Parameter calibration method, device and equipment for laser radar and IMU (Inertial Measurement Unit)
By acquiring the candidate frame point cloud data of lidar loop detection and the corrected time measurement data of IMU, the time synchronization error parameters of lidar and IMU are calculated, which solves the problem of inaccurate dynamic calibration accuracy and improves the accuracy of laser inertia odometer and stability.
Patent Information
- Application Number
- CN202411388129.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-30
- Publication Date
- 2024-12-03
AI Technical Summary
In the prior art, the dynamic calibration method between lidar and IMU is inaccurate, especially the time synchronization error cannot be accurately calibrated, which affects the accuracy and stability of the laser inertia odometer.
By acquiring the candidate frame point cloud data detected by lidar loopback, the rotation angular velocity between the candidate frame point cloud data and the target point cloud data is calculated, and the rotation angular velocity measured at the corrected time of the IMU is obtained, the time between the lidar and the IMU is obtained based on the rotation angular velocity measured at the corrected time of the IMU is obtained. Synchronize error parameters and optimize calibration process to improve accuracy.
Accurate external parameter calibration and time synchronization error calibration between lidar and IMU are realized, improving the accuracy of dynamic calibration and reducing calibration costs.
Smart Images

Figure CN119063766A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of autonomous driving technology, and in particular to a method, device and equipment for calibrating parameters of a laser radar and an IMU. Background Art
[0002] In the field of autonomous driving, especially in the research and application of simultaneous mapping and positioning, LiDAR and Inertial Measurement Unit (IMU), as two important sensors that can effectively perceive the motion and spatial posture information of the carrier, are widely used in engineering applications and research in related fields.
[0003] In actual engineering applications of LiDAR, when using LiDAR for positioning to obtain real-time position information of the vehicle, it is often necessary to configure and use it in combination with IMU. The gravity acceleration and angular velocity information in the IMU coordinate system can be measured in its own coordinate system, which can be used to dedistort the LiDAR point cloud data. At the same time, the combination of LiDAR can run the laser inertial mileage calculation method to improve the stability and accuracy of the positioning process of a single LiDAR. In applications in this field, it is necessary to provide a spatial position relationship between the LiDAR sensor coordinate system and the IMU coordinate system, which is generally referred to as the external parameter between the LiDAR and the IMU. At the same time, it is also necessary to calibrate the time synchronization error between the LiDAR measurement data and the IMU measurement data. Accurate evaluation of external parameters and time synchronization errors has an important impact on the accuracy and stability of the laser inertial odometer.
[0004] In the related art, the calibration methods of LiDAR and IMU include static calibration and dynamic calibration. The static calibration method usually requires a set of calibration scenes to be arranged in advance. Each time the parameters of LiDAR and IMU are calibrated, the vehicle needs to be driven to the calibration workshop, which is not only inconvenient to operate, but also has high scene arrangement costs. In addition, the static calibration method cannot calibrate the time synchronization error between IMU and LiDAR. The dynamic calibration method is based on the vehicle running a specific motion trajectory, collecting LiDAR point cloud data and IMU data in the process, setting the initial external parameter conditions, and combining Kalman filtering or nonlinear optimization methods to complete the calibration results of the external parameters and time synchronization errors of LiDAR and IMU. Although the dynamic calibration method does not rely on the static scenes arranged in advance, the cumulative error of the running laser inertial odometer itself will affect the estimation accuracy of the dynamic calibration. Summary of the invention
[0005] In view of this, the present invention provides a parameter calibration method, device and equipment for laser radar and IMU to solve the problem of inaccurate precision in the dynamic calibration method of laser radar and IMU.
[0006] According to a first aspect, an embodiment of the present disclosure provides a parameter calibration method for a laser radar and an IMU, the method comprising:
[0007] Based on the pose transformation result corresponding to the target point cloud data set sampled by the laser radar within the first preset time, the candidate frame point cloud data for the laser radar loop closure detection is obtained;
[0008] Obtaining point cloud pose mapping data between the candidate frame point cloud data and each frame of target point cloud data in the target point cloud data set;
[0009] Based on the point cloud pose mapping data, obtaining a first rotation angular velocity between the candidate frame point cloud data and each frame target point cloud data;
[0010] Obtaining a second rotation angular velocity measured by the IMU at each correction moment within a second preset time, wherein the correction moment within the second preset time is obtained based on the target time deviation adjustment;
[0011] Based on the first rotation angular velocity between the candidate frame point cloud data and each frame target point cloud data and the second rotation angular velocity measured by the IMU at each correction moment within the second preset time, the time synchronization error parameter between the laser radar and the IMU is obtained.
[0012] The disclosed embodiment obtains candidate frame point cloud data of lidar loop detection, further calculates the first rotation angular velocity between the candidate frame point cloud data and each frame of target point cloud data, and obtains the second rotation angular velocity measured by the IMU at each correction moment within a second preset time, and finally obtains the time synchronization error parameter between the lidar and the IMU. It can not only accurately obtain the external parameters of the lidar and the IMU, but also accurately calibrate the time synchronization error between the IMU and the lidar, and optimize the calibration cost.
[0013] In some optional implementations, obtaining a pose transformation result corresponding to a target point cloud data set sampled by a laser radar within a first preset time includes:
[0014] Through the ICP registration algorithm, each time the point cloud data of the current frame and its adjacent previous frame are used as the registration object, and each point cloud data in the point cloud data set collected by the laser radar within the first preset time is successively registered;
[0015] The pose transformation result corresponding to the target point cloud data set sampled by the laser radar within the first preset time is obtained from each registration result.
[0016] The disclosed embodiment implements the above-mentioned implementation mode, and can ensure the accuracy of the target point cloud data set sampled by the laser radar through the ICP registration algorithm.
