A precise and robust camera-IMU calibration method and system using GNSS assistance

Abstract

This invention discloses a precise and robust camera-IMU calibration method and system using GNSS assistance, belonging to the field of sensor calibration technology. The method includes: obtaining camera pose using visual odometry and IMU pose using a loose combination of GNSS / INS; constructing camera-IMU rotation and translation extrinsic parameter relationships; performing preliminary solutions based on the camera-IMU rotation extrinsic parameter relationships using least squares and inter-epoch iterations, and then performing preliminary solutions again using least squares and iterations in conjunction with the camera-IMU translation extrinsic parameter relationships; and optimizing the observability extrinsic parameters using data filtering and truncated singular value decomposition methods based on the preliminary solutions of the camera-IMU rotation and translation extrinsic parameter relationships to obtain precise solutions for the camera-IMU rotation and translation extrinsic parameters, thus completing the calibration. This invention effectively suppresses IMU error drift and filters out poorly observable sections, improving the overall observation quality and the accuracy and robustness of the calibration.

Description

Technical Field

[0001] This invention belongs to the field of sensor calibration technology, specifically relating to a precise and robust camera-IMU calibration method and system using GNSS assistance. Background Technology

[0002] Accurate calibration of cameras and inertial measurement units (IMUs) is not only fundamental to the effective fusion of image information and inertial data, but also a crucial step in the field of multi-sensor fusion, widely used in high-precision navigation, attitude estimation, robot perception, and unmanned systems. However, current calibration methods still face many challenges in terms of accuracy, stability, and adaptability to complex environments.

[0003] Existing camera-IMU calibration methods are mainly divided into external marker-based calibration and odometry-based calibration. External marker-based methods rely on specific visual markers such as Apriltag QR codes to provide reference points for the camera and IMU, but their dependence on external features and cumbersome experimental procedures limit their applicability. Odometry-based calibration methods can complete calibration in featureless environments, but due to the IMU's zero-bias error and the uncertainty of monocular vision scale, the calibration accuracy of these methods is easily affected. Furthermore, blindly processing all available data may not only use invalid observations with poor observability and gross errors, but also lead to unnecessarily high computational costs. This patent specifically focuses on data with poor observability. For example, commonly used mobile platforms (such as unmanned vehicles, robots, and unmanned boats) may experience observation degradation when moving vertically and around the horizontal axis, resulting in insufficient observability and limited solution stability and accuracy.

[0004] Therefore, it is necessary to design a precise and robust camera-IMU calibration method and system that uses Global Navigation Satellite System (GNSS) to improve positioning accuracy in order to address the above problems. Summary of the Invention

[0005] The purpose of this invention is to address the problems existing in the prior art by providing a precise and robust camera-IMU calibration method and system using GNSS assistance. By introducing GNSS assistance, an information theory-based data filtering method and a truncated singular value decomposition (TSVD) method are proposed, which not only improves the accuracy and quality of the data, but also effectively solves the problem of heterogeneous observation data in complex environments.

[0006] According to one aspect of this specification, a precise and robust camera-IMU calibration method using GNSS assistance is provided, comprising:

[0007] The camera pose is obtained by visual odometry and the IMU pose is obtained by GNSS / INS loose combination method. The two poses are then obtained by inter-epoch difference to obtain their respective inter-epoch relative poses.

[0008] Based on their relative poses between epochs, the camera-IMU rotation extrinsic parameter relationship and the camera-IMU translation extrinsic parameter relationship are constructed.

[0009] Based on the camera-IMU rotation extrinsic parameter relationship, the camera-IMU rotation extrinsic parameter is initially solved by least squares and interepoch iteration. Combining the results of the initial solution of the camera-IMU rotation extrinsic parameter and the camera-IMU translation extrinsic parameter relationship, the camera-IMU translation extrinsic parameter is initially solved by least squares calculation and iteration.

[0010] Based on the preliminary solutions of the camera-IMU rotation extrinsic parameters and the camera-IMU translation extrinsic parameters, observable extrinsic parameters are optimized using data filtering and truncated singular value decomposition methods to obtain accurate solutions for the camera-IMU rotation and translation extrinsic parameters, thus completing the calibration.

[0011] Furthermore, the camera-IMU rotation extrinsic parameters are initially solved using least squares and inter-epoch iteration, including:

[0012] The rotation matrix in the camera-IMU rotation extrinsic parameter relationship is converted into a quaternion, and the quaternion is solved by least squares.

[0013] A system of rotation equations is constructed based on K time points. The system of rotation equations is solved through inter-epoch iterations to obtain the preliminary solution of the camera-IMU rotation extrinsic parameters.

[0014] Furthermore, the camera-IMU translation extrinsic parameters are initially solved using least squares and iteration, including:

[0015] The camera-IMU translation extrinsic parameter relationship is transformed into a system of translation equations. The least squares method is used to calculate the formula. The formula is then iteratively solved based on the preliminary solution of the camera-IMU rotation extrinsic parameter to obtain the preliminary solution of the camera-IMU translation extrinsic parameter.

