Method and device for parameter correction

Abstract

This specification discloses a method and device for parameter correction. By optimizing the external parameters corresponding to each laser radar according to the point cloud data collected by the laser radar and the posture based on which the laser radar collects the point cloud data, and correcting the external parameters of each laser radar based on the optimized external parameters of each laser radar, the external parameters of each laser radar can be corrected in real time, and the problem of lack of consistency between the external parameters of different laser radars can be avoided, thereby providing a guarantee for the execution of tasks that require the use of external parameters.

Description

Technical Field

[0001] This specification relates to the field of unmanned driving technology, and in particular to a method and device for parameter correction. Background Art

[0002] When an autonomous driving device performs a task based on the point cloud data collected by the lidar and the posture data collected by the integrated navigation device, it is usually necessary to fuse the data collected by the lidar and the integrated navigation device based on the preset external parameters between the lidar and the integrated navigation device, and then perform the task based on the fused data.

[0003] However, since the sensors installed in the acquisition equipment often become loose during daily operation, there will be a deviation between the preset external parameters and the actual external parameters, which in turn affects the tasks performed by the autonomous driving equipment.

[0004] In the existing technology, staff usually recalibrate the preset external parameters at regular intervals. However, since the manually set period may not match the actual correction period required when the external parameters are affected by loose equipment, when there is a large deviation between the preset external parameters and the actual external parameters, they cannot be recalibrated in time. When the external parameters are recalibrated, it is found that the deviation between the recalibrated external parameters and the original external parameters is small and does not affect the task performed by the autonomous driving equipment, resulting in a waste of resources. In addition, since the manual recalibration of the preset external parameters is performed independently for each lidar, it also leads to a lack of consistency between the external parameters of each radar sensor, which in turn affects the task performed by the autonomous driving equipment.

[0005] Therefore, how to calibrate the external parameters of the lidar in a timely and effective manner to ensure that the autonomous driving equipment can perform its tasks normally is an urgent problem to be solved. Summary of the Invention

[0006] This specification provides a parameter calibration method and device to partially solve the above-mentioned problems existing in the prior art.

[0007] This manual adopts the following technical solutions:

[0008] This specification provides a parameter calibration method, including:

[0009] Obtaining point cloud data collected by each lidar on the autonomous driving device, as well as the pose based on which the point cloud data was collected, where the pose is determined by other sensors on the autonomous driving device;

[0010] For each laser radar, based on each point cloud data collected by the laser radar and the posture based on which each point cloud data is collected by the laser radar, the external parameters corresponding to the laser radar are optimized to obtain the optimized external parameters;

[0011] Multiple rounds of parameter correction are performed on each laser radar on the autonomous driving device to obtain the corrected external parameters corresponding to each laser radar, wherein for each round of parameter correction, the optimized external parameters of the second laser radar are corrected according to the optimized external parameters of the first laser radar in this round of parameter correction, the point cloud data collected by the first laser radar, and the point cloud data collected by the second laser radar that needs to be optimized in this round of parameter correction to obtain the corrected external parameters, and it is determined whether the preset correction termination conditions are met. If it is determined that the correction termination conditions are not met, the second laser radar is used as the first laser radar for the next round of parameter correction, and the corrected external parameters are used as the optimized external parameters of the first laser radar in the next round of parameter correction.

[0012] Optionally, based on each point cloud data collected by the laser radar and the posture based on which each point cloud data is collected by the laser radar, the external parameters corresponding to the laser radar are optimized to obtain optimized external parameters, including:

[0013] Determine a target area from an area covered by each point cloud data collected by the laser radar according to each point cloud data collected by the laser radar;

[0014] According to the point cloud data corresponding to the target area and the posture based on which the point cloud data corresponding to the target area is collected by the laser radar, the external parameters corresponding to the laser radar are optimized to obtain optimized external parameters.

[0015] Optionally, the other sensors include: a global navigation satellite system GNSS;

[0016] According to each point cloud data collected by the laser radar, a target area is determined from an area covered by each point cloud data collected by the laser radar, including:

[0017] Obtain the GNSS signal distribution map corresponding to each point cloud data collected by the lidar;

[0018] According to the GNSS signal distribution map, the target area is determined from the area covered by each point cloud data collected by the lidar.

[0019] Optionally, determining the target area from the area covered by each point cloud data collected by the lidar according to the GNSS signal distribution map includes:

[0020] The point cloud data collected by the lidar and the GNSS signal distribution map are input into a preset screening model to determine the score corresponding to each area covered by the point cloud data collected by the lidar through the screening model, and based on the score, determine the target area from the area covered by the point cloud data collected by the lidar.

[0021] Optionally, based on each point cloud data collected by the laser radar and the posture based on which each point cloud data is collected by the laser radar, the external parameters corresponding to the laser radar are optimized to obtain optimized external parameters, including:

[0022] Determining, based on each point cloud data collected by the laser radar, a first relative pose corresponding to a target object involved in each point cloud data collected by the laser radar;

[0023] Determining a second relative pose corresponding to the target object based on the pose based on which the autonomous driving device collects each point cloud data through the laser radar;

[0024] According to the posture deviation between the first relative posture and the second relative posture, the external parameters corresponding to the laser radar are optimized to obtain optimized external parameters.

[0025] Optionally, determining a second relative pose corresponding to the target object according to a pose based on which the autonomous driving device collects each point cloud data through the laser radar includes:

[0026] Determining point cloud data related to the target object from each point cloud data collected by the laser radar as basic point cloud data;

[0027] Determining, based on the point cloud data related to the target object except the basic point cloud data collected by the laser radar, a set of point cloud data associated with the basic point cloud data as a target point cloud data set;

[0028] The second relative posture corresponding to the target object is determined based on the basic point cloud data, the target point cloud data set, and the posture based on which the autonomous driving device collects each point cloud data through the laser radar.