[0017] In some optional implementations, obtaining candidate frame point cloud data for lidar loop closure detection includes:
[0018] From the pose transformation results corresponding to the target point cloud data set sampled by the laser radar within the first preset time, searching for the first pose transformation data corresponding to the current frame point cloud data;
[0019] Searching for the remaining frame point cloud data sets excluding the current frame point cloud data sets from the target point cloud data sets sampled by the laser radar within the first preset time;
[0020] From the pose transformation result corresponding to the target point cloud dataset sampled by the laser radar within the first preset time, obtain the pose transformation dataset corresponding to the remaining frame point cloud dataset;
[0021] Get each moment corresponding to the remaining frame point cloud data set and the current moment corresponding to the current frame point cloud data;
[0022] Calculate the time difference between each moment corresponding to the remaining frame point cloud data set and the current moment corresponding to the current frame point cloud data;
[0023] From the pose transformation data set corresponding to the remaining frame point cloud data set, obtain the second pose transformation data corresponding to each frame point cloud data of the remaining frame point cloud data set;
[0024] Calculate the distance difference between the second pose transformation data corresponding to each frame of point cloud data of the remaining frame point cloud data set and the first pose transformation data corresponding to the current frame point cloud data;
[0025] The point cloud data corresponding to a time difference greater than a first preset threshold and a distance difference less than a second preset threshold is taken as a candidate point cloud data set;
[0026] Through the ICP registration algorithm, the current frame point cloud data and each frame point cloud data in the candidate point cloud data set are used as the registration object each time, and each point cloud data in each frame point cloud data in the candidate point cloud data set is registered one by one;
[0027] From all the registration results of the candidate point cloud data set, the point cloud data with the largest number of successful registrations is obtained as the candidate frame point cloud data.
[0028] The disclosed embodiment obtains candidate frame point cloud data of lidar loop detection by executing the above-mentioned implementation mode, which is equivalent to calculating the candidate frame point cloud data through loop detection, and further utilizing the candidate frame point cloud data to optimize the error of parameter calibration of lidar and IMU and improve the estimation accuracy of dynamic calibration parameters.
[0029] In some optional implementations, obtaining point cloud pose mapping data includes:
[0030] Obtain candidate time and third pose transformation data corresponding to the candidate frame point cloud data;
[0031] Obtain the fourth pose transformation data and target time corresponding to each frame of target point cloud data;
[0032] Point cloud pose mapping data is generated based on the candidate time, the third pose transformation data, the fourth pose transformation data corresponding to each frame of target point cloud data, and the target time corresponding to each frame of target point cloud data.
[0033] By executing the above implementation, the embodiment of the present disclosure can calculate the first rotational angular velocity between the candidate frame point cloud data and each frame target point cloud data.
[0034] In some optional implementations, obtaining a second rotation angular velocity measured by the IMU at each correction moment within a second preset time, wherein the correction moment within the second preset time is obtained based on the target time deviation adjustment, includes:
[0035] Set the preset time error adjustment range and preset time adjustment step;
[0036] Starting from the left end point value of the preset time error adjustment range, the target time deviation is gradually increased according to the preset time adjustment step length until it increases to the right end point value of the preset time error adjustment range;
[0037] Subtract the target time deviation from the original time corresponding to each IMU measurement data to obtain the corrected time corresponding to each IMU measurement data;
[0038] At a correction moment corresponding to each IMU measurement data, a second rotation angular velocity measured by the IMU at each correction moment within a second preset time is obtained.
[0039] The disclosed embodiment executes the above-mentioned implementation manner and gradually optimizes the measurement moment of each IMU measurement data, thereby facilitating improving the accuracy of measuring the second rotation angular velocity measured by the IMU at each correction moment within the second preset time.
[0040] In some optional embodiments, based on the first rotation angular velocity between the candidate frame point cloud data and each frame target point cloud data and the second rotation angular velocity measured by the IMU at each correction moment within the second preset time, obtaining the time synchronization error parameter between the laser radar and the IMU includes:
[0041] Finding a first rotation angular velocity between the candidate frame point cloud data and each frame target point cloud data at each correction moment within a second preset time based on the IMU;
[0042] Obtain the target time deviation corresponding to the maximum autocorrelation function value between the first rotation angular velocity between the candidate frame point cloud data and each frame target point cloud data and the second rotation angular velocity measured by the IMU at each correction moment within the second preset time through the following autocorrelation function;
[0043]
[0044] Among them, D Z is the autocorrelation function value, d m is the target time deviation, ω1 is the first rotation angular velocity between the candidate frame point cloud data and each frame target point cloud data, and ω2 is the second rotation angular velocity measured by the IMU at each correction moment within the second preset time;
[0045] The target time deviation corresponding to the autocorrelation is taken as the time synchronization error parameter.
[0046] The disclosed embodiment executes the above-mentioned implementation mode, and autocorrelates the first rotation angular velocity between the candidate frame point cloud data and each frame target point cloud data with the second rotation velocity of the IMU at each correction moment within the second preset time, so as to accurately calibrate the time synchronization error parameters between the lidar and the IMU.
[0047] In some optional implementations, the parameter calibration method of the laser radar and the IMU in the embodiment of the present disclosure further includes:
[0048] Based on the time synchronization error parameters between the lidar and the IMU, correct each current moment of the IMU;
[0049] Based on the corrected IMU, obtaining a third rotation angular velocity measured by the IMU at each current moment within a third preset time;
[0050] Based on the first rotation angular velocity between the candidate frame point cloud data and each frame of target point cloud data, and the third rotation angular velocity measured by the IMU at each current moment within a third preset time, the rotation extrinsic parameter data of the IMU is obtained by constructing a rotation residual equation.
[0051] By executing the above-mentioned implementation methods, the embodiment of the present disclosure corrects each current moment of the IMU based on the time synchronization error parameters between the lidar and the IMU, and can accurately calibrate the rotation external parameter data of the IMU.
[0052] According to the second aspect, the embodiment of the present disclosure further provides a parameter calibration device for a laser radar and an IMU, the device comprising:
[0053] A candidate frame data acquisition module, used to acquire candidate frame point cloud data for lidar loop detection based on a pose transformation result corresponding to a target point cloud data set sampled by the lidar within a first preset time;
[0054] A point cloud pose data acquisition module is used to acquire point cloud pose mapping data between candidate frame point cloud data and each frame of target point cloud data in a target point cloud data set;
[0055] A first angular velocity acquisition module, used to acquire a first rotational angular velocity between the candidate frame point cloud data and each frame target point cloud data based on the point cloud posture mapping data;
[0056] A second angular velocity acquisition module is used to acquire a second rotation angular velocity measured by the IMU at each correction moment within a second preset time, where the correction moment within the second preset time is obtained based on the target time deviation adjustment;
[0057] The time synchronization error acquisition module is used to obtain the time synchronization error parameters between the laser radar and the IMU based on the first rotation angular velocity between the candidate frame point cloud data and each frame target point cloud data and the second rotation angular velocity measured by the IMU at each correction moment within the second preset time.
[0058] In a third aspect, the present invention provides a computer device, comprising: a memory and a processor, the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the laser radar and IMU parameter calibration method of the first aspect or any corresponding embodiment thereof by executing the computer instructions.