[0016] Furthermore, observability extrinsic parameter optimization is performed using data filtering and truncated singular value decomposition methods, including:

[0017] Based on the preliminary solutions of the camera-IMU rotation extrinsic parameters and the camera-IMU translation extrinsic parameters, normal equations and Fisher information matrices are constructed, and data filtering is performed.

[0018] Based on the filtered data segments, the truncated singular value decomposition method is used to obtain singular values ​​that meet the threshold limit for solving, resulting in optimized camera-IMU rotation extrinsic parameters and camera-IMU translation extrinsic parameters.

[0019] Further, data filtering is performed, including:

[0020] Calculate the minimum singular value of the Fisher information matrix for different data segments, and filter out the data segments with the most information.

[0021] According to one aspect of this specification, a precise and robust camera-IMU calibration system using GNSS assistance is provided, comprising:

[0022] The pose acquisition module is used to acquire camera pose using visual odometry and IMU pose using GNSS / INS loose combination method. The two poses are then used to obtain their respective epoch relative poses through epoch difference.

[0023] The module for constructing extrinsic parameter relationships is used to construct camera-IMU rotation extrinsic parameter relationships and camera-IMU translation extrinsic parameter relationships based on their respective relative poses between epochs.

[0024] The preliminary solution module is used to perform preliminary solution of camera-IMU rotation extrinsic parameters based on the camera-IMU rotation extrinsic parameter relationship, using least squares and interepoch iteration. Combining the results of the preliminary solution of camera-IMU rotation extrinsic parameters with the camera-IMU translation extrinsic parameter relationship, the module uses least squares calculation and iteration to perform preliminary solution of camera-IMU translation extrinsic parameters.

[0025] The extrinsic parameter optimization and calibration module is used to optimize the observable extrinsic parameters based on the preliminary solutions of the camera-IMU rotation extrinsic parameters and the preliminary solutions of the camera-IMU translation extrinsic parameters. It employs data filtering and truncated singular value decomposition methods to obtain the exact solutions of the camera-IMU rotation extrinsic parameters and the camera-IMU translation extrinsic parameters, thus completing the calibration.

[0026] According to one aspect of this specification, an electronic device is provided, including a memory and a processor, the memory storing a computer program, characterized in that the processor executes the computer program to implement the steps of the GNSS-assisted precise and robust camera-IMU calibration method.

[0027] According to one aspect of this specification, a computer-readable storage medium is provided having a computer program stored thereon, characterized in that, when the computer program is executed by a processor, it implements the steps of the GNSS-assisted precise and robust camera-IMU calibration method.

[0028] Compared with the prior art, the beneficial effects of the present invention are:

[0029] 1. The present invention proposes a precise and robust camera-IMU calibration method and system using GNSS assistance. The IMU pose is calculated by GNSS / INS loose combination algorithm, which effectively suppresses the error drift of the IMU during long-term use and improves the calibration accuracy.

[0030] 2. The present invention proposes a precise and robust camera-IMU calibration method and system using GNSS assistance. It adopts a data filtering method based on information theory to divide camera and IMU observations into different segments, effectively filtering out segments with poor observability and improving the overall observation quality.

[0031] 3. The present invention proposes a precise and robust camera-IMU calibration method and system using GNSS assistance. It employs a truncated singular value decomposition method to filter singular values, effectively reducing the impact of low-quality data on the calculation results and improving the accuracy and robustness of the calibration. Attached Figure Description

[0032] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0033] Figure 1 This is a flowchart of the camera-IMU calibration method according to an embodiment of the present invention;

[0034] Figure 2 This is a schematic diagram of the camera-IMU extrinsic parameters according to an embodiment of the present invention;

[0035] Figure 3 This is a schematic diagram of the combined pose of VO and GNSS / INS according to an embodiment of the present invention. Detailed Implementation

[0036] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only 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 ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0037] This invention provides a precise and robust camera-IMU calibration method using GNSS assistance, such as... Figure 1 As shown, the process includes: obtaining camera pose using visual odometry and IMU pose using a GNSS / INS loose combination method; obtaining the relative pose between each pose through inter-epoch difference; constructing camera-IMU rotation extrinsic parameter relationships and camera-IMU translation extrinsic parameter relationships based on the respective relative poses; performing preliminary solutions for camera-IMU rotation extrinsic parameters using least squares and inter-epoch iterations based on the camera-IMU rotation extrinsic parameter relationships; and performing preliminary solutions for camera-IMU translation extrinsic parameters using least squares and iterations, combining the results of the preliminary solutions for camera-IMU rotation extrinsic parameters and the camera-IMU translation extrinsic parameter relationships. Based on the preliminary solutions of the camera-IMU rotation extrinsic parameters and the camera-IMU translation extrinsic parameters, observable extrinsic parameters are optimized using data filtering and truncated singular value decomposition methods to obtain accurate solutions for the camera-IMU rotation and translation extrinsic parameters, thus completing the calibration.

[0038] Specifically, suppose the pose The pose of coordinate system a in coordinate system b is: ,in This represents rotation along the three coordinate axes. This represents the translation of the origin of the coordinate system. Let the camera-IMU extrinsic parameters be... Where I represents the IMU coordinate system and C represents the camera coordinate system, the calibration is to determine The value, such as Figure 2 As shown.