[0029] Optionally, based on the point cloud data related to the target object other than the basic point cloud data collected by the laser radar, a set of point cloud data associated with the basic point cloud data is determined as the target point cloud data set, including:

[0030] Determining target point cloud data from each point cloud data collected by the laser radar except the basic point cloud data;

[0031] Determine, from each point cloud data collected by the laser radar except the basic point cloud data, point cloud data whose collection time corresponding to the target point cloud data does not exceed a set time interval, as associated point cloud data corresponding to the target point cloud data;

[0032] According to the target point cloud data and the associated point cloud data corresponding to the target point cloud data, a set of point cloud data associated with the basic point cloud data is determined as the target point cloud data set.

[0033] Optionally, the method further includes:

[0034] For each laser radar of the autonomous driving device, determining whether a deviation between an external parameter after correction corresponding to the laser radar and an external parameter before correction corresponding to the laser radar exceeds a preset threshold;

[0035] If so, a calibration request is sent to the server to manually recalibrate the external parameters;

[0036] If not, perform the tasks required to control the autonomous driving equipment according to the corrected external parameters corresponding to each of the laser radars.

[0037] Optionally, for each round of parameter correction, the first laser radar in the round of parameter correction and the second laser radar in the round of parameter correction are laser radars arranged in different areas of the autonomous driving device.

[0038] This specification provides a parameter calibration device, including:

[0039] An acquisition module, configured to acquire point cloud data collected by each lidar on the autonomous driving device, and a pose based on which the point cloud data is collected, the pose being determined by other sensors on the autonomous driving device;

[0040] An optimization module is used to optimize the external parameters corresponding to each laser radar based on the point cloud data collected by the laser radar and the posture based on which the laser radar collects the point cloud data, so as to obtain the optimized external parameters;

[0041] A correction module is used to perform multiple rounds of parameter correction on each laser radar on the autonomous driving device to obtain the corrected external parameters corresponding to each laser radar, wherein for each round of parameter correction, the optimized external parameters of the second laser radar are corrected according to the optimized external parameters of the first laser radar in this round of parameter correction, the point cloud data collected by the first laser radar, and the point cloud data collected by the second laser radar that needs to be optimized in this round of parameter correction to obtain the corrected external parameters, and determine whether the preset correction termination conditions are met. If it is determined that the correction termination conditions are not met, the second laser radar is used as the first laser radar for the next round of parameter correction, and the corrected external parameters are used as the optimized external parameters of the first laser radar in the next round of parameter correction.

[0042] This specification provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the above-mentioned parameter correction method is implemented.

[0043] This specification provides an autonomous driving device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements a parameter correction method when executing the program.

[0044] At least one of the above technical solutions adopted in this specification can achieve the following beneficial effects:

[0045] The parameter calibration method provided in this specification first obtains each point cloud data collected by each laser radar on the autonomous driving device, as well as the pose based on which each point cloud data is collected, wherein the pose is determined by other sensors on the autonomous driving device. Then, for each laser radar, based on the each point cloud data collected by the laser radar and the pose based on which each point cloud data is collected by the laser radar, the extrinsic parameters corresponding to the laser radar are optimized to obtain optimized extrinsic parameters. Further, multiple rounds of parameter calibration are performed on each laser radar on the autonomous driving device to obtain corrected extrinsic parameters corresponding to each laser radar. For each round of parameter calibration, based on the optimized extrinsic parameters of the first laser radar in the round of parameter calibration, the each point cloud data collected by the first laser radar, and the each point cloud data collected by the second laser radar to be optimized in the round of parameter calibration, the optimized extrinsic parameters of the second laser radar are calibrated to obtain corrected extrinsic parameters, and it is determined whether a preset calibration termination condition is met. If it is determined that the calibration termination condition is not met, the second laser radar is used as the first laser radar for the next round of parameter calibration, and the corrected extrinsic parameters are used as the optimized extrinsic parameters of the first laser radar in the next round of parameter calibration.

[0046] It can be seen from the above method that by optimizing the external parameters corresponding to each laser radar according to the point cloud data collected by the laser radar and the posture based on which the laser radar collects the point cloud data, and correcting the external parameters of each laser radar according to the optimized external parameters of each laser radar, the external parameters of each laser radar can be corrected in real time, and the problem of lack of consistency between the external parameters of different laser radars is avoided, thereby providing a guarantee for the execution of tasks that require the use of external parameters. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The drawings described herein are used to provide a further understanding of this specification and constitute a part of this specification. The exemplary embodiments and descriptions of this specification are used to explain this specification and do not constitute an improper limitation of this specification. In the drawings:

[0048] Figure 1 A flow chart of a parameter correction method provided in this specification;

[0049] Figure 2 A schematic diagram of the overall process of an external reference calibration method provided in this manual;

[0050] Figure 3 A schematic diagram of a parameter correction device provided in this specification;

[0051] Figure 4 This manual provides a corresponding Figure 1 Schematic diagram of the electronic device. DETAILED DESCRIPTION

[0052] To make the objectives, technical solutions, and advantages of this specification more clear, the following will clearly and completely describe the technical solutions of this specification in conjunction with the specific embodiments of this specification and the corresponding drawings. Obviously, the embodiments described are only part of the embodiments of this specification, not all of the embodiments. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of this specification.

[0053] The technical solutions provided by the embodiments of this specification are described in detail below with reference to the accompanying drawings.

