Roadside laser radar calibration method, device and electronic equipment

Abstract

The present application discloses a road-side laser radar calibration method, device and electronic device, wherein the road-side laser radar calibration method can be applied to the field of vehicle-road cooperative automatic driving, and the method comprises: obtaining the first point cloud data collected by the road-side laser radar to be calibrated; based on the initial posture of the road-side laser radar, determining a local point cloud map that meets a preset distance condition from the point cloud map corresponding to the road-side laser radar; the point cloud map is obtained based on the second point cloud data within the preset range of the road-side laser radar collected by the laser radar on the acquisition vehicle; optimizing the initial posture of the road-side laser radar based on the first point cloud data and the local point cloud map until the preset optimization condition is met, and obtaining the target posture of the road-side laser radar; when the target posture meets the preset matching condition, using the target posture as an external parameter of the road-side laser radar. The present application provides accurate calibration parameters for road-side perception, which can achieve accurate positioning of obstacles.

Description

Technical Field

[0001] The present application relates to the field of autonomous driving technology, and in particular to a road-side laser radar calibration method, device and electronic equipment. Background Art

[0002] At present, more and more autonomous driving research has been invested in the field of autonomous driving based on V2X (Vehicle to Everything). In order to reduce the cost of autonomous driving vehicles, installing LiDAR on roadside equipment and then providing perception results to vehicles through V2X technology has become the mainstream solution for most. In order to ensure that the vehicle LiDAR can have accurate perception results, the calibration accuracy of the roadside LiDAR is particularly important.

[0003] Since road-side LiDARs are fixed on road-side equipment, immovable and high above the ground, traditional vehicle-mounted calibration solutions are not suitable for the calibration of road-side LiDARs. Therefore, a calibration solution that can solve the problem of road-side LiDARs is needed. Summary of the invention

[0004] In order to solve the problems of the prior art, the embodiments of the present application provide a roadside laser radar calibration method, device and electronic device. The technical solution is as follows:

[0005] On the one hand, a roadside laser radar calibration method is provided, the method comprising:

[0006] Obtaining the first point cloud data collected by the road-side laser radar to be calibrated;

[0007] Based on the initial position of the roadside laser radar, a local point cloud map satisfying a preset distance condition is determined from the point cloud map corresponding to the roadside laser radar; the point cloud map is obtained based on second point cloud data within a preset range of the roadside laser radar collected by the laser radar on the collection vehicle;

[0008] Optimizing the initial position and posture of the road-end laser radar based on the first point cloud data and the local point cloud map until a preset optimization condition is met, thereby obtaining a target position and posture of the road-end laser radar;

[0009] When the target posture meets the preset matching condition, the target posture is used as the external parameter of the road-side laser radar.

[0010] On the other hand, a roadside laser radar calibration device is provided, the device comprising:

[0011] A first acquisition module, used to acquire first point cloud data collected by a road-side laser radar to be calibrated;

[0012] A map extraction module, configured to determine a local point cloud map satisfying a preset distance condition from a point cloud map corresponding to the roadside laser radar based on an initial position of the roadside laser radar; the point cloud map is obtained based on second point cloud data within a preset range of the roadside laser radar collected by a laser radar on a collection vehicle;

[0013] An optimization processing module, used for optimizing the initial posture of the road-end laser radar based on the first point cloud data and the local point cloud map until a preset optimization condition is met, thereby obtaining a target posture of the road-end laser radar;

[0014] The matching judgment module is used to use the target posture as an external parameter of the road-side laser radar when the target posture meets the preset matching condition.

[0015] In an exemplary embodiment, the optimization processing module includes:

[0016] A feature point extraction module, used to extract at least one feature point based on the first point cloud data;

[0017] A feature point conversion module, used to use the initial posture as the current posture, and based on the current posture, convert the at least one feature point to a coordinate system corresponding to the local point cloud map to obtain at least one target feature point;

[0018] A first nearest neighbor point extraction module is used to extract a preset number of first nearest neighbor points corresponding to each target feature point from the local point cloud map based on the at least one target feature point, wherein the first nearest neighbor point is a point on the local point cloud map that is closest to the corresponding target feature point;

[0019] A first distance determination module is used to determine the distance between each target feature point and a corresponding preset number of first nearest neighbor points to obtain a first distance corresponding to each target feature point;

[0020] The target posture determination module is used to determine the target posture of the road-side laser radar based on the first distance corresponding to each target feature point.

[0021] In an exemplary embodiment, when the feature points include corner points, a preset number of first nearest neighbor points corresponding to each target corner point are linearly distributed; the first distance determination module includes:

[0022] A first determination module is used to determine, for each target corner point, a center point corresponding to a corresponding preset number of first nearest neighbor points and a main direction vector of the corresponding preset number of first nearest neighbor points;

[0023] A second determination module is used to determine a straight line which is parallel to the main direction vector and passes through the center point, and obtain a target straight line corresponding to the target corner point;

[0024] The third determination module is used to determine the distance between each target corner point and the corresponding target straight line to obtain the first distance corresponding to each target corner point.

[0025] In an exemplary embodiment, when the feature points include surface points, the first distance determination module includes:

[0026] A fourth determination module is used to determine, for each target surface point, a target plane where a corresponding preset number of first nearest neighbor points are located;

[0027] A fifth determination module, used to determine the normal direction of the target plane;

[0028] The sixth determination module is used to obtain a first distance corresponding to the target surface point based on the distance between the target surface point and the target plane in the normal direction.

[0029] In an exemplary embodiment, the target posture determination module includes:

[0030] A sum value determination module, used to determine the sum value of the first distance corresponding to each target feature point;

[0031] A current posture optimization module is used to update the current posture when the sum value is greater than or equal to a first preset threshold value, and execute the steps of converting the at least one feature point based on the current posture to the coordinate system corresponding to the local point cloud map to obtain at least one target feature point; extracting a preset number of first nearest neighbor points corresponding to each target feature point from the local point cloud map based on the at least one target feature point, wherein the first nearest neighbor point is a point on the local point cloud map that is closest to the corresponding target feature point; determining the distance between each target feature point and the corresponding preset number of first nearest neighbor points to obtain a first distance corresponding to each target feature point, until the sum value is less than the first preset threshold value;

[0032] The seventh determination module is used to use the current posture when the sum value is less than the first preset threshold as the target posture of the road-end laser radar.

