a first non-ground point cloud included in the first original point cloud and a second non-ground point cloud included in the second original point cloud; Based on a registration algorithm, registering the first non-ground point cloud and the second non-ground point cloud to obtain a target extrinsic parameter between the first non-ground point cloud and the second non-ground point cloud; The first radar and the second radar are calibrated based on the target extrinsic parameters.

2. The method according to claim 1, characterized in that: The determining of the first non-ground point cloud included in the first original point cloud comprises: Based on a first initial external parameter of the first radar, converting the first original point cloud into a first converted point cloud; According to a preset height requirement, the first converted point cloud is screened to obtain a first candidate point cloud; The first non-ground point cloud is obtained by performing ground correction on the first candidate point cloud.

3. The method according to claim 2, characterized in that The step of screening the first conversion point cloud according to the preset height requirement to obtain a first candidate point cloud includes: Based on a first height threshold, the first converted point cloud is screened to obtain a first preliminary selected point cloud; Based on the first height interval, the first preliminary point cloud is screened to obtain the first candidate point cloud within the first height interval.

4. The method according to claim 3, characterized in that The first height threshold is determined based on the first initial external parameter of the first radar; The first height interval is determined based on height information of the first preliminary point cloud.

5. The method according to claim 2, characterized in that: The step of obtaining the first non-ground point cloud by performing ground correction on the first candidate point cloud includes: Calculating an actual ground normal vector of the first candidate point cloud; Determining an adjustment extrinsic parameter corresponding to the first candidate point cloud based on a difference between the actual ground normal vector and the ideal ground normal vector; Iteratively adjusting the first candidate point cloud based on the adjustment external parameter until the ground height of the adjusted first candidate point cloud is less than a preset target height threshold; Based on the second height interval, the first non-ground point cloud is determined from the adjusted first candidate point cloud.

6. The method according to claim 5, characterized in that The determining, based on the difference between the actual ground normal vector and the ideal ground normal vector, an adjustment extrinsic parameter corresponding to the first candidate point cloud includes: Determine a roll angle and a pitch angle based on the actual ground normal vector and the ideal ground normal vector; Based on the flip angle and the pitch angle, adjusting the first candidate point cloud so that the angle between the adjusted first candidate point cloud and the ground is less than a preset angle threshold; Calculating the height of the adjusted first candidate point cloud from the ground; The roll angle, the pitch angle and the ground height are used as the adjustment external parameters.

7. The method according to claim 1, characterized in that The method further comprises: Displaying a first initial external parameter of the first radar and / or a second initial external parameter of the second radar on a graphical user interface; In response to a radar calibration operation, calibration parameters related to the calibration operation are dynamically displayed through the graphical user interface.

8. A radar calibration device, characterized in that: The device comprises: An acquisition module, used to acquire a first original point cloud collected by a first radar and a second original point cloud collected by a second radar on a target vehicle; a distinguishing module, configured to determine a first non-ground point cloud included in the first original point cloud and a second non-ground point cloud included in the second original point cloud; A registration module, configured to perform registration processing on the first non-ground point cloud and the second non-ground point cloud based on a registration algorithm to obtain a target extrinsic parameter between the first non-ground point cloud and the second non-ground point cloud; A calibration module is used to calibrate the first radar and the second radar based on the target external parameters.

9. An electronic device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate via the bus, and when the machine-readable instructions are executed by the processor, the steps of the radar calibration method as described in any one of claims 1 to 7 are performed.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the radar calibration method according to any one of claims 1 to 7 are executed.

Images