[0054] Figure 1 This is a flow chart of a parameter calibration method provided in this specification, comprising the following steps:

[0055] S101: Acquire each point cloud data collected by each laser radar on the autonomous driving device, and the posture based on which the point cloud data is collected, where the posture is determined by other sensors on the autonomous driving device.

[0056] In this specification, the autonomous driving device can promptly correct the external parameters (parameters used to convert data collected by each sensor between different sensor coordinate systems to the coordinate systems corresponding to other sensors, i.e., external parameters) between each laser radar and other sensor devices in the autonomous driving device based on the looseness of the equipment that occurs during daily operation of the sensor devices installed on the autonomous driving device.

[0057] Prior to this, the autonomous driving device needs to first obtain the point cloud data collected by each laser radar installed on the autonomous driving device, as well as the posture based on which the autonomous driving device collects each point cloud data, determined by other sensors on the autonomous driving device. Then, based on the point cloud data collected by each laser radar on the autonomous driving device, as well as the posture based on which the autonomous driving device collects each point cloud data, the external parameters between the laser radars of the autonomous driving device and other sensors can be calibrated.

[0058] The poses based on which the autonomous driving device collects each point cloud data include: the pose of the lidar relative to the world coordinate system, and the pose of the lidar relative to the coordinate systems corresponding to other sensors. Other sensors collect pose information. The other sensors mentioned here may include: an integrated navigation device composed of an inertial measurement unit (IMU), a global navigation satellite system (GNSS), a camera, millimeter-wave radar, etc.

[0059] In this specification, the execution entity of the method for implementing parameter correction may refer to an autonomous driving device or a designated device such as a server set up on a business platform. For the sake of ease of description, the parameter correction method provided in this specification is described below using the autonomous driving device as an example of the execution entity.

[0060] Among them, when the executor of the method for implementing parameter correction is the above-mentioned designated device, the designated device receives the point cloud data sent by the autonomous driving device, and corrects the external parameters of each lidar through the above-mentioned method to obtain the corrected external parameters, and sends the corrected external parameters to the autonomous driving device to perform the task through the autonomous driving device.

[0061] S102: For each laser radar, according to each point cloud data collected by the laser radar and the posture based on which each point cloud data is collected by the laser radar, the external parameters corresponding to the laser radar are optimized to obtain the optimized external parameters.

[0062] After the autonomous driving device obtains the point cloud data collected by each laser radar installed on the autonomous driving device, as well as the posture based on which the autonomous driving device collects each point cloud data determined by other sensors on the autonomous driving device, the autonomous driving device can optimize the external parameters corresponding to each laser radar according to the point cloud data collected by the laser radar and the posture based on which the laser radar collects each point cloud data, to obtain the optimized external parameters.

[0063] In this specification, the autonomous driving device can determine the target object from the point cloud data collected by the laser radar before optimizing the external parameters corresponding to the laser radar based on the point cloud data collected by the laser radar and the posture based on which the point cloud data is collected by the laser radar. Then, based on the posture of the corresponding target object in different point cloud data, the external parameters of each laser radar of the autonomous driving device can be calibrated, wherein the target object refers to the target object contained in one or more frames of point cloud data in each laser radar, or the matching point cloud points in one or more frames of point cloud data in each laser radar.

[0064] For example, if one or more frames of point cloud data collected by a laser radar installed in an HD map collection vehicle contain point cloud data corresponding to a house, the house is the target object. For another example, if one frame of point cloud data collected by a laser radar installed in an acquisition device contains a specified point cloud point, and another frame or more frames of data collected by the laser radar contain point cloud points that correspond to the specified point cloud point in terms of positional relationship, the corresponding point cloud points in these two or more frames of point cloud data can be considered to be matching point cloud points, wherein the specified point cloud point and the point cloud points that match the specified point cloud point are the target objects.

[0065] Based on this, the autonomous driving device can determine, based on the point cloud data collected by the lidar and using a preset algorithm, a first relative pose corresponding to the target object involved in each point cloud data collected by the lidar. Furthermore, based on the pose based on which the autonomous driving device collected each point cloud data using the lidar, it can determine a second relative pose corresponding to the target object. Then, based on the pose deviation between the first relative pose and the second relative pose, the extrinsic parameters corresponding to the lidar are optimized to obtain optimized extrinsic parameters.

[0066] The preset algorithm may be, for example, a Normal Distributions Transform (NDT) algorithm, an Iterative Closest Point (ICP) algorithm, or the like.

[0067] Specifically, the autonomous driving device can select any frame of point cloud data from the point cloud data containing the target object collected by the laser radar as the basic point cloud data, and then select any frame of point cloud data from the point cloud data containing the target object collected by the laser radar except the basic point cloud data as the target point cloud data, and then optimize the external parameters of the laser radar based on the basic point cloud data and target point cloud data of the laser radar.

[0068] In addition, in order to make the final parameter optimization result more accurate, the autonomous driving device can also select one or more frames of point cloud data corresponding to the target point cloud data from the point cloud data collected by the lidar except the basic point cloud data, and the collection time does not exceed the set time interval as the associated point cloud data corresponding to the target point cloud data. According to the target point cloud data and the associated point cloud data corresponding to the target point cloud data, a target point cloud data set is created.

[0069] For example: from the point cloud data collected by the laser radar, select the i-th frame point cloud data as the basic point cloud data, and then select the j-th frame point cloud data from the point cloud data collected by the laser radar except the basic point cloud data as the target point cloud data, and from the point cloud data collected by the laser radar except the basic point cloud data, select the j-1-th frame point cloud data and the j+1-th frame point cloud data with a collection time interval of 1 second corresponding to the target point cloud data as the associated point cloud data corresponding to the target point cloud data, and create a target point cloud data set based on the target point cloud data and the associated point cloud data corresponding to the target point cloud data.