[0033] In an exemplary embodiment, the matching judgment module includes:

[0034] A data conversion module, configured to convert the first point cloud data into a coordinate system corresponding to the local point cloud map based on the target posture to obtain a first target point cloud;

[0035] An external parameter determination module is used to determine the target posture as the external parameter of the road-end laser radar when the overlap between the point cloud map formed by the first target point cloud and the local point cloud map reaches a preset overlap.

[0036] In an exemplary embodiment, the external parameter determination module includes:

[0037] A second nearest neighbor point extraction module, configured to extract a second nearest neighbor point from the local point cloud map based on the first target point cloud, wherein the second nearest neighbor point is a point on the local point cloud map that is closest to the first target point cloud;

[0038] A second distance determination module, used to determine the distance between the first target point cloud and the second nearest neighbor point to obtain a second distance;

[0039] A mean value determination module, used to determine the mean value of the second distance;

[0040] An eighth determination module is used to determine the target posture as an external parameter of the road-end laser radar when the mean is less than a second preset threshold.

[0041] In an exemplary embodiment, the device further includes a map making module for making a point cloud map corresponding to the roadside laser radar, and the map making module includes:

[0042] A second acquisition module is used to acquire second point cloud data collected by the laser radar on the acquisition vehicle at multiple moments during the movement; the second point cloud data is point cloud data within a preset range of the road-side laser radar;

[0043] A map creation module, configured to select a first moment from the plurality of moments, and determine a current point cloud map based on second point cloud data collected at the first moment;

[0044] A map updating module, configured to select a second moment from the remaining moments, and update the current point cloud map based on second point cloud data collected at the second moment; the remaining moments refer to moments that are not selected from the multiple moments;

[0045] a map optimization module, configured to repeat the step of selecting a second moment from the remaining moments and updating the current point cloud map based on second point cloud data collected at the second moment, if the coverage degree between the updated current point cloud map and the preset range does not reach the preset coverage degree, until the coverage degree between the updated current point cloud map and the preset range reaches the preset coverage degree;

[0046] The point cloud map determination module is used to use the updated current point cloud map when the preset coverage level is reached as the point cloud map corresponding to the road-side laser radar.

[0047] On the other hand, an electronic device is provided, including a processor and a memory, wherein the memory stores at least one instruction or at least one program, and the at least one instruction or the at least one program is loaded and executed by the processor to implement the road-end lidar calibration method of any of the above aspects.

[0048] On the other hand, a computer-readable storage medium is provided, in which at least one instruction or at least one program is stored, and the at least one instruction or the at least one program is loaded and executed by a processor to implement a road-end lidar calibration method as described in any of the above aspects.

[0049] On the other hand, a computer program product or a computer program is provided, the computer program product or the computer program includes computer instructions, the computer instructions are stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the electronic device performs the road-side laser radar calibration method of any aspect above.

[0050] The embodiment of the present application produces a point cloud map by collecting point cloud data within a preset range of a road-end laser radar to be calibrated collected by a laser radar on a vehicle, extracts a local point cloud map that meets a preset distance condition from the point cloud map based on the initial posture of the road-end laser radar to be calibrated, optimizes the initial posture of the road-end laser radar to be calibrated based on the local point cloud map and the point cloud data collected by the road-end laser radar to be calibrated until the preset optimization condition is met, and obtains the target posture of the road-end laser radar to be calibrated. When the target posture meets the preset matching condition, the target posture is used as an external parameter of the road-end laser radar to be calibrated, thereby completing the calibration using the point cloud map when the road-end laser radar to be calibrated is not moving, thereby greatly improving the accuracy of the calibration parameters of the road-end laser radar, and achieving accurate positioning of obstacles. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments 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 these drawings without creative work.

[0052] Figure 1 It is a flowchart of a road-side laser radar calibration method provided in an embodiment of the present application;

[0053] Figure 2 is a schematic diagram of a process for optimizing the current posture provided in an embodiment of the present application;

[0054] Figure 3 is a schematic diagram of a process for calculating a target pose matching score provided in an embodiment of the present application;

[0055] Figure 4 It is a schematic diagram of the process of making a point cloud map provided in an embodiment of the present application;

[0056] Figure 5 It is a structural block diagram of a road-side laser radar calibration device provided in an embodiment of the present application;

[0057] Figure 6 It is a hardware structure block diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0058] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0059] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0060] It is understandable that in the specific implementation of this application, related data such as user information is involved. When the above embodiments of this application are applied to specific products or technologies, user permission or consent is required, and the collection, use and processing of relevant data need to comply with relevant laws, regulations and standards of relevant countries and regions.

[0061] See also Figure 1, which shows a flow chart of a road-side lidar calibration method provided in an embodiment of the present application. It should be noted that this specification provides method operation steps as described in the embodiment or flow chart, but may include more or fewer operation steps based on conventional or non-creative labor. The order of steps listed in the embodiment is only one way of executing the steps among many orders, and does not represent the only order of execution. When the system or product is executed in practice, it can be executed in the order of the method shown in the embodiment or the accompanying drawings or in parallel (for example, in a parallel processor or multi-threaded processing environment). Specifically, Figure 1 As shown, the method may include:

[0062] S101, obtaining first point cloud data collected by a road-side laser radar to be calibrated.

[0063] Among them, the first point cloud data is the point cloud data of the object within the scanning range of the road-end laser radar to be calibrated.

[0064] S103, based on the initial position and posture of the road-end laser radar, determine a local point cloud map that meets a preset distance condition from the point cloud map corresponding to the road-end laser radar.