[0059] In a fourth aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the laser radar and IMU parameter calibration method of the first aspect or any corresponding embodiment thereof. BRIEF DESCRIPTION OF THE DRAWINGS
[0060] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0061] Figure 1 is a flow chart of a parameter calibration method of a laser radar and an IMU according to an embodiment of the present invention;
[0062] Figure 2 is a flow chart of another method for calibrating parameters of a laser radar and an IMU according to an embodiment of the present invention;
[0063] Figure 3is a flow chart of another method for calibrating parameters of a laser radar and an IMU according to an embodiment of the present invention;
[0064] Figure 4 is a flow chart of a method for calibrating parameters of a laser radar and an IMU according to an embodiment of the present invention;
[0065] Figure 5 is a flow chart of a method for calibrating parameters of a laser radar and an IMU according to an embodiment of the present invention;
[0066] Figure 6 is a flow chart of a method for calibrating parameters of a laser radar and an IMU according to an embodiment of the present invention;
[0067] Figure 7 is a flow chart of a method for calibrating parameters of a laser radar and an IMU according to an embodiment of the present invention;
[0068] Figure 8 is a structural block diagram of a parameter calibration device for a laser radar and an IMU according to an embodiment of the present invention;
[0069] Fig. 9 It is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION
[0070] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.
[0071] According to an embodiment of the present invention, an embodiment of a parameter calibration method for a laser radar and an IMU is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.
[0072] In this embodiment, a parameter calibration method of a laser radar and an IMU is provided, which can be used in mobile terminals such as mobile phones, tablet computers, etc. Figure 1 is a flow chart of a parameter calibration method of a laser radar and an IMU according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps:
[0073] Step S101, based on the pose transformation result corresponding to the target point cloud data set sampled by the laser radar within the first preset time, obtain the candidate frame point cloud data for the laser radar loop detection.
[0074] Specifically, before executing the parameter calibration method of the laser radar and IMU in this embodiment, fix the laser radar and IMU modules in advance to ensure that the two are fixed during the calibration process and no relative movement occurs. Then, the data acquisition frame rate of the laser radar and IMU modules is as high as possible. It is necessary to ensure that the laser radar data sampling frequency is greater than 10Hz, and the IMU data sampling frequency is greater than 100Hz. Finally, the surrounding environment is required to have a good structured scene during the data acquisition process. For example, there are relatively regular plane scenes such as walls and ground around.
[0075] Exemplarily, the target point cloud data set sampled by the laser radar within the first preset time is the target point cloud data set continuously collected by the laser radar at each timestamp (each moment) within a continuous time (the first preset time). For example, in each moment of time1, time2, time3, time4, and time5, if each time corresponds to 1s, then 5 times are 5s. Each timestamp (each moment) of the laser radar odometer corresponds to a sampling point, that is, corresponds to a frame of target point cloud. Each timestamp (each moment) within a period of time (the first preset time) corresponds to multiple frames of target point cloud data, thereby constituting a target point cloud data set. For example, the target point cloud data sets corresponding to the above 5s are P1, P2, P3, P4, and P5, respectively. For example, the two frames of target point clouds corresponding to the two adjacent moments time1 and time2 are P1 and P2, respectively.
[0076] Furthermore, the pose transformation result corresponding to the target point cloud data set sampled by the laser radar within the first preset time can be represented by T, and the pose transformation result includes: rotation transformation and translation transformation. The pose transformation result here is obtained by the ICP registration algorithm based on the registration of the target point cloud data of two adjacent frames.
[0077] Furthermore, the candidate frame point cloud data of the lidar loop detection is used to optimize the pose transformation result corresponding to the target point cloud data set sampled by the lidar within the first preset time. That is, the cumulative error existing in the running laser inertial odometer itself will affect the estimation accuracy of dynamic calibration, so as to further achieve the purpose of optimizing the pose error.
[0078] Step S102, obtaining point cloud pose mapping data between the candidate frame point cloud data and each frame of target point cloud data in the target point cloud data set.
[0079] Specifically, the candidate frame point cloud data is represented by P nn Indicates that the candidate frame point cloud data Pnn The target point cloud data of each frame in the target point cloud dataset is obtained by the following detailed loop registration calculation. b Indicates that the point cloud pose mapping data between the candidate frame point cloud data and each frame of the target point cloud data in the target point cloud data set includes: the candidate frame point cloud data P nn Corresponding candidate time -nn , candidate frame point cloud data P nn The corresponding third posture transformation data T nn , each frame of point cloud data P in the target point cloud dataset b The corresponding time is time -ii , each frame of point cloud data P in the target point cloud dataset b The corresponding fourth posture transformation data T ii Among them, the third posture transformation data T nn is the candidate frame point cloud data P nn The pose transformation data corresponding to the current timestamp, the fourth pose transformation data T ii is the point cloud data P of each frame in the target point cloud dataset b The pose transformation data corresponding to the current timestamp.
[0080] Step S103, based on the point cloud pose mapping data, obtaining a first rotational angular velocity between the candidate frame point cloud data and each frame target point cloud data.
[0081] Specifically, the candidate frame point cloud data P nn With each frame target point cloud data P b The first rotational angular velocity between can be represented by ω1, that is, the rotational angular velocity corresponding to each frame of target point cloud at each timestamp after the laser radar is calculated through loop closure detection. Since the target point cloud data set is multi-frame point cloud data, the first rotational angular velocity ω1 between the candidate frame point cloud data and each frame of target point cloud data is also multiple.
[0082] Step S104, obtaining a second rotation angular velocity measured by the IMU at each correction moment within a second preset time, wherein the correction moment within the second preset time is obtained based on the target time deviation adjustment.
[0083] Specifically, each correction moment is each timestamp corrected by the target time deviation, the second preset time is a continuous period of time after the IMU measures the data, and the second rotational angular velocity is the second rotational angular velocity corresponding to the timestamp after the IMU corrects the time deviation. The second rotational angular velocity can be represented by ω2.
[0084] Step S105, based on the first rotation angular velocity between the candidate frame point cloud data and each frame target point cloud data and the second rotation angular velocity measured by the IMU at each correction moment within the second preset time, obtain the time synchronization error parameter between the laser radar and the IMU.
[0085] Specifically, based on the first rotation angular velocity ω1 between the candidate frame point cloud data and each frame of target point cloud data, and the second rotation angular velocity ω2 measured by the IMU at each correction moment within the second preset time, the target time deviation corresponding to the time when the autocorrelation function value between the first rotation angular velocity ω1 between the candidate frame point cloud data and each frame of target point cloud data and the second rotation angular velocity ω2 measured by the IMU at each correction moment within the second preset time is maximum is obtained through the autocorrelation function formula, and the target time deviation is used as the time synchronization error parameter.