[0070] Furthermore, the point cloud data of each frame in the target point cloud data set can be fused to obtain fused point cloud data, wherein the fusion process refers to superimposing and fusing the point clouds in each frame of point cloud data to construct a map (that is, the fused point cloud data mentioned above), and then optimizing the external parameters of the lidar according to the relative position of the target object in the basic point cloud data and the target object in the constructed map.

[0071] The autonomous driving device can calculate the first relative pose of the target point cloud data and the target object in the fused point cloud data through the above-mentioned preset algorithm, and set an initial value as an extrinsic parameter residual that may exist between the extrinsic parameters of the laser radar after optimization (the residual is the difference between the extrinsic parameters of the laser radar after optimization). It can be understood that it is first assumed that there is a difference between the external parameters set in the laser radar and the actual external parameters (that is, the optimized external parameters), and then an initial value is set as this difference. In subsequent calculations, this difference is gradually modified to obtain the actual difference between the external parameters set in the laser radar and the actual external parameters.

[0072] Then, the autonomous driving device can convert the corresponding pose of the target object in the basic point cloud data collected by other sensors in the coordinate system of other sensors, as well as the corresponding pose of the target object in the fused point cloud data in the coordinate system of other sensors, into the lidar coordinate system through the external parameters between the lidar and other sensors, and obtain the pose of the target object in the basic point cloud data and the fused point cloud data in the lidar coordinate system, and calculate the relative pose between the poses of the target object in the basic point cloud data and the fused point cloud data in the lidar coordinate system as the second relative pose.

[0073] It is worth noting that the first relative pose is calculated by a preset algorithm based on the point cloud data collected by the lidar, and the point cloud data collected by the lidar is in the lidar coordinate system. Therefore, in the above content, it is necessary to convert the basic point cloud data and the pose of the target object in the fused point cloud data in other sensor coordinate systems into the lidar coordinate system through external parameters to ensure that the first relative pose and the second relative pose of the basic point cloud data and the fused point cloud data are in the same coordinate system, and then subsequent optimization of the external parameters can be carried out.

[0074] In addition, since the external parameters used in the process of converting the poses of the target objects in the basic point cloud data and the fused point cloud data in other sensor coordinate systems to the lidar coordinate system are the initial external parameters before optimization between the lidar and other sensors, if the initial external parameters before optimization are inaccurate, there will be a large deviation between the calculated second relative pose and the first relative pose obtained directly based on the point cloud data collected by the lidar. Based on this, the external parameters of the lidar can be optimized according to the deviation between the first relative pose and the second relative pose.

[0075] Specifically, a preset algorithm can be used to optimize the initial value of the preset extrinsic parameter residual with the goal of minimizing the difference between the first relative pose and the second relative pose of the target object. Then, the optimized extrinsic parameter of the lidar can be obtained based on the extrinsic parameter of the lidar and the extrinsic parameter residual of the lidar. The preset algorithm can be an algorithm such as the least squares method. The following takes the least squares method as the preset algorithm and the combined navigation device as the other sensor as an example to further illustrate the above content, with specific reference to the following formula:

[0076]

[0077] In the above formula, T′ lidar→imu Represents the residual of the external parameter to be optimized, Represents the first relative pose calculated by the NDT algorithm, represents the second relative posture, Represents the pose of the target object in the fused point cloud data in the laser radar coordinate system based on the pose of the target object in the combined navigation device coordinate system. Represents the pose of the target object in the basic point cloud data in the lidar coordinate system based on the pose of the target object in the basic point cloud data in the combined navigation device coordinate system.

[0078] It can be seen from here that the optimized value of the initial extrinsic parameter residual can be calculated by the least squares method, and then the optimized extrinsic parameter corresponding to the laser radar can be calculated based on the optimized value of the extrinsic parameter residual corresponding to the laser radar.

[0079] In this specification, in order to improve the efficiency and accuracy of optimizing the external parameters of each lidar, the autonomous driving equipment can also determine the target area with rich feature information from the area covered by the point cloud data collected by the lidar, and then optimize the external parameters corresponding to the lidar through the above-mentioned lidar external parameter optimization method based on the corresponding point cloud data in the target area to obtain the optimized external parameters.

[0080] For example, in high-precision map drawing tasks, areas with strong GNSS signals, containing more trajectories of acquisition equipment, and rich point cloud data features (point cloud data features can be corner point features, etc.) can be selected from various point cloud data as target areas.

[0081] Specifically, the autonomous driving device can obtain the GNSS signal distribution map corresponding to the point cloud data collected by the lidar, and determine the target area from the area covered by the point cloud data collected by the lidar based on the GNSS signal distribution map.

[0082] Of course, the autonomous driving device can also input the point cloud data and GNSS signal distribution map collected by the lidar into a preset screening model to determine the trajectory characteristics, point cloud data characteristics, and GNSS signal distribution characteristics of the collection device for each area covered by the point cloud data collected by the lidar through the screening model, and then determine the score corresponding to each area covered by the point cloud data collected by the lidar, and determine the target area from the area covered by the point cloud data collected by the lidar according to the score.

[0083] Among them, the training method of the screening model can be to input the point cloud data and GNSS signal distribution map in an area into a preset screening model, obtain the score of the area output by the screening model, and then train the screening model with the optimization goal of minimizing the deviation between the score of the area output by the screening model and the actual score of the area.