[0065] The point cloud map is obtained based on the second point cloud data within the preset range of the road-end laser radar collected by the laser radar on the collection vehicle. In a specific implementation, the collection vehicle can move around the road-end laser radar to be calibrated to collect point cloud data within the preset range of the road-end laser radar through the laser radar on the collection vehicle, and the point cloud data is used to generate a point cloud map corresponding to the road-end laser radar.

[0066] Among them, the preset range is the range of point cloud data around the road-side laser radar to be calibrated that needs to be collected by the laser radar on the collection vehicle.

[0067] Among them, the initial posture is obtained by using GPS measurement at the installation location of the road-side laser radar to be calibrated. It is the posture of the road-side laser radar to be calibrated with a large error compared to the actual posture of the road-side laser radar. It is expressed in the form of a transformation matrix, including the position of the road-side laser radar to be calibrated and the posture of the road-side laser radar to be calibrated. The position is expressed by the coordinates of the road-side laser radar to be calibrated in the world coordinate system, and the posture indicates the direction of the road-side laser radar to be calibrated. The external parameters of the road-side laser radar to be calibrated are determined by optimizing the initial posture.

[0068] Among them, the preset distance condition needs to be set in advance, and is set according to the size of the required local point cloud map. The larger the distance, the greater the amount of calculation in the optimization process and matching process of this embodiment, and of course the more accurate the final calibration parameters. Take the coordinate information in the initial posture as the center of the circle, and the preset distance condition as the radius to extract the local point cloud map from the point cloud map. In this embodiment, the preset distance condition is 100 meters. According to the initial posture of the road-end laser radar to be calibrated, the local point cloud map within 100 meters around the road-end laser radar to be calibrated is extracted from the point cloud map.

[0069] S105, optimizing the initial posture of the road-end laser radar based on the first point cloud data and the local point cloud map until a preset optimization condition is met, thereby obtaining a target posture of the road-end laser radar.

[0070] Among them, optimization processing refers to using the initial pose to transform the points in the first point cloud data into the coordinate system of the local point cloud map, that is, the world coordinate system, and obtaining a gradually optimized pose by continuously reducing the error of the conversion result compared to the actual coordinates of the first point cloud data in the world coordinate system.

[0071] Among them, the preset optimization condition is to use the optimized posture to convert the points in the first point cloud data to the coordinate system of the local point cloud map, and the error generated has been minimized.

[0072] In an exemplary embodiment, Figure 2 As shown, the above step S105 may include:

[0073] S201: extract at least one feature point based on the first point cloud data.

[0074] The first point cloud data is traversed, and points meeting the feature conditions are extracted as feature points. The feature conditions are used to screen points in the first point cloud data to obtain feature points. The specific feature conditions can be set based on the type of feature points, and different feature conditions can be set for different types of feature points.

[0075] S203: Taking the initial posture as the current posture.

[0076] To optimize the initial pose of the road-side LiDAR to be calibrated, this initial pose with a large error must enter the first round of optimization process.

[0077] S205: Based on the current posture, convert the at least one feature point to a coordinate system corresponding to the local point cloud map to obtain at least one target feature point.

[0078] Use the current pose to convert the feature points to the world coordinate system corresponding to the local point cloud map to obtain the target feature points corresponding to the feature points. The error between the coordinates of the target feature points and the actual coordinates of the feature points in the world coordinate system reflects the error of the initial pose, which is the idea of ​​​​subsequent optimization of the initial pose.

[0079] S207: extracting a preset number of first nearest neighbor points corresponding to each target feature point from the local point cloud map based on the at least one target feature point.

[0080] The first nearest neighbor point is the point on the local point cloud map that is closest to the corresponding target feature point.

[0081] Traverse each target feature point, traverse the local point cloud map starting from the current target feature point, and extract a preset number of first nearest neighbor points corresponding to each target feature point, wherein the corresponding preset number of first nearest neighbor points need to be able to determine at least one edge or one plane.

[0082] S209, determining the distance between each target feature point and a corresponding preset number of first nearest neighboring points, and obtaining a first distance corresponding to each target feature point.

[0083] The corresponding preset number of first nearest neighbor points may form a plane or an edge, and the first distance corresponding to each target feature point is the distance between each target feature point and the corresponding plane or edge.

[0084] When the feature points include corner points, a preset number of first nearest neighbor points corresponding to each target corner point are linearly distributed, and the above step S209 may include the following steps:

[0085] For each target corner point, determine the center point corresponding to a corresponding preset number of first nearest neighbor points and the main direction vectors of the corresponding preset number of first nearest neighbor points;

[0086] Determine a straight line that is parallel to the main direction vector and passes through the center point, and obtain a target straight line corresponding to the target corner point;

[0087] The distance between each target corner point and the corresponding target straight line is determined to obtain a first distance corresponding to each target corner point.

[0088] Among them, the corner point is an edge point with a larger curvature, and the edge is the edge of an object within the scanning range of the road-side laser radar, such as the edge of a tree trunk, etc. The curvature here refers to the curvature of the edge at the corner point.

[0089] When extracting corner points, it is necessary to determine whether a certain point is a corner point. Whether it is a corner point is determined by calculating the curvature. A certain number of points around the point are extracted on the laser beam where the point is located. An edge is determined based on the point and these surrounding points. The curvature of the edge at the point is calculated. If the curvature is greater than the preset value, the point is determined to be a corner point.

[0090] Among them, before determining the preset number of first nearest neighbor points corresponding to a target corner point, it is necessary to calculate the covariance matrix of the preset number of points extracted from the local point cloud map that are closest to the target corner point, calculate the eigenvalues ​​and eigenvectors of the covariance matrix, and determine whether the preset number of points are distributed in a linear manner. If so, the preset number of points are determined as the preset number of first nearest neighbor points corresponding to the target corner point.