[0086] The disclosed embodiment obtains candidate frame point cloud data of lidar loop detection, further calculates the first rotation angular velocity between the candidate frame point cloud data and each frame of target point cloud data, and obtains the second rotation angular velocity measured by the IMU at each correction moment within a second preset time, and finally obtains the time synchronization error parameter between the lidar and the IMU. It can not only accurately obtain the external parameters of the lidar and the IMU, but also accurately calibrate the time synchronization error between the IMU and the lidar, and optimize the calibration cost.
[0087] In this embodiment, a parameter calibration method of a laser radar and an IMU is provided, which can be used in mobile terminals such as mobile phones, tablet computers, etc. Figure 2 is a flow chart of a parameter calibration method of a laser radar and an IMU according to an embodiment of the present invention. Figure 2 As shown, in the above step S101, the pose transformation result corresponding to the target point cloud data set sampled by the laser radar within the first preset time is obtained by the following steps. The process includes the following steps:
[0088] Step S201, through the ICP registration algorithm, each time using the current frame and the previous frame of point cloud data adjacent to it as the registration object, successively registering each point cloud data in the point cloud data set collected by the laser radar within the first preset time.
[0089] Step S202, obtaining from each registration result a pose transformation result corresponding to a target point cloud data set sampled by the laser radar within a first preset time.
[0090] Specifically, the ICP registration algorithm, full name Iterative Closest Point, is a geometric registration algorithm widely used in computer vision and image processing. It is mainly used to match two or more point cloud data sets, find the optimal coordinate transformation relationship between point clouds through iterative optimization, so that point cloud data under different coordinates can be merged into the same coordinate system.
[0091] For example, the specific flow chart of the ICP registration algorithm is shown in the following steps:
[0092] Step a1, the point cloud data of the current frame is P2, the point cloud data of the previous frame of the current frame is P1, and each point ps in the point cloud data of the current frame P2 is transformed according to the currently estimated relative transformation T. And according to the transformed position, the point pt closest to it is found in the point cloud data of the previous frame of the current frame P1. When the distance between the two is less than the preset threshold distance_th, it is considered that a pair of matching points ps and pt is found.
[0093] Step a2: Based on the matching relationship of the point pairs found above, the following optimization equation is constructed based on the minimum distance error after matching the two frames of point clouds:
[0094]
[0095] Among them, ps and pt are one-to-one corresponding points in the source point cloud and the target point cloud.
[0096] Step a3: According to the constraint equation in step a2, the relative rotation transformation and translation transformation between the two frame point clouds can be obtained;
[0097] Specifically, we first give the conclusion. When the corresponding relationship of the points is known, let Represent the centroid of the source point cloud and the target point cloud respectively, let make This is a 3×3 matrix. Performing singular value decomposition (SVD) on H yields H = UΣV T , then the optimal rotation of the point-to-point ICP problem is: R * =VU T , the optimal translation is:
[0098] Step a4, after calculating the current rotation and translation transformation T in the above step a3, continue to repeat the process of step a1-step a3, iteratively solve and obtain the final transformation T. (The iteration stop condition can be set to the maximum number of iterations, for example, it can be set to 5, that is, the iteration is stopped after 5 rounds. In addition, the rotation and translation transformation T at the beginning of the first iteration can be set to the unit transformation, and the two frames of point cloud data are considered to be overlapped).
[0099] The disclosed embodiment executes the above steps S201 and S202, and the accuracy of the target point cloud data set sampled by the laser radar can be ensured through the ICP registration algorithm.
[0100] In this embodiment, a parameter calibration method of a laser radar and an IMU is provided, which can be used in the above-mentioned mobile terminals, such as mobile phones, tablet computers, etc. Figure 3 is a flow chart of a parameter calibration method of a laser radar and an IMU according to an embodiment of the present invention. Figure 3 As shown, in the above step S101, candidate frame point cloud data for lidar loop detection is obtained, and the process includes the following steps:
[0101] Step S301, searching for the first position transformation data corresponding to the current frame point cloud data from the position transformation results corresponding to the target point cloud data set sampled by the laser radar within the first preset time.
[0102] Specifically, the pose transformation results corresponding to the target point cloud data set of the laser radar odometer within a continuous period of time (the first preset time) are traversed to find the current frame point cloud data P d The corresponding pose transformation data, that is, the first pose transformation data, can be obtained by T d Indicates that the current frame point cloud data P d The corresponding current time is time_d.
[0103] Step S302, searching for remaining frame point cloud data sets excluding the current frame point cloud data set from the target point cloud data set sampled by the laser radar within the first preset time.
[0104] Specifically, for example, the target point cloud dataset includes {P1, P2, P3, P4...P d ...P n}, remove the current frame point cloud data P d The remaining frame point cloud datasets include {P1, P2, P3, P4...P n}.
[0105] Step S303, obtaining a pose transformation data set corresponding to the remaining frame point cloud data sets from the pose transformation results corresponding to the target point cloud data set sampled by the laser radar within the first preset time.
[0106] Specifically, for example, the target point cloud dataset includes {P1, P2, P3, P4...P d ...P n}, the corresponding pose transformation dataset includes {T1, T2, T3, T4...T d ...T n}, the remaining frame point cloud datasets {P1, P2, P3, P4...P n} corresponding to the pose transformation dataset, including {T1, T2, T3, T4...T n}.
[0107] Step S304, obtaining each moment corresponding to the remaining frame point cloud data sets and the current moment corresponding to the current frame point cloud data sets.
[0108] Specifically, the remaining frame point cloud datasets {P1, P2, P3, P4...P n} Each moment corresponding to is {time1, time2, time3, time4...time_n}.
[0109] Step S305, calculating the time difference between each moment corresponding to the remaining frame point cloud data set and the current moment corresponding to the current frame point cloud data set.
[0110] Specifically, the current moment corresponding to the current frame point cloud data is time_d, and each moment corresponding to the remaining frame point cloud data sets is {time1, time2, time3, time4...time_n}, and the side calculates time1-time_d=t1, time2-time_d=t2, time3-time_d=t3, time4-time_d=t4,...time_n-time_d=tn.
[0111] Step S306: Obtain second pose transformation data corresponding to each frame of point cloud data of the remaining frame point cloud data sets from the pose transformation data sets corresponding to the remaining frame point cloud data sets.
[0112] Specifically, the remaining frame point cloud datasets {P1, P2, P3, P4...P n} corresponding to the pose transformation dataset, including {T1, T2, T3, T4...T n}, the second pose transformation data corresponding to each frame of the remaining point cloud data set is T1, T2, T3, T4, ...T n .