[0084] It should be noted that the size of the target area can be determined according to actual needs. For example, the target area is an area with a size of 100 meters × 100 meters.

[0085] S103: Perform multiple rounds of parameter calibration on each laser radar on the autonomous driving device to obtain the corrected external parameters corresponding to each laser radar, wherein for each round of parameter calibration, the optimized external parameters of the second laser radar are calibrated according to the optimized external parameters of the first laser radar in this round of parameter calibration, the point cloud data collected by the first laser radar, and the point cloud data collected by the second laser radar that needs to be optimized in this round of parameter calibration to obtain the corrected external parameters, and determine whether the preset calibration termination conditions are met. If it is determined that the calibration termination conditions are not met, the second laser radar is used as the first laser radar for the next round of parameter calibration, and the corrected external parameters are used as the optimized external parameters of the first laser radar in the next round of parameter calibration.

[0086] Because in the above content, the process of optimizing the external parameters of each lidar of the autonomous driving equipment is only performed on the lidar. Therefore, the optimized external parameters corresponding to the lidar lack consistency with other lidars during actual application, which leads to the optimized external parameters not being well applicable to subsequent tasks. Therefore, after optimizing the external parameters corresponding to each lidar, the autonomous driving equipment can also perform multiple rounds of parameter correction on each lidar to obtain the external parameters corresponding to multiple rounds of parameter correction for each lidar, so that the data collected by each lidar at the same time after the external parameters are corrected are consistent.

[0087] Among them, the autonomous driving equipment can determine the first laser radar in each round of parameter correction, and use the first laser radar as the basic laser radar in the round of parameter correction to perform parameter correction on the second laser radar.

[0088] Specifically, the autonomous driving device can calibrate the optimized external parameters of the second laser radar based on the optimized external parameters of the first laser radar in this round of parameter correction, the point cloud data collected by the first laser radar, and the point cloud data collected by the second laser radar that needs to be optimized in this round of parameter correction, obtain the corrected external parameters, and judge whether the preset correction termination conditions are met. If it is determined that the correction termination conditions are not met, the second laser radar is used as the first laser radar for the next round of parameter correction, and the corrected external parameters are used as the optimized external parameters of the first laser radar in the next round of parameter correction.

[0089] Among them, for each round of parameter correction, the first laser radar in the round of parameter correction and the second laser radar in the round of parameter correction are laser radars set in different areas of the autonomous driving device, for example, a top-view laser radar set on the top of the autonomous driving device, and a side-view laser radar set on the side of the autonomous driving device, that is, laser radars set in different areas of the autonomous driving device.

[0090] In this specification, if the autonomous driving device determines that the number of parameter correction rounds has reached a preset number, it can be determined that the above-mentioned preset condition has been met, and the correction of the extrinsic parameter is stopped. Alternatively, the autonomous driving device can determine that the above-mentioned preset condition has been met, and the correction of the extrinsic parameter is stopped, when it determines that the difference between the extrinsic parameter after correction of the second laser radar in the previous round of parameter correction and the extrinsic parameter before correction of the second laser radar in the current round of parameter correction is less than a preset threshold. Other forms of preset conditions are not illustrated here one by one.

[0091] When the parameter correction meets the preset conditions, the optimized extrinsic parameters of the first laser radar and the corrected extrinsic parameters of the second laser radar in this round of parameter correction are used as the corrected extrinsic parameters corresponding to the first laser radar and the second laser radar. The following takes the preset algorithm as the least squares method and the other sensors as the combined navigation device as an example to further illustrate the above content. Please refer to the following formula for details.

[0092]

[0093] In the above formula, Represents the residual error of the external parameters to be optimized for the second lidar, represents the first relative posture, represents the second relative posture, Represents the pose of the target object in the first lidar coordinate system obtained based on the pose of the target object in the basic point cloud data in the coordinate system corresponding to the integrated navigation device. Represents the pose of the target object in the second lidar coordinate system obtained based on the pose of the target object in the fused point cloud data in the coordinate system corresponding to the integrated navigation device. Represents the extrinsic parameter residual of the first lidar.

[0094] It can be seen from the above formula that the difference between the extrinsic parameters before and after optimization of the fixed first laser radar can remain unchanged, and through algorithms such as the least squares method, with the goal of minimizing the posture deviation between the first relative posture and the second relative posture of the target object, the initial value of the extrinsic parameter residual value corresponding to the second laser radar is optimized, and then the corrected extrinsic parameter corresponding to the second laser radar can be calculated based on the extrinsic parameter residual value corresponding to the optimized second laser radar.

[0095] From the above content, it can be seen that before the autonomous driving equipment needs to process the point cloud data collected by the lidar according to the lidar's external parameters to perform subsequent tasks, it can first calibrate the lidar's external parameters. However, the external parameters corrected by the above method are only the corrected external parameters calculated by the algorithm, and there is no guarantee that the corrected external parameters are the actual external parameters. That is, through the above method, the corrected external parameters can be made as close to the actual external parameters as possible.

[0096] Of course, in this specification, the autonomous driving device can also determine the use of the corrected external parameters in subsequent tasks by comparing the differences between the corrected external parameters and the external parameters before correction. Specifically, before the autonomous driving device obtains the corrected external parameters corresponding to each laser radar through the above method and actually applies them to the subsequent task, it can be determined whether the deviation between the corrected external parameters corresponding to the laser radar and the external parameters before correction corresponding to the laser radar exceeds a preset threshold. If so, the external parameters corrected by the above method can no longer guarantee the accuracy of the external parameters, which may affect the subsequent tasks. Therefore, it is necessary to send a calibration request to the server to manually recalibrate the external parameters. If not, the corrected external parameters can be used directly in the subsequent task, and then the task can be executed according to the corrected external parameters corresponding to each laser radar.