[0091] Among them, the first distance is the distance between the target corner point and the edge where the corresponding preset number of first nearest neighbor points are located. This is the error generated in the process of converting the corner point into the target corner point on the local point cloud map using the current posture.

[0092] It can be seen from the above technical solutions of the embodiments of the present invention that the embodiments of the present invention correspond to the error of the current posture by the distance between the target corner point and the edge where the corresponding preset number of first nearest neighbor points are located, thereby providing an entry point for subsequent optimization of the current posture.

[0093] When the feature points include surface points, the above step S209 may include the following steps:

[0094] For each target surface point, determining the target plane where a corresponding preset number of first nearest neighbor points are located;

[0095] determining a normal direction of the target plane;

[0096] Based on the distance between the target surface point and the target plane in the normal direction, a first distance corresponding to the target surface point is obtained.

[0097] Among them, the surface point is a plane point with a smaller curvature, and the plane is a plane within the scanning range of the road-end laser radar, such as the ground, wall, etc. Here, the curvature refers to the curvature of the plane at the surface point.

[0098] When extracting surface points, it is necessary to determine whether a certain point is a surface point. Whether it is a surface point is determined by calculating the curvature. A certain number of points around the point are extracted on the laser beam where the point is located. A surface is determined based on the point and these surrounding points. The curvature of the surface at the point is calculated. If the curvature is less than the preset value, the point is a surface point.

[0099] Among them, the first distance is the distance between the target surface point and the plane where the corresponding preset number of first nearest neighbor points are located. This is the error generated in the process of converting the surface point into the target surface point on the local point cloud map using the current posture.

[0100] When the feature points in the above step S209 include corner points and surface points at the same time, the specific implementation method corresponds to the methods provided in the above-mentioned embodiments. Therefore, the implementation method of the aforementioned road-side laser radar calibration method is also applicable to the road-side laser radar calibration method provided in this embodiment, and will not be described in detail in this embodiment.

[0101] It can be seen from the above technical solutions of the embodiment of the present invention that the embodiment of the present invention provides an entry point for subsequent optimization of the current posture by corresponding the error of the current posture to the distance between the target surface point and the plane where the corresponding preset number of first nearest neighbor points are located.

[0102] S2011, determining the target posture of the road-end laser radar based on the first distance corresponding to each target feature point.

[0103] Among them, the first distance can reflect the error of the initial posture to a certain extent. Based on this error, the initial posture is optimized to obtain the target posture of the road-side laser radar.

[0104] A first distance is obtained based on the current posture, the current posture is optimized based on the first distance, a smaller first distance is obtained based on the optimized current posture, and the corresponding current posture is optimized based on the smaller first distance until the first distance is minimized. The current posture that minimizes the first distance is determined as the target posture of the road-end laser radar to be calibrated.

[0105] It can be seen from the above technical solutions of the embodiment of the present invention that the embodiment of the present invention converts feature points into target feature points in the local point cloud map through the current posture, aligns the distance between the target feature points and the corresponding preset number of first nearest neighbor points with the error of the current posture, and continuously reduces the distance, thereby realizing continuous optimization of the current posture.

[0106] In an exemplary embodiment, please continue to see Figure 2 , the above step S2011 may include:

[0107] S20111, determine the sum of the first distances corresponding to each target feature point.

[0108] Among them, the target feature point is the point obtained by converting the feature point to the coordinate system of the local point cloud map, that is, the world coordinate system, using the current posture. The sum value is an overall reflection of the error between the coordinates of the target feature point and the actual coordinates of the feature point in the world coordinate system, and this error reflects the error between the posture of converting the feature point to the target feature point and the actual posture of the roadside laser radar. Based on the sum value of the first distance, it is judged whether the error generated by converting the feature point to the coordinate system corresponding to the local point cloud map using the current posture has reached the minimum.

[0109] S20113, determine whether the sum is less than a first preset threshold.

[0110] Specifically, if the result of the judgment is no, step S20115 can be executed; otherwise, if the result of the judgment is yes, step S20117 can be executed.

[0111] Among them, the first preset threshold is set according to the error between the coordinates of the acceptable target feature point and the actual coordinates of the feature point in the world coordinate system. The smaller the first preset threshold is set, the more accurate the target posture is, and the more accurate the calibration parameters are.

[0112] S20115, update the current posture and execute step S205.

[0113] This is a cyclic optimization process. The optimized current pose is used to transform the feature points to the coordinate system corresponding to the local point cloud map. The current pose is optimized by the error of the transformation result until the current pose can minimize the error of the above transformation result. Then the optimization of the initial pose is completed.

[0114] S20117, taking the current posture when the sum value is less than the first preset threshold as the target posture of the road-end laser radar.

[0115] Among them, the current posture when the sum value is less than the first preset threshold refers to the current posture that can minimize the error caused by converting the feature point to the coordinate system corresponding to the local point cloud map, that is, the optimized posture.

[0116] It can be seen from the above technical scheme of the embodiment of the present invention that the embodiment of the present invention determines whether the distance between the target feature point and the corresponding preset number of first nearest neighbor points has reached the minimum, determines whether the current posture has reached the optimal one, and decides whether to continue to optimize the current posture based on this, thereby ensuring that the error of the target posture is minimized.

[0117] S107, when the target posture meets a preset matching condition, using the target posture as an external parameter of the road-end laser radar.

[0118] Among them, the preset matching condition is that the matching score of the target posture is less than a specific threshold.

[0119] The matching score of the target pose is calculated to verify the result of the optimization process. If the matching score is less than a specific threshold, the verification is passed and the optimization result, i.e., the target pose, is used as the external parameter of the road-side lidar to be calibrated.

[0120] It can be seen from the above technical scheme of the embodiment of the present invention that the embodiment of the present invention places the second point cloud data within the preset range of the road-end laser radar to be calibrated collected by the laser radar on the collecting vehicle and the first point cloud data collected by the road-end laser radar to be calibrated in the same coordinate system, compares the degree of matching between the two in the form of a point cloud map, and optimizes the initial posture of the road-end laser radar to be calibrated by continuously reducing the error between the two, thereby obtaining the external parameter parameters of the road-end laser radar to be calibrated, which greatly improves the accuracy of the calibration parameters of the road-end laser radar and can achieve accurate positioning of obstacles.