[0113] Step S307, calculating the distance difference between the second pose transformation data corresponding to each frame of point cloud data of the remaining frame point cloud data set and the first pose transformation data corresponding to the current frame point cloud data.
[0114] Specifically, for example, T1-T d =d1,T2-T d =d2,T3-T d =d3,T4-T d =d4, ..., T n -T d =dn.
[0115] Step S308: taking the point cloud data corresponding to the time difference being greater than the first preset threshold and the distance difference being less than the second preset threshold as the candidate point cloud data set.
[0116] Specifically, the first preset threshold is t0, and from the above calculated time differences t1, t2, t3, t4, ... tn, find the time difference greater than the first preset threshold t0; at the same time, the second preset threshold is d0, and from the above d1, d2, d3, d4 ... dn, find the point cloud data corresponding to the point cloud data less than the second preset threshold d0 as the candidate point cloud data set. For example, the point cloud data that meet the conditions at the same time are P1, P3, ... P n , then the candidate point cloud dataset is {P1, P3, ...P n}.
[0117] Step S309 , through the ICP registration algorithm, each time using the current frame point cloud data and each frame point cloud data in the candidate point cloud data set as the registration object, and successively registering each point cloud data in each frame point cloud data in the candidate point cloud data set.
[0118] Step S310 , obtaining point cloud data with the largest number of successful registrations from all registration results of the candidate point cloud data sets as candidate frame point cloud data.
[0119] Specifically, the candidate point cloud dataset {P1, P3, ...P n} and the point cloud data of each frame in the current frame P d As the registration object, each point cloud data in each frame of point cloud data in the candidate point cloud data set is successively registered by the ICP registration algorithm. The above steps a1 to a4 of the ICP registration algorithm have been introduced and will not be repeated here.
[0120] The disclosed embodiment obtains candidate frame point cloud data of lidar loop detection by executing the above steps S301 to S310, which is equivalent to calculating the candidate frame point cloud data through loop detection, and further utilizing the candidate frame point cloud data to optimize the error of parameter calibration of the lidar and IMU and improve the estimation accuracy of dynamic calibration parameters.
[0121] In this embodiment, a parameter calibration method of a laser radar and an IMU is provided, which can be used in the above-mentioned mobile terminals, such as mobile phones, tablet computers, etc. Figure 4 is a flow chart of a parameter calibration method of a laser radar and an IMU according to an embodiment of the present invention. Figure 4 As shown, the above step S103, obtaining point cloud pose mapping data, the process includes the following steps:
[0122] S401, obtaining candidate time and third pose transformation data corresponding to the candidate frame point cloud data.
[0123] Specifically, for example, if the candidate frame point cloud data is P1, the candidate time corresponding to the candidate frame point cloud data is time1, the posture transformation data corresponding to the candidate frame point cloud data is the third posture transformation data, and the third posture transformation data is T1.
[0124] S402, obtaining fourth posture transformation data and target time corresponding to each frame of target point cloud data.
[0125] Specifically, the target point cloud dataset includes {P1, P2, P3, P4...P d ...P n}, the corresponding pose transformation data of each frame of the target point cloud data is the fourth pose transformation data, for example, P1 corresponds to T1, P2 corresponds to T2, P3 corresponds to T3, P4 corresponds to T4, P d Corresponding to T d , P n Corresponding to T n , T1, T2, T3, and T4 all belong to the fourth pose transformation data separately, and the target time corresponding to each frame of target point cloud data is time1, time2, time3, time4...time_n respectively.
[0126] S403, generating point cloud pose mapping data based on the candidate time, the third pose transformation data, the fourth pose transformation data corresponding to each frame of target point cloud data, and the target time corresponding to each frame of target point cloud data.
[0127] Specifically, for example, the point cloud pose mapping data includes: {T nn 、time_ nn , T1, time1}, or, {T nn 、time_ nn , T2, time2}, or, {T nn 、time_ nn , T3, time3}, or, {T nn 、time_ nn , T4, time4}, or, {T nn 、time_ nn , T d , time_d}, or, {T nn 、time_nn , T n 、time_ n}.
[0128] In a specific example, the above step S103, based on the point cloud pose mapping data, obtains the first rotation angular velocity between the candidate frame point cloud data and each frame target point cloud data, which is performed by the following formula:
[0129]
[0130] Among them, T ii is the fourth pose transformation data corresponding to each frame of target point cloud data, T nn The third point cloud data corresponding to the candidate frame is the pose transformation data, time _nn is the candidate time corresponding to the candidate frame point cloud data, time_ ii is the target moment corresponding to each frame of target point cloud data, and ω1 is the first rotation angular velocity between the candidate frame point cloud data and each frame of target point cloud data.
[0131] By calculating the first rotation angular velocity between the candidate frame point cloud data and each frame target point cloud data in the laser odometer, the first rotation angular velocity corresponding to each timestamp in the laser odometer is obtained.
[0132] In this embodiment, a parameter calibration method of a laser radar and an IMU is provided, which can be used in the above-mentioned mobile terminals, such as mobile phones, tablet computers, etc. Figure 5 is a flow chart of a parameter calibration method of a laser radar and an IMU according to an embodiment of the present invention. Figure 5 As shown, in the above step S104, the second rotation angular velocity measured by the IMU at each correction moment within the second preset time is obtained, and the correction moment within the second preset time is obtained based on the target time deviation adjustment. The process includes the following steps:
[0133] Step S501, setting a preset time error adjustment range and a preset time adjustment step.
[0134] Specifically, the preset time error adjustment range is [-time1, time1], for example, [-1s, 1s], and the preset time adjustment step is represented by time_step, for example, set to 0.001s. The preset time adjustment range and the preset time adjustment step can be flexibly set according to actual needs.
[0135] Step S502 , starting from the left end point value of the preset time error adjustment range, the target time deviation is gradually increased according to the preset time adjustment step length until it increases to the right end point value of the preset time error adjustment range.
[0136] Step S503: subtract the target time deviation from the original time corresponding to each IMU measurement data to obtain the corrected time corresponding to each IMU measurement data.
[0137] Specifically, for each original moment of the IMU measurement data in the time range of [-time1, time1], iterate from -time1 and gradually increase the preset time step time_step to obtain the target time deviation d m . (and d m =n*time_step, n is an integer between 0 and 2*time1 / time_step. For example, in the interval [-time1, time1], sampling is performed according to the step length of time_step. For example, if the interval is set to [-1s, 1s], sampling is performed according to the step length of 0.001s. Then the candidate time deviations di that can be obtained are: -0.999, -0.998, -0.997......0.995, 0.996, 0.997, 0.998, 0.999, 1.0. Further, according to the target time deviation d m Correct each timestamp of the IMU by subtracting d from the sampling time of the IMU measurement data. m , and then obtain the correction time corresponding to each IMU measurement data.