[0097] The method provided in this specification can be used in many scenarios. For example, in the process of drawing high-precision maps, since the manual recalibration of the preset external parameters is performed independently for each lidar, it will lead to a lack of consistency between the external parameters of each radar sensor, and thus cause the drawn high-precision map to have ghosting problems. Therefore, the autonomous driving device can correct the external parameters of each lidar set on the autonomous driving device according to the point cloud data collected by each lidar on each autonomous driving device and the corresponding posture of each point cloud data, and then draw the corresponding high-precision map based on the corrected external parameters.

[0098] For another example, during trajectory planning for an autonomous driving device, if there is a deviation between the external parameters of the lidar and the actual external parameters of the lidar, and if this deviation is not promptly corrected, the accuracy of the trajectory planned for the autonomous driving device may be affected, thereby posing a safety hazard to the operation of the autonomous driving device. Therefore, before planning a trajectory for the autonomous driving device based on the external parameters of the lidar, the autonomous driving device can calibrate the external parameters of each lidar installed on the autonomous driving device using the methods provided in this specification, and then plan a corresponding trajectory for the autonomous driving device based on the corrected external parameters.

[0099] Figure 2 This is a schematic diagram of the overall process of an external reference calibration method provided in this manual.

[0100] like Figure 2 As shown, the autonomous driving device can score each area covered by each point cloud data collected by each laser radar set on the autonomous driving device based on the acquired point cloud data collected by each laser radar, the posture based on which the autonomous driving device collects each point cloud data, the external parameters of each laser radar, and the GNSS signal distribution map, and determine a target area based on the score, and then optimize the external parameters of each laser radar based on the point cloud data in the target area collected by each laser radar to obtain the optimized external parameters.

[0101] Furthermore, the first laser radar and the second laser radar can be selected in turn from each laser radar. Then, in order to ensure that the optimized external parameters of the first laser radar and the second laser radar are more consistent, the optimized external parameters of the first laser radar and the second laser radar can be extrinsically corrected to obtain the corrected external parameters of each laser radar.

[0102] If the difference between the extrinsic parameters of each LiDAR after calibration and the extrinsic parameters before calibration does not exceed a preset threshold, the extrinsic parameters of each LiDAR may be output for use in subsequent tasks required to control the autonomous driving device. If the difference between the extrinsic parameters of each LiDAR after calibration and the extrinsic parameters before calibration does not exceed a preset threshold, a request for manual recalibration of the extrinsic parameters is generated and sent to the R&D personnel.

[0103] The above is a parameter correction method provided in one or more embodiments of this specification. Based on the same idea, this specification also provides a corresponding parameter correction device, such as Figure 3 shown.

[0104] Figure 3 A schematic diagram of a parameter correction device provided in this specification, including:

[0105] An acquisition module 301 is configured to acquire point cloud data collected by each lidar on the autonomous driving device, as well as a pose based on which the point cloud data is collected, the pose being determined by other sensors on the autonomous driving device.

[0106] The optimization module 302 is configured to optimize the extrinsic parameters corresponding to each laser radar based on the point cloud data collected by the laser radar and the pose based on which the laser radar collects the point cloud data, thereby obtaining optimized extrinsic parameters.

[0107] The correction module 303 is used to perform multiple rounds of parameter correction on each laser radar on the autonomous driving device to obtain the corrected external parameters corresponding to each laser radar, wherein, for each round of parameter correction, the optimized external parameters of the second laser radar are corrected according to the optimized external parameters of the first laser radar in this round of parameter correction, the point cloud data collected by the first laser radar, and the point cloud data collected by the second laser radar that needs to be optimized in this round of parameter correction to obtain the corrected external parameters, and determine whether the preset correction termination conditions are met. If it is determined that the correction termination conditions are not met, the second laser radar is used as the first laser radar for the next round of parameter correction, and the corrected external parameters are used as the optimized external parameters of the first laser radar in the next round of parameter correction.

[0108] Optionally, the optimization module 302 is specifically used to determine the target area from the area covered by the point cloud data collected by the laser radar based on the point cloud data collected by the laser radar; optimize the external parameters corresponding to the laser radar based on the point cloud data corresponding to the target area and the posture based on which the point cloud data corresponding to the target area is collected by the laser radar to obtain optimized external parameters.

[0109] Optionally, the optimization module 302 is specifically used to obtain a GNSS signal distribution map corresponding to each point cloud data collected by the laser radar; and determine a target area from an area covered by each point cloud data collected by the laser radar according to the GNSS signal distribution map.

[0110] Optionally, the optimization module 302 is specifically used to input the point cloud data collected by the lidar and the GNSS signal distribution map into a preset screening model, so as to determine the score corresponding to each area covered by the point cloud data collected by the lidar through the screening model, and determine the target area from the area covered by the point cloud data collected by the lidar based on the score.

[0111] Optionally, the optimization module 302 is specifically used to determine, based on the point cloud data collected by the laser radar, a first relative posture corresponding to the target object involved in each point cloud data collected by the laser radar; determine a second relative posture corresponding to the target object based on the posture based on which the autonomous driving device collects each point cloud data through the laser radar; and optimize the external parameters corresponding to the laser radar based on the posture deviation between the first relative posture and the second relative posture to obtain optimized external parameters.

[0112] Optionally, the optimization module 302 is specifically used to determine target point cloud data from each point cloud data collected by the laser radar except the basic point cloud data; determine point cloud data corresponding to the target point cloud data whose collection time does not exceed a set time interval from each point cloud data collected by the laser radar except the basic point cloud data, as associated point cloud data corresponding to the target point cloud data; and determine a set of point cloud data associated with the basic point cloud data as a target point cloud data set based on the target point cloud data and the associated point cloud data corresponding to the target point cloud data.