[0121] In an exemplary embodiment, Figure 3 As shown, the above step S107 may include:

[0122] S301: Convert the first point cloud data into a coordinate system corresponding to the local point cloud map based on the target posture to obtain a first target point cloud.

[0123] Among them, the target posture is used to convert the first point cloud data into the coordinate system corresponding to the local point cloud map, the error generated by this conversion process is calculated, and it is judged whether the error is within an acceptable range, so as to determine whether the target posture meets the standards as the external parameter of the road-side laser radar to be calibrated.

[0124] S303: When the degree of overlap between the point cloud map formed by the first target point cloud and the local point cloud map reaches a preset degree of overlap, the target posture is determined as an external parameter of the road-end laser radar.

[0125] The overlap between the point cloud map and the local point cloud map corresponds to the error generated by the above conversion process, that is, the error of the target pose, which is also the matching score. The preset overlap is a specific threshold, and when the error is less than the threshold, the match is qualified.

[0126] It can be seen from the above technical scheme of the embodiment of the present invention that the embodiment of the present invention determines whether the error of the target posture is within an acceptable range by judging whether the overlap between the point cloud map composed of the first target point cloud and the point cloud map of the headquarters is less than a preset overlap, and thereby judges whether the target posture meets the standards for the external parameter parameters of the road-side laser radar to be calibrated, thereby greatly improving the accuracy of the external parameter parameters of the road-side laser radar to be calibrated.

[0127] In an exemplary embodiment, please continue to see Figure 3 , the above step S303 may include:

[0128] S3031: Extract a second nearest neighbor point from the local point cloud map based on the first target point cloud.

[0129] The second nearest neighbor point is the point on the local point cloud map that is closest to the first target point cloud.

[0130] Traverse the first target point cloud, traverse the local point cloud map from the current target point, and extract the second nearest neighbor point corresponding to each target point.

[0131] S3033: Determine the distance between the first target point cloud and the second nearest neighbor point to obtain a second distance.

[0132] The second distance is an error generated in a process of converting the first point cloud data into the first target point cloud on the local point cloud map using the target pose.

[0133] S3035: Determine the mean of the second distances.

[0134] Among them, the mean is the matching score. Based on the average value of the second distance, it is judged whether the error generated by using the target posture to convert the first point cloud data to the coordinate system corresponding to the local point cloud map is within an acceptable range, that is, whether the error of the target posture is within an acceptable range.

[0135] S3037: When the mean is less than a second preset threshold, the target posture is determined as an external parameter of the road-end laser radar.

[0136] Among them, if the mean is less than the second preset threshold, the error of the target posture is within an acceptable range, and the target posture meets the standard of being an external parameter of the road-side laser radar to be calibrated.

[0137] It can be seen from the above technical scheme of the embodiment of the present invention that the embodiment of the present invention calculates the matching score of the target posture by the distance between the first target point cloud and the second nearest neighbor point on the corresponding local point cloud map, and judges whether the matching score of the target posture meets the standard, and judges whether the target posture meets the standard as the external parameter parameter of the road-side laser radar to be calibrated, thereby greatly improving the accuracy of the external parameter parameters of the road-side laser radar to be calibrated.

[0138] The following describes the preparation of a point cloud map before determining a local point cloud map that meets a preset distance condition from a point cloud map corresponding to the road-side laser radar based on the initial posture of the road-side laser radar. Specifically, the following steps may be included:

[0139] S401, obtaining second point cloud data collected by a laser radar on a collection vehicle at multiple times during movement; the second point cloud data is point cloud data within a preset range of the road-end laser radar.

[0140] Among them, the second point cloud data is point cloud data of objects collected by the laser radar on the collection vehicle when the collection vehicle moves near the road-end laser radar to be calibrated.

[0141] S403: Select a first moment from the multiple moments, and determine a current point cloud map based on second point cloud data collected at the first moment.

[0142] A first moment is randomly selected from multiple moments, and positioning data of second point cloud data collected at the first moment is obtained based on RTK real-time dynamic positioning technology. A current point cloud map is produced based on the second point cloud data and the positioning data. The current point cloud map obtained based on the second point cloud data collected at the first moment is the base map for subsequent map updates.

[0143] S405: Select a second moment from the remaining moments, and update the current point cloud map based on second point cloud data collected at the second moment.

[0144] The remaining moments refer to the moments that are not selected from the multiple moments;

[0145] Based on the RTK real-time dynamic positioning technology, the positioning data of the second point cloud data collected at the second moment is obtained, the point cloud map at the second moment is determined based on the second point cloud data and the positioning data, and the point cloud map is supplemented to the current point cloud map, thereby completing the update of the current point cloud map.

[0146] S407, determining whether the coverage degree between the updated current point cloud map and the preset range reaches a preset coverage degree.

[0147] The preset coverage degree is a specific value used to determine whether the coverage rate of the current point cloud map for the preset range meets the requirements.

[0148] Specifically, if the result of the judgment is no, step S405 can be executed; otherwise, if the result of the judgment is yes, step S409 can be executed.

[0149] S409, using the updated current point cloud map when the preset coverage level is reached as the point cloud map corresponding to the road-end laser radar.

[0150] This is a cyclic updating process, which continuously expands the coverage of the current point cloud map by collecting the second point cloud data at different times until the current point cloud map can cover the preset range near the road-end laser radar to be calibrated.