[0138] Step S504, at the correction time corresponding to each IMU measurement data, obtain the second rotation angular velocity measured by the IMU at each correction time within the second preset time.
[0139] Specifically, at the correction moment corresponding to each IMU measurement data, the second rotation angular velocity measured by the IMU at each correction moment within a continuous time (a second preset time) is obtained and represented by ω2.
[0140] The embodiment of the present disclosure executes the above steps S501 to S504 and gradually optimizes the measurement moment of each IMU measurement data, thereby facilitating improving the accuracy of measuring the second rotation angular velocity measured by the IMU at each correction moment within the second preset time.
[0141] In this embodiment, a parameter calibration method of a laser radar and an IMU is provided, which can be used in the above-mentioned mobile terminals, such as mobile phones, tablet computers, etc. Figure 6 is a flow chart of a parameter calibration method of a laser radar and an IMU according to an embodiment of the present invention. Figure 6 As shown, in the above step S105, the time synchronization error parameter between the laser radar and the IMU is obtained based on the first rotation angular velocity between the candidate frame point cloud data and each frame target point cloud data and the second rotation angular velocity measured by the IMU at each correction moment within the second preset time. The process includes the following steps:
[0142] Step S601, searching for a first rotational angular velocity between the candidate frame point cloud data and each frame target point cloud data at each correction moment within a second preset time based on the IMU.
[0143] Specifically, at each correction moment within the second preset time, the IMU searches the laser radar odometer for a first rotation angular velocity ω1 between the candidate point cloud data and each frame of target point cloud data.
[0144] Step S602, obtaining the target time deviation corresponding to the maximum autocorrelation function value between the first rotation angular velocity between the candidate frame point cloud data and each frame target point cloud data and the second rotation angular velocity measured by the IMU at each correction moment within the second preset time through the following autocorrelation function.
[0145]
[0146] Among them, D Z is the autocorrelation function value, d m is the target time deviation, ω1 is the first rotation angular velocity between the candidate frame point cloud data and each frame of target point cloud data, and ω2 is the second rotation angular velocity measured by the IMU at each correction moment within the second preset time.
[0147] Step S603: taking the target time deviation corresponding to the autocorrelation as a time synchronization error parameter.
[0148] The disclosed embodiment executes the above steps S601 to S603 to autocorrelate the first rotation angular velocity between the candidate frame point cloud data and each frame target point cloud data with the second rotation velocity of the IMU at each correction moment within the second preset time, thereby accurately calibrating the time synchronization error parameters between the lidar and the IMU.
[0149] In this embodiment, a parameter calibration method of a laser radar and an IMU is provided, which can be used in the above-mentioned mobile terminals, such as mobile phones, tablet computers, etc. Figure 7 is a flow chart of a parameter calibration method of a laser radar and an IMU according to an embodiment of the present invention. Figure 7 As shown, the following steps are also included:
[0150] Step S701, based on the time synchronization error parameter between the laser radar and the IMU, correct each current moment of the IMU.
[0151] Specifically, for example, the time synchronization error parameter is expressed as d t Indicates that d t = 0.005s, subtract d from each current moment of the IMU t , so as to achieve the purpose of correcting the IMU at each current moment.
[0152] Step S702: Based on the corrected IMU, a third rotation angular velocity measured by the IMU at each current moment within a third preset time is obtained.
[0153] Specifically, the third rotation angular velocity measured by the IMU at each timestamp (each current moment) within another continuous period of time (within the third preset time) can be represented by ω3.
[0154] Step S703, based on the first rotation angular velocity between the candidate frame point cloud data and each frame target point cloud data, and the third rotation angular velocity measured by the IMU at each current moment within a third preset time, a rotation residual equation is constructed to obtain the rotation extrinsic parameter data of the IMU.
[0155] In a specific example, the rotation residual equation is constructed by the following formula:
[0156] ErrorA=Rω1-ω3
[0157] Among them, ErrorA is the function value of the rotation residual equation, R is the rotation matrix, ω1 is the first rotation angular velocity between the candidate frame point cloud data and each frame of target point cloud data, and ω3 is the third rotation angular velocity measured by the IMU at each current moment within the third preset time.
[0158] The rotation external parameter data R corresponding to each current moment of the IMU within the third preset time is calculated by the above formula.
[0159] By executing the above steps S701-S703, based on the time synchronization error parameters between the lidar and the IMU, each current moment of the IMU is corrected, and the rotation external parameter data of the IMU can be accurately calibrated.
[0160] In this embodiment, a parameter calibration device for a laser radar and an IMU is also provided, which is used to implement the above-mentioned embodiments and preferred implementation modes, and will not be repeated hereafter. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.
[0161] This embodiment provides a parameter calibration device for a laser radar and an IMU, such as Figure 8 As shown, including:
[0162] The candidate frame data acquisition module 81 is used to acquire candidate frame point cloud data for lidar loop detection based on the pose transformation result corresponding to the target point cloud data set sampled by the lidar within the first preset time;
[0163] The point cloud pose data acquisition module 82 is used to acquire point cloud pose mapping data between the candidate frame point cloud data and each frame of target point cloud data in the target point cloud data set;
[0164] A first angular velocity acquisition module 83 is used to acquire a first rotational angular velocity between the candidate frame point cloud data and each frame target point cloud data based on the point cloud pose mapping data;
[0165] A second angular velocity acquisition module 84 is used to acquire a second rotation angular velocity measured by the IMU at each correction moment within a second preset time, where the correction moment within the second preset time is obtained based on the target time deviation adjustment;
[0166] The time synchronization error acquisition module 85 is used to obtain the time synchronization error parameters between the laser radar and the IMU based on the first rotation angular velocity between the candidate frame point cloud data and each frame target point cloud data and the second rotation angular velocity measured by the IMU at each correction moment within the second preset time.
[0167] In some optional implementations, the candidate frame data acquisition module 81 includes:
[0168] The first registration submodule is used to use the ICP registration algorithm to use the current frame and the previous frame of point cloud data adjacent to it as the registration object each time, and to successively register each point cloud data in the point cloud data set collected by the laser radar within the first preset time;
[0169] The first acquisition submodule is used to obtain, from each registration result, a pose transformation result corresponding to a target point cloud data set sampled by the laser radar within a first preset time.