[0113] Optionally, the optimization module 302 is specifically used to determine target point cloud data from each point cloud data collected by the laser radar except the basic point cloud data; determine point cloud data corresponding to the target point cloud data whose collection time does not exceed a set time interval from each point cloud data collected by the laser radar except the basic point cloud data, as associated point cloud data corresponding to the target point cloud data; and determine a set of point cloud data associated with the basic point cloud data as a target point cloud data set based on the target point cloud data and the associated point cloud data corresponding to the target point cloud data.

[0114] Optionally, the determination module 304 is specifically used to determine, for each laser radar of the autonomous driving device, whether the deviation between the corrected external parameters corresponding to the laser radar and the pre-corrected external parameters corresponding to the laser radar exceeds a preset threshold; if so, send a calibration request to the server to manually recalibrate the external parameters; if not, execute the tasks required to control the autonomous driving device according to the corrected external parameters corresponding to each laser radar.

[0115] Optionally, for each round of parameter correction, the first laser radar in the round of parameter correction and the second laser radar in the round of parameter correction are laser radars arranged in different areas of the autonomous driving device.

[0116] This specification also provides a computer-readable storage medium, which stores a computer program that can be used to execute the above Figure 1 A parameter correction method is provided.

[0117] This manual also provides Figure 4 The one shown corresponds to Figure 1 Schematic diagram of the electronic equipment. Figure 4 As mentioned above, at the hardware level, the electronic device includes a processor, an internal bus, a network interface, a memory and a non-volatile memory, and may also include other hardware required for the business. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to achieve the above Figure 1 Of course, in addition to software implementation, this specification does not exclude other implementation methods, such as logic devices or a combination of software and hardware, etc., that is, the execution subject of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0118] In the 1990s, technological improvements could be clearly distinguished as either hardware improvements (for example, improvements to circuit structures like diodes, transistors, and switches) or software improvements (improvements to process flows). However, with the advancement of technology, many process flow improvements today can now be considered direct improvements to hardware circuit structures. Designers almost always create the corresponding hardware circuit structure by programming the improved process flow into the hardware circuit. Therefore, it cannot be said that a process flow improvement cannot be implemented using hardware modules. For example, a programmable logic device (PLD), such as a field programmable gate array (FPGA), is an integrated circuit whose logical function is determined by user programming. Designers can "integrate" a digital system on a PLD through their own programming, without having to hire a chip manufacturer to design and manufacture a dedicated integrated circuit chip. Moreover, nowadays, instead of manually fabricating integrated circuit chips, this programming is mostly done using "logic compiler" software. This is similar to the software compiler used when developing programs. Before compilation, the original code must also be written in a specific programming language, called a hardware description language (HDL). There is not just one HDL, but many, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc. The most commonly used ones are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art will also understand that by simply programming the method flow in one of these hardware description languages ​​and then programming it into an integrated circuit, a hardware circuit that implements the logic method flow can be easily obtained.

[0119] The controller can be implemented in any suitable manner. For example, the controller can take the form of a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers. Examples of controllers include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320. The memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art will also know that in addition to implementing the controller in a purely computer-readable program code format, the controller can be implemented in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, such a controller can be considered a hardware component, and the devices included therein for implementing various functions can also be considered as structures within the hardware component. Or even, the devices for implementing various functions can be considered as both software modules that implement the method and structures within the hardware component.

[0120] The systems, devices, modules, or units described in the above embodiments may be implemented by computer chips or entities, or by products having certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smartphone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0121] For the convenience of description, the above devices are described as being divided into various units according to their functions. Of course, when implementing this specification, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0122] Those skilled in the art will appreciate that the embodiments of this specification may be provided as methods, systems, or computer program products. Therefore, this specification may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, this specification may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0123] This specification is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of this specification. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0124] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0125] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0126] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0127] Memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

[0128] Computer-readable media includes permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0129] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0130] Those skilled in the art will appreciate that the embodiments of this specification may be provided as methods, systems, or computer program products. Thus, this specification may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, this specification may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0131] This specification may be described in the general context of computer-executable instructions, such as program modules, executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, and the like that perform specific tasks or implement specific abstract data types. This specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media, including storage devices.

[0132] The various embodiments in this specification are described in a progressive manner. Similar parts between the various embodiments can be referred to in conjunction with each other. Each embodiment focuses on the differences between the other embodiments. In particular, the system embodiments are generally similar to the method embodiments, so the description is relatively simple. For relevant parts, refer to the description of the method embodiments.

[0133] The foregoing is merely an example of the present invention and is not intended to limit the present invention. Various modifications and variations are possible for those skilled in the art. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be included within the scope of the claims of the present invention.