[0151] It can be seen from the above technical scheme of the embodiment of the present invention that the embodiment of the present invention produces a point cloud map by collecting point cloud data within a preset range of the road-end laser radar to be calibrated collected by the laser radar on the vehicle, and uses the point cloud map as a reference for optimizing the initial posture of the road-end laser radar to be calibrated. Point cloud data of different ranges at different times are collected by collecting the movement of the vehicle, and point cloud maps of different ranges are corresponded to point cloud maps of different ranges at different times. The current point cloud map is updated by producing point cloud maps at different times, and the point cloud maps of different ranges are spliced ​​together, thereby ensuring that the final point cloud map covers the preset range of the road-end laser radar to be calibrated.

[0152] Corresponding to the road-end lidar calibration methods provided in the above-mentioned embodiments, the embodiment of the present application also provides a road-end lidar calibration device. Since the road-end lidar calibration device provided in the embodiment of the present application corresponds to the road-end lidar calibration methods provided in the above-mentioned embodiments, the implementation methods of the aforementioned road-end lidar calibration methods are also applicable to the road-end lidar calibration device provided in this embodiment and will not be described in detail in this embodiment.

[0153] See also Figure 5 , which is a schematic diagram of the structure of a roadside laser radar calibration device provided in an embodiment of the present application, the device has the function of implementing the roadside laser radar calibration method in the above method embodiment, and the function can be implemented by hardware or by hardware executing corresponding software. Figure 5 As shown, the roadside laser radar calibration device 500 may include:

[0154] A first acquisition module 510 is used to acquire first point cloud data collected by a road-side laser radar to be calibrated;

[0155] A map extraction module 520 is used to determine a local point cloud map that meets a preset distance condition from a point cloud map corresponding to the roadside laser radar based on an initial position of the roadside laser radar; the point cloud map is obtained based on second point cloud data within a preset range of the roadside laser radar collected by the laser radar on the collection vehicle;

[0156] An optimization processing module 530 is used to optimize the initial posture of the road-end laser radar based on the first point cloud data and the local point cloud map until a preset optimization condition is met to obtain a target posture of the road-end laser radar;

[0157] The matching judgment module 540 is used to use the target posture as an external parameter of the road-side laser radar when the target posture meets the preset matching condition.

[0158] In an exemplary embodiment, the optimization processing module includes:

[0159] A feature point extraction module, used to extract at least one feature point based on the first point cloud data;

[0160] A feature point conversion module, used to use the initial posture as the current posture, and based on the current posture, convert the at least one feature point to a coordinate system corresponding to the local point cloud map to obtain at least one target feature point;

[0161] A first nearest neighbor point extraction module is used to extract a preset number of first nearest neighbor points corresponding to each target feature point from the local point cloud map based on the at least one target feature point, wherein the first nearest neighbor point is a point on the local point cloud map that is closest to the corresponding target feature point;

[0162] A first distance determination module is used to determine the distance between each target feature point and a corresponding preset number of first nearest neighbor points to obtain a first distance corresponding to each target feature point;

[0163] The target posture determination module is used to determine the target posture of the road-side laser radar based on the first distance corresponding to each target feature point.

[0164] In an exemplary embodiment, when the feature points include corner points, a preset number of first nearest neighbor points corresponding to each target corner point are linearly distributed; the first distance determination module includes:

[0165] A first determination module is used to determine, for each target corner point, a center point corresponding to a corresponding preset number of first nearest neighbor points and a main direction vector of the corresponding preset number of first nearest neighbor points;

[0166] A second determination module is used to determine a straight line which is parallel to the main direction vector and passes through the center point, and obtain a target straight line corresponding to the target corner point;

[0167] The third determination module is used to determine the distance between each target corner point and the corresponding target straight line to obtain the first distance corresponding to each target corner point.

[0168] In an exemplary embodiment, when the feature points include surface points, the first distance determination module includes:

[0169] A fourth determination module is used to determine, for each target surface point, a target plane where a corresponding preset number of first nearest neighbor points are located;

[0170] A fifth determination module, used to determine the normal direction of the target plane;

[0171] The sixth determination module is used to obtain a first distance corresponding to the target surface point based on the distance between the target surface point and the target plane in the normal direction.

[0172] In an exemplary embodiment, the target posture determination module includes:

[0173] A sum value determination module, used to determine the sum value of the first distance corresponding to each target feature point;

[0174] A current posture optimization module is used to update the current posture when the sum value is greater than or equal to a first preset threshold value, and execute the steps of converting the at least one feature point based on the current posture to the coordinate system corresponding to the local point cloud map to obtain at least one target feature point; extracting a preset number of first nearest neighbor points corresponding to each target feature point from the local point cloud map based on the at least one target feature point, wherein the first nearest neighbor point is a point on the local point cloud map that is closest to the corresponding target feature point; determining the distance between each target feature point and the corresponding preset number of first nearest neighbor points to obtain a first distance corresponding to each target feature point, until the sum value is less than the first preset threshold value;

[0175] The seventh determination module is used to use the current posture when the sum value is less than the first preset threshold as the target posture of the road-end laser radar.

[0176] In an exemplary embodiment, the matching judgment module includes:

[0177] A data conversion module, configured to convert the first point cloud data into a coordinate system corresponding to the local point cloud map based on the target posture to obtain a first target point cloud;

[0178] An external parameter determination module is used to determine the target posture as the external parameter of the road-end laser radar when the overlap between the point cloud map formed by the first target point cloud and the local point cloud map reaches a preset overlap.

[0179] In an exemplary embodiment, the external parameter determination module includes:

[0180] A second nearest neighbor point extraction module, configured to extract a second nearest neighbor point from the local point cloud map based on the first target point cloud, wherein the second nearest neighbor point is a point on the local point cloud map that is closest to the first target point cloud;

[0181] A second distance determination module, used to determine the distance between the first target point cloud and the second nearest neighbor point to obtain a second distance;

[0182] A mean value determination module, used to determine the mean value of the second distance;

[0183] An eighth determination module is used to determine the target posture as an external parameter of the road-end laser radar when the mean is less than a second preset threshold.