[0170] In some optional implementations, the candidate frame data acquisition module 81 includes:
[0171] A first search submodule is used to search for the first position transformation data corresponding to the current frame point cloud data from the position transformation results corresponding to the target point cloud data set sampled by the laser radar within the first preset time;
[0172] A second search submodule is used to search for remaining frame point cloud data sets excluding the current frame point cloud data from the target point cloud data set sampled by the laser radar within the first preset time;
[0173] The second acquisition submodule is used to obtain the pose transformation data set corresponding to the remaining frame point cloud data set from the pose transformation result corresponding to the target point cloud data set sampled by the laser radar within the first preset time,
[0174] The third acquisition submodule is used to obtain each moment corresponding to the remaining frame point cloud data set and the current moment corresponding to the current frame point cloud data;
[0175] The first calculation submodule is used to calculate the time difference between each moment corresponding to the remaining frame point cloud data set and the current moment corresponding to the current frame point cloud data;
[0176] The fourth acquisition submodule is used to obtain the second posture transformation data corresponding to each frame of point cloud data of the remaining frame point cloud data set from the posture transformation data set corresponding to the remaining frame point cloud data set;
[0177] The second calculation submodule is used to calculate the distance difference between the second pose transformation data corresponding to each frame of point cloud data of the remaining frame point cloud data set and the first pose transformation data corresponding to the current frame point cloud data;
[0178] A fifth acquisition submodule, configured to take point cloud data corresponding to a time difference greater than a first preset threshold and a distance difference less than a second preset threshold as a candidate point cloud data set;
[0179] The second registration submodule is used to use the ICP registration algorithm to use the current frame point cloud data and each frame point cloud data in the candidate point cloud data set as the registration object each time, and to successively register each point cloud data in each frame point cloud data in the candidate point cloud data set;
[0180] The sixth acquisition submodule is used to acquire point cloud data with the largest number of successful registrations from all the registration results of the candidate point cloud data sets as candidate frame point cloud data.
[0181] In some optional implementations, the first angular velocity acquisition module 83 includes:
[0182] The first acquisition submodule is used to obtain the candidate time and the third posture transformation data corresponding to the candidate frame point cloud data;
[0183] The second acquisition submodule is used to obtain the fourth posture transformation data and target time corresponding to each frame of target point cloud data;
[0184] The posture generation submodule is used to generate point cloud posture mapping data based on the candidate time, the third posture transformation data, the fourth posture transformation data corresponding to each frame of target point cloud data, and the target time corresponding to each frame of target point cloud data.
[0185] In some optional implementations, the second angular velocity acquisition module 84 includes:
[0186] A setting submodule is used to set a preset time error adjustment range and a preset time adjustment step;
[0187] An adjustment submodule, used to gradually increase the target time deviation from the left end point value of the preset time error adjustment range according to the preset time adjustment step length until it increases to the right end point value of the preset time error adjustment range;
[0188] The acquisition submodule is used to subtract the target time deviation from the original time corresponding to each IMU measurement data to obtain the corrected time corresponding to each IMU measurement data;
[0189] The correction submodule is used to obtain the second rotation angular velocity measured by the IMU at each correction moment within a second preset time at the correction moment corresponding to each IMU measurement data.
[0190] In some optional implementations, the time synchronization error acquisition module 85 includes:
[0191] A search submodule, used to search for a first rotation angular velocity between the candidate frame point cloud data and each frame target point cloud data based on the IMU at each correction moment within a second preset time;
[0192] An association submodule is used to obtain a target time deviation corresponding to the maximum autocorrelation function value between a first rotation angular velocity between the candidate frame point cloud data and each frame target point cloud data and a second rotation angular velocity measured by the IMU at each correction moment within a second preset time through the following autocorrelation function;
[0193]
[0194] Among them, D Z is the autocorrelation function value, d m is the target time deviation, ω1 is the first rotation angular velocity between the candidate frame point cloud data and each frame target point cloud data, and ω2 is the second rotation angular velocity measured by the IMU at each correction moment within the second preset time;
[0195] The parameter determination submodule is used to use the target time deviation corresponding to the autocorrelation as a time synchronization error parameter.
[0196] In some optional implementations, the parameter calibration device of the laser radar and the IMU in the embodiment of the present disclosure further includes:
[0197] The IMU correction submodule is used to correct each current moment of the IMU based on the time synchronization error parameter between the lidar and the IMU;
[0198] An angular velocity acquisition submodule, used to acquire, based on the corrected IMU, a third rotation angular velocity measured by the IMU at each current moment within a third preset time;
[0199] The rotation extrinsic parameter acquisition submodule obtains the rotation extrinsic parameter data of the IMU by constructing a rotation residual equation based on the first rotation angular velocity between the candidate frame point cloud data and each frame target point cloud data and the third rotation angular velocity measured by the IMU at each current moment within a third preset time.
[0200] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.
[0201] The parameter calibration device of the laser radar and IMU in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.
[0202] An embodiment of the present invention also provides a computer device having a parameter calibration device for a laser radar and an IMU.
[0203] See also Fig. 9 , Fig. 9 is a schematic diagram of the structure of a computer device provided by an optional embodiment of the present invention, such as Fig. 9 As shown, the computer device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components are connected to each other using different buses for communication, and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in or on the memory to display the graphical information of the GUI on an external input / output device (such as, a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Fig. 9 A processor 10 is taken as an example.
[0204] The processor 10 may be a central processing unit, a network processor or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic or any combination thereof.
[0205] The memory 20 stores instructions executable by at least one processor 10, so that the at least one processor 10 executes the method shown in the above embodiment.
[0206] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely arranged relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.
[0207] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 20 may also include a combination of the above types of memory.
[0208] The computer device further comprises a communication interface 30 for the computer device to communicate with other devices or a communication network.
[0209] The embodiment of the present invention also provides a computer-readable storage medium. The method according to the embodiment of the present invention can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium through a network download, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state hard disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.
[0210] Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations are all within the scope defined by the appended claims.
Claims
1. A parameter calibration method for laser radar and IMU, It is characterized in that The method comprises: Based on the pose transformation result corresponding to the target point cloud data set sampled by the laser radar within the first preset time, the candidate frame point cloud data for the laser radar loop closure detection is obtained; Acquire point cloud pose mapping data between the candidate frame point cloud data and each frame of target point cloud data in the target point cloud data set; Based on the point cloud pose mapping data, obtaining a first rotational angular velocity between the candidate frame point cloud data and each frame target point cloud data; Obtain a second rotation angular velocity measured by the IMU at each correction moment within a second preset time, wherein the correction moment within the second preset time is obtained based on the target time deviation adjustment; Based on the first rotational angular velocity between the candidate frame point cloud data and the target point cloud data of each frame and the second rotational angular velocity measured by the IMU at each correction moment within the second preset time, the time synchronization error parameter between the laser radar and the IMU is obtained.