Claims

1. A parameter calibration method, characterized in that: include: Obtaining point cloud data collected by each lidar on the autonomous driving device, as well as a pose based on which the point cloud data was collected, the pose being determined by other sensors on the autonomous driving device; For each laser radar, based on each point cloud data collected by the laser radar, determine a first relative pose corresponding to a target object involved in each point cloud data collected by the laser radar; based on the pose based on which the autonomous driving device collects each point cloud data through the laser radar, determine a second relative pose corresponding to the target object; based on the pose deviation between the first relative pose and the second relative pose, optimize the extrinsic parameter corresponding to the laser radar to obtain an optimized extrinsic parameter; the relative pose is the relative pose of the target object contained in the basic point cloud data in the each point cloud data and the target object contained in the fused point cloud data in the each point cloud data; Multiple rounds of parameter correction are performed on each laser radar on the autonomous driving device to obtain the corrected external parameters corresponding to each laser radar, wherein for each round of parameter correction, the optimized external parameters of the second laser radar are corrected according to the optimized external parameters of the first laser radar in this round of parameter correction, the point cloud data collected by the first laser radar, and the point cloud data collected by the second laser radar that needs to be optimized in this round of parameter correction to obtain the corrected external parameters, and it is determined whether the preset correction termination conditions are met. If it is determined that the correction termination conditions are not met, the second laser radar is used as the first laser radar for the next round of parameter correction, and the corrected external parameters are used as the optimized external parameters of the first laser radar in the next round of parameter correction.

2. The method according to claim 1, wherein According to each point cloud data collected by the laser radar and the posture based on which each point cloud data is collected by the laser radar, the external parameters corresponding to the laser radar are optimized to obtain the optimized external parameters, including: Determine a target area from an area covered by each point cloud data collected by the laser radar according to each point cloud data collected by the laser radar; According to the point cloud data corresponding to the target area and the posture based on which the point cloud data corresponding to the target area is collected by the laser radar, the external parameters corresponding to the laser radar are optimized to obtain optimized external parameters.

3. The method according to claim 2, wherein The other sensors include: Global Navigation Satellite System GNSS; According to each point cloud data collected by the laser radar, a target area is determined from an area covered by each point cloud data collected by the laser radar, including: Obtain the GNSS signal distribution map corresponding to each point cloud data collected by the lidar; According to the GNSS signal distribution map, the target area is determined from the area covered by each point cloud data collected by the lidar.

4. The method according to claim 3, wherein Determining a target area from an area covered by each point cloud data collected by the lidar according to the GNSS signal distribution map includes: Inputting each point cloud data collected by the lidar and the GNSS signal distribution map into a preset screening model to determine, through the screening model, a score corresponding to each area covered by each point cloud data collected by the lidar; According to the score, a target area is determined from the area covered by each point cloud data collected by the lidar.

5. The method according to claim 1, wherein Determining a second relative pose corresponding to the target object based on the pose based on which the autonomous driving device collects each point cloud data through the laser radar includes: Determining point cloud data related to the target object from each point cloud data collected by the laser radar as basic point cloud data; Determining, based on the point cloud data related to the target object except the basic point cloud data collected by the laser radar, a set of point cloud data associated with the basic point cloud data as a target point cloud data set; The second relative posture corresponding to the target object is determined based on the basic point cloud data, the target point cloud data set, and the posture based on which the autonomous driving device collects each point cloud data through the laser radar.

6. The method according to claim 5, wherein According to each point cloud data related to the target object except the basic point cloud data collected by the laser radar, a set of point cloud data associated with the basic point cloud data is determined as a target point cloud data set, including: Determining target point cloud data from each point cloud data collected by the laser radar except the basic point cloud data; Determine, from each point cloud data collected by the laser radar except the basic point cloud data, point cloud data whose collection time corresponding to the target point cloud data does not exceed a set time interval, as associated point cloud data corresponding to the target point cloud data; According to the target point cloud data and the associated point cloud data corresponding to the target point cloud data, a set of point cloud data associated with the basic point cloud data is determined as the target point cloud data set.

7. The method according to claim 1, wherein The method further comprises: For each laser radar of the autonomous driving device, determining whether a deviation between an external parameter after correction corresponding to the laser radar and an external parameter before correction corresponding to the laser radar exceeds a preset threshold; If so, a calibration request is sent to the server to manually recalibrate the external parameters; If not, perform the tasks required to control the autonomous driving equipment according to the corrected external parameters corresponding to each of the laser radars.

8. The method according to any one of claims 1 to 7, wherein For each round of parameter correction, the first laser radar in the round of parameter correction and the second laser radar in the round of parameter correction are laser radars arranged in different areas of the autonomous driving device.

9. A parameter correction device, characterized in that: include: an acquisition module, configured to acquire point cloud data collected by each lidar on the autonomous driving device, and a pose based on which the point cloud data is collected, the pose being determined by other sensors on the autonomous driving device; an optimization module for determining, for each laser radar, a first relative pose corresponding to a target object involved in each point cloud data collected by the laser radar based on each point cloud data collected by the laser radar; determining a second relative pose corresponding to the target object based on the pose based on which the autonomous driving device collects each point cloud data through the laser radar; optimizing an extrinsic parameter corresponding to the laser radar based on a pose deviation between the first relative pose and the second relative pose to obtain an optimized extrinsic parameter; the relative pose being the relative pose of the target object contained in the basic point cloud data in each point cloud data and the target object contained in the fused point cloud data in each point cloud data; A correction module is used to perform multiple rounds of parameter correction on each laser radar on the autonomous driving device to obtain the corrected external parameters corresponding to each laser radar, wherein for each round of parameter correction, the optimized external parameters of the second laser radar are corrected according to the optimized external parameters of the first laser radar in this round of parameter correction, the point cloud data collected by the first laser radar, and the point cloud data collected by the second laser radar that needs to be optimized in this round of parameter correction to obtain the corrected external parameters, and determine whether the preset correction termination conditions are met. If it is determined that the correction termination conditions are not met, the second laser radar is used as the first laser radar for the next round of parameter correction, and the corrected external parameters are used as the optimized external parameters of the first laser radar in the next round of parameter correction.

10. A computer-readable storage medium, characterized in that The storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

11. An autonomous driving device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the method according to any one of claims 1 to 8 is implemented.

Images