[0184] In an exemplary embodiment, the device further includes a map making module for making a point cloud map corresponding to the roadside laser radar, and the map making module includes:

[0185] A second acquisition module is used to acquire second point cloud data collected by the laser radar on the acquisition vehicle at multiple moments during the movement; the second point cloud data is point cloud data within a preset range of the road-side laser radar;

[0186] A map creation module, configured to select a first moment from the plurality of moments, and determine a current point cloud map based on second point cloud data collected at the first moment;

[0187] A map updating module, configured to select a second moment from the remaining moments, and update the current point cloud map based on second point cloud data collected at the second moment; the remaining moments refer to moments that are not selected from the multiple moments;

[0188] a map optimization module, configured to repeat the step of selecting a second moment from the remaining moments and updating the current point cloud map based on second point cloud data collected at the second moment, if the coverage degree between the updated current point cloud map and the preset range does not reach the preset coverage degree, until the coverage degree between the updated current point cloud map and the preset range reaches the preset coverage degree;

[0189] The point cloud map determination module is used to use the updated current point cloud map when the preset coverage level is reached as the point cloud map corresponding to the road-side laser radar.

[0190] It should be noted that the device provided in the above embodiment, when implementing its functions, is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the device and method embodiments provided in the above embodiment belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0191] An embodiment of the present application provides an electronic device, which includes a processor and a memory, wherein the memory stores at least one instruction or at least one program, and the at least one instruction or the at least one program is loaded and executed by the processor to implement any one of the road-end lidar calibration methods provided in the above method embodiments.

[0192] The memory can be used to store software programs and modules. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, application programs required for functions, etc.; the data storage area can store data created according to the use of the device, etc. In addition, the memory can include a high-speed random access memory and can also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage devices. Accordingly, the memory can also include a memory controller to provide the processor with access to the memory.

[0193] The method embodiments provided in the embodiments of the present application may be executed in a computer terminal, a server or a similar computing device, that is, the above-mentioned electronic device may include a computer terminal, a server or a similar computing device. Figure 6 is a hardware structure block diagram of an electronic device for running a road-side laser radar calibration method provided in an embodiment of the present application, such as Figure 6 As shown, the internal structure of the electronic device may include but is not limited to: a processor, a network interface and a memory. The processor, network interface and memory in the electronic device may be connected via a bus or other means. Figure 6 The example of connecting through bus is taken in the following.

[0194] Among them, the processor (or CPU (Central Processing Unit)) is the computing core and control core of the computer device. The network interface may optionally include a standard wired interface, a wireless interface (such as WI-FI, a mobile communication interface, etc.). The memory is a memory device in a computer device for storing programs and data. It is understandable that the memory here can be a high-speed RAM storage device or a non-volatile memory device (non-volatile memory), such as at least one disk storage device; optionally, it can also be at least one storage device located away from the aforementioned processor. The memory provides a storage space, which stores the operating system of the electronic device, which may include but is not limited to: Windows system (an operating system), Linux (an operating system), Android (Android, a mobile operating system) system, IOS (a mobile operating system) system, etc., and the present invention is not limited to this; and, in the storage space, one or more instructions suitable for being loaded and executed by the processor are also stored, and these instructions may be one or more computer programs (including program codes). In the embodiment of this specification, the processor loads and executes one or more instructions stored in the memory to implement the road-side laser radar calibration method provided in the above method embodiment.

[0195] An embodiment of the present application also provides a computer-readable storage medium, which can be set in an electronic device to store at least one instruction or at least one program related to implementing a road-end lidar calibration method. The at least one instruction or the at least one program is loaded and executed by the processor to implement any road-end lidar calibration method provided by the above method embodiments.

[0196] Optionally, in this embodiment, the above-mentioned storage medium may include but is not limited to: a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and other media that can store program codes.

[0197] It should be noted that the above-mentioned sequence of the embodiments of the present application is for description only and does not represent the advantages and disadvantages of the embodiments. The above-mentioned specific embodiments of this specification are described. Other embodiments are within the scope of the attached claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0198] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0199] A person skilled in the art will understand that all or part of the steps to implement the above embodiments may be accomplished by hardware or by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a disk or an optical disk, etc.

[0200] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A roadside laser radar calibration method, characterized in that: The method comprises: Obtaining the first point cloud data collected by the road-side laser radar to be calibrated; Based on the initial position of the roadside laser radar, a local point cloud map satisfying a preset distance condition is determined from the point cloud map corresponding to the roadside laser radar; the point cloud map is obtained based on second point cloud data within a preset range of the roadside laser radar collected by the laser radar on the collection vehicle; Optimizing the initial position and posture of the road-end laser radar based on the first point cloud data and the local point cloud map until a preset optimization condition is met, thereby obtaining a target position and posture of the road-end laser radar; When the target posture meets the preset matching condition, the target posture is used as an external parameter of the road-side laser radar; Among them, the initial posture of the road-end laser radar is optimized based on the first point cloud data and the local point cloud map until the preset optimization conditions are met to obtain the target posture of the road-end laser radar, including: extracting at least one feature point based on the first point cloud data; taking the initial posture as the current posture, and converting the at least one feature point to the coordinate system corresponding to the local point cloud map based on the current posture to obtain at least one target feature point; extracting a preset number of first nearest neighbor points corresponding to each target feature point from the local point cloud map based on the at least one target feature point, the first nearest neighbor point being the point on the local point cloud map that is closest to the corresponding target feature point; determining the distance between each target feature point and the corresponding preset number of first nearest neighbor points to obtain the first distance corresponding to each target feature point; determining the target posture of the road-end laser radar based on the first distance corresponding to each target feature point; The preset number of first nearest neighbor points are used to determine an edge or a plane, and the first distance corresponding to each target feature point is the distance between each target feature point and the corresponding edge or plane.