2. The method according to claim 1, It is characterized in that Obtaining a pose transformation result corresponding to a target point cloud data set sampled by the laser radar within a first preset time, including: By using the ICP registration algorithm, each time using the current frame and the previous frame of point cloud data adjacent to it as the registration object, each point cloud data in the point cloud data set collected by the laser radar within the first preset time is successively registered; The pose transformation result corresponding to the target point cloud data set sampled by the laser radar within the first preset time is obtained from each registration result.
3. The method according to claim 1 or 2, It is characterized in that Obtaining candidate frame point cloud data for the lidar loop detection, including: From the pose transformation results corresponding to the target point cloud data set sampled by the laser radar within the first preset time, searching for the first pose transformation data corresponding to the current frame point cloud data; Searching for the remaining frame point cloud data sets excluding the current frame point cloud data from the target point cloud data set sampled by the laser radar within the first preset time; Obtaining a pose transformation dataset corresponding to the remaining frame point cloud dataset from a pose transformation result corresponding to the target point cloud dataset sampled by the laser radar within a first preset time; Obtain each moment corresponding to the remaining frame point cloud data sets and the current moment corresponding to the current frame point cloud data; Calculate the time difference between each moment corresponding to the remaining frame point cloud data set and the current moment corresponding to the current frame point cloud data; From the pose transformation data set corresponding to the remaining frame point cloud data set, obtain second pose transformation data corresponding to each frame point cloud data of the remaining frame point cloud data set; Calculate the distance difference between the second pose transformation data corresponding to each frame of point cloud data of the remaining frame point cloud data set and the first pose transformation data corresponding to the current frame point cloud data; The point cloud data corresponding to the time difference being greater than a first preset threshold and the distance difference being less than a second preset threshold is taken as a candidate point cloud data set; By using the ICP registration algorithm, each time using the current frame point cloud data and each frame point cloud data in the candidate point cloud data set as the registration object, each point cloud data in each frame point cloud data in the candidate point cloud data set is successively registered; The point cloud data with the largest number of successful registrations is obtained from all the registration results of the candidate point cloud data set as the candidate frame point cloud data.
4. The method according to claim 1, It is characterized in that Obtaining the point cloud pose mapping data, including: Obtaining candidate time and third pose transformation data corresponding to the candidate frame point cloud data; Obtaining fourth posture transformation data and target time corresponding to each frame of target point cloud data; Point cloud pose mapping data is generated based on the candidate time, the third pose transformation data, the fourth pose transformation data corresponding to each frame of target point cloud data, and the target time corresponding to each frame of target point cloud data.
5. The method according to claim 1, It is characterized in that The step of obtaining a second rotation angular velocity measured by the IMU at each correction moment within a second preset time, wherein the correction moment within the second preset time is obtained based on a target time deviation adjustment, comprises: Set the preset time error adjustment range and preset time adjustment step; Starting from the left end point value of the preset time error adjustment range, the target time deviation is obtained by gradually increasing the preset time adjustment step size until it increases to the right end point value of the preset time error adjustment range; Subtract the target time deviation from the original time corresponding to each IMU measurement data to obtain the corrected time corresponding to each IMU measurement data; At a correction moment corresponding to each IMU measurement data, a second rotation angular velocity measured by the IMU at each correction moment within a second preset time is obtained.
6. The method according to claim 1, It is characterized in that Based on the first rotation angular velocity between the candidate frame point cloud data and the target point cloud data of each frame and the second rotation angular velocity measured by the IMU at each correction moment within the second preset time, obtaining the time synchronization error parameter between the laser radar and the IMU includes: Based on the IMU at each correction moment within the second preset time, searching for a first rotation angular velocity between the candidate frame point cloud data and each frame target point cloud data; Obtaining the target time deviation corresponding to the maximum autocorrelation function value between the first rotation angular velocity between the candidate frame point cloud data and the target point cloud data of each frame and the second rotation angular velocity measured by the IMU at each correction moment within the second preset time through the following autocorrelation function; Among them, D Z is the autocorrelation function value, d m is the target time deviation, ω 1 is the first rotation angular velocity between the candidate frame point cloud data and the target point cloud data of each frame, ω 2 A second rotation angular velocity measured by the IMU at each correction moment within a second preset time; The target time deviation corresponding to the autocorrelation is used as the time synchronization error parameter.
7. The method according to claim 4, It is characterized in that Also includes: Based on the time synchronization error parameter between the laser radar and the IMU, correct each current moment of the IMU; Based on the corrected IMU, obtaining a third rotation angular velocity measured by the IMU at each current moment within a third preset time; Based on the first rotation angular velocity between the candidate frame point cloud data and the target point cloud data of each frame, and the third rotation angular velocity measured by the IMU at each current moment within a third preset time, the rotation external parameter data of the IMU is obtained by constructing a rotation residual equation.
8. A parameter calibration device for laser radar and IMU, It is characterized in that The device comprises: A candidate frame data acquisition module, used to acquire candidate frame point cloud data for lidar loop detection based on a pose transformation result corresponding to a target point cloud data set sampled by the lidar within a first preset time; A point cloud pose data acquisition module, used to acquire point cloud pose mapping data between the candidate frame point cloud data and each frame of target point cloud data in the target point cloud data set; A first angular velocity acquisition module, used for acquiring a first rotational angular velocity between the candidate frame point cloud data and each frame target point cloud data based on the point cloud pose mapping data; A second angular velocity acquisition module is used to acquire a second rotation angular velocity measured by the IMU at each correction moment within a second preset time, where the correction moment within the second preset time is obtained based on the target time deviation adjustment; The time synchronization error acquisition module is used to obtain the time synchronization error parameter between the laser radar and the IMU based on the first rotation angular velocity between the candidate frame point cloud data and the target point cloud data of each frame and the second rotation angular velocity measured by the IMU at each correction moment within the second preset time.
9. A computer device, It is characterized in that include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the parameter calibration method of the laser radar and IMU according to any one of claims 1 to 6 by executing the computer instructions.
10. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the parameter calibration method of the laser radar and IMU according to any one of claims 1 to 6.
Citation Information
Patent Citations
Method and system for laser-IMU external parameter calibration
CN111207774A
Laser radar and inertial measurement unit calibration method and system
CN114518111A
Positioning mapping method and device based on template matching and laser inertial navigation loose coupling
CN115077519A
Method for determining calibration parameters of laser radar and inertial measurement unit and related equipment
CN115265581A
Space-time external parameter calibration method, device and system for laser radar and combined inertial navigation
CN116182905A