2. A roadside laser radar calibration method according to claim 1, characterized in that: When the feature points include corner points, a preset number of first nearest neighbor points corresponding to each target corner point are linearly distributed; determining the distance between each target feature point and the corresponding preset number of first nearest neighbor points to obtain the first distance corresponding to each target feature point includes: For each target corner point, determine the center point corresponding to a corresponding preset number of first nearest neighbor points and the main direction vectors of the corresponding preset number of first nearest neighbor points; Determine a straight line that is parallel to the main direction vector and passes through the center point, and obtain a target straight line corresponding to the target corner point; The distance between each target corner point and the corresponding target straight line is determined to obtain a first distance corresponding to each target corner point.

3. A roadside laser radar calibration method according to claim 1 or 2, characterized in that: When the feature points include surface points, determining the distance between each target feature point and a corresponding preset number of first nearest neighboring points to obtain a first distance corresponding to each target feature point includes: For each target surface point, determining the target plane where a corresponding preset number of first nearest neighbor points are located; determining a normal direction of the target plane; Based on the distance between the target surface point and the target plane in the normal direction, a first distance corresponding to the target surface point is obtained.

4. The roadside laser radar calibration method according to claim 1, characterized in that: The determining the target posture of the roadside laser radar based on the first distance corresponding to each target feature point includes: Determine the sum of the first distances corresponding to each target feature point; In the case where the sum is greater than or equal to a first preset threshold, the current posture is updated, and the steps of converting the at least one feature point based on the current posture to a coordinate system corresponding to the local point cloud map to obtain at least one target feature point are performed; based on the at least one target feature point, a preset number of first nearest neighbor points corresponding to each target feature point are extracted from the local point cloud map, where the first nearest neighbor point is the point on the local point cloud map that is closest to the corresponding target feature point; the distance between each target feature point and the corresponding preset number of first nearest neighbor points is determined to obtain a first distance corresponding to each target feature point, until the sum is less than the first preset threshold; The current posture when the sum value is less than the first preset threshold is used as the target posture of the road-end laser radar.

5. The roadside laser radar calibration method according to claim 1, characterized in that: When the target posture meets the preset matching condition, the target posture is used as an external parameter of the road-side laser radar, including: Based on the target posture, the first point cloud data is converted into a coordinate system corresponding to the local point cloud map to obtain a first target point cloud; When the degree of overlap between the point cloud map formed by the first target point cloud and the local point cloud map reaches a preset degree of overlap, the target posture is determined as the external parameter of the road-end laser radar.

6. A roadside laser radar calibration method according to claim 5, characterized in that: When the overlap between the point cloud map formed by the first target point cloud and the local point cloud map reaches a preset overlap, determining the target pose as an external parameter of the roadside laser radar includes: Extracting a second nearest neighbor point from the local point cloud map based on the first target point cloud, where the second nearest neighbor point is a point on the local point cloud map that is closest to the first target point cloud; Determine the distance between the first target point cloud and the second nearest neighbor point to obtain a second distance; determining a mean of the second distances; When the mean is less than a second preset threshold, the target posture is determined as an external parameter of the road-end laser radar.

7. A roadside laser radar calibration method according to claim 1, characterized in that: Before determining a local point cloud map that meets a preset distance condition from a point cloud map corresponding to the roadside laser radar based on the initial position of the roadside laser radar, the method further includes: Acquire second point cloud data collected by a laser radar on the collection vehicle at multiple times during the movement; the second point cloud data is point cloud data within a preset range of the road-side laser radar; Selecting a first moment from the multiple moments, and determining a current point cloud map based on second point cloud data collected at the first moment; Select a second moment from the remaining moments, and update the current point cloud map based on the second point cloud data collected at the second moment; the remaining moments refer to moments that are not selected from the multiple moments; In the case that the coverage degree between the updated current point cloud map and the preset range does not reach the preset coverage degree, repeating the step of selecting a second moment from the remaining moments and updating the current point cloud map based on the second point cloud data collected at the second moment, until the coverage degree between the updated current point cloud map and the preset range reaches the preset coverage degree; The updated current point cloud map when the preset coverage level is reached is used as the point cloud map corresponding to the road-side laser radar.

8. A roadside laser radar calibration device, characterized in that: The device comprises: A first acquisition module, used to acquire first point cloud data collected by a road-side laser radar to be calibrated; A map extraction module, configured to determine a local point cloud map satisfying a preset distance condition from a point cloud map corresponding to the roadside laser radar based on an initial position of the roadside laser radar; the point cloud map is obtained based on second point cloud data within a preset range of the roadside laser radar collected by a laser radar on a collection vehicle; An optimization processing module, used for optimizing the initial posture of the road-end laser radar based on the first point cloud data and the local point cloud map until a preset optimization condition is met, thereby obtaining a target posture of the road-end laser radar; A matching judgment module, configured to use the target posture as an external parameter of the road-side laser radar when the target posture meets a preset matching condition; Wherein, the optimization processing module includes: A feature point extraction module, used to extract at least one feature point based on the first point cloud data; A feature point conversion module, used to use the initial posture as the current posture, and based on the current posture, convert the at least one feature point to a coordinate system corresponding to the local point cloud map to obtain at least one target feature point; A first nearest neighbor point extraction module is used to extract a preset number of first nearest neighbor points corresponding to each target feature point from the local point cloud map based on the at least one target feature point, wherein the first nearest neighbor point is a point on the local point cloud map that is closest to the corresponding target feature point; A first distance determination module is used to determine the distance between each target feature point and a corresponding preset number of first nearest neighbor points to obtain a first distance corresponding to each target feature point; A target posture determination module, used to determine the target posture of the road-side laser radar based on the first distance corresponding to each target feature point; The preset number of first nearest neighbor points are used to determine an edge or a plane, and the first distance corresponding to each target feature point is the distance between each target feature point and the corresponding edge or plane.

9. An electronic device, characterized in that: It includes a processor and a memory, wherein the memory stores at least one instruction or at least one program, and the at least one instruction or the at least one program is loaded and executed by the processor to implement the road-end laser radar calibration method as described in any one of claims 1 to 7.

Images