[0039] Specifically, the camera pose is calculated using the open-source visual odometry (VO) algorithm ORB-SLAM2, and is represented in the world coordinate system (W coordinate system) as follows: The pose is calculated using a loose combination of GNSS / INS methods, that is, combining GNSS RTK position results with IMU data, and represented in the navigation coordinate system (N frame) as follows: The W-series and N-series definitions differ, and due to the existence of camera-IMU extrinsic parameters, the camera trajectory and IMU trajectory do not coincide, such as... Figure 3As shown. Although the camera and IMU trajectories do not overlap, and their shapes are similar, they differ by one scale in translation.

[0040] Specifically, set for Tie The change in system position affects the camera pose. and posture The transformation relationship is as follows:

[0041] (1)

[0042] For two consecutive epochs The pose relationship between the camera and the IMU is as follows:

[0043] (2)

[0044] (3)

[0045] Inverting both sides of equation (2), we get:

[0046] (4)

[0047] Multiplying equation (4) by the left and right sides of equation (3) respectively, we get:

[0048]

[0049] (5)

[0050]

[0051] This eliminates the unknown quantity. , obtained the shape of Hand-eye alignment formula. Position Write as rotation Peaceful relocation By combining the forms, we can obtain:

[0052] (6)

[0053] in, This represents the monocular visual scale. Expanding equation (6), deriving the results, and extracting the relationships corresponding to rotation and translation, we can obtain:

[0054] (7)

[0055] (8)

[0056] Equations (7) and (8) are derived from the hand-eye calibration relationship and include camera-IMU extrinsic parameters. , The relation is also the solution. and The foundation.

[0057] Specifically, according to equation (7), let , The two represent consecutive epochs, respectively. The camera and IMU rotate relative to each other. The right side of equation (7) is... Moving it to the left yields:

[0058] (9)

[0059] Equation (9) is also The form is a hand-eye calibration relation for rotation. Rotation matrix The special orthogonal group SO(3) on the manifold is not easy to solve directly, so the rotation matrix is ​​used. Write as quaternions From the form, we can obtain:

[0060] (10)

[0061] in, , represent identity matrix Representing quaternions The real part, Representing quaternions Composed of the imaginary part vector, Representative vector Construct an antisymmetric matrix.

[0062] According to equation (10), At each moment, the following system of equations can be constructed:

[0063] (11)

[0064] Among them, quaternions Solving using weighted least squares, we obtain:

[0065] (12)

[0066] Specifically, in the actual solution process, each additional epoch of observation, i.e., the expansion of the equation system... And solving it again, we get:

[0067] (13)

[0068] The relative rotation relationship between the camera and the IMU is accurately solved through inter-epoch iteration. After the solution is obtained, Convert to rotation matrix This is used for the next step of adjusting the translation parameters. The estimate.

[0069] Specifically, if Figure 3 As shown, because a monocular camera cannot provide accurate scale information, The given translation and The translations provided by integrated navigation systems may have scale differences, where the red, green, and blue coordinate axes represent the corresponding sensor values. Axis. In , After obtaining two sets of poses by combining them, according to the formula 8) The translation parts of the two are related as follows:

[0070] (14)

[0071] in, Equation (14) includes the scale. Peaceful relocation Two unknown parameters, written in the following form:

[0072] (15)

[0073] Equation (15) can be simplified as , A sequence of keyframes can be composed of The system of equations can be calculated using least squares as follows:

[0074] (16)

[0075] In the actual solution, the camera-IMU relative rotation is solved in each iteration. Subsequently, based on historical observations, a new translational observation equation was added and iteratively solved to obtain... The exact solution.

[0076] Specifically, to overcome potential degradation due to insufficient motion stimulation or scene constraints, and to effectively reduce computational costs, the method proposed in this embodiment automatically selects the data segment with the highest information content for solving. Assuming that the measurements are independent and contain only Gaussian white noise, the parameter estimation is modeled as a maximum likelihood estimation (MLE) problem, which can be re-formulated as follows:

[0077] (17)

[0078] in, , Indicates from the first The residuals introduced in this measurement have a covariance of: . It is a diagonal matrix, with elements on the diagonal. Representing the The weights of each observation equation are calculated as follows: ,in Represents an exponential function. This is the scaling factor.

[0079] Specifically, in this embodiment of the invention, a variance matrix is ​​introduced into the observation equation of equation (17). (i.e., weights), and solve using iterative methods (such as the Gaussian-Newton method or the Levenberg-Marquard method). The initial value is obtained using the least squares estimate from the previous step, and in each iteration, the current estimate is used. Linearize the above nonlinear system: ,in This is the Jacobian matrix. From this, we can obtain the update increment (or error state estimate). The linear least squares problem is as follows:

[0080] (18)

[0081] The optimal solution is given by the normal equation:

[0082] (19)

[0083] Once calculated , then The estimate has been updated to ,in This represents the addition operation on a manifold.

[0084] Specifically, this embodiment of the invention proposes a data segment filtering method based on information theory, in equation (19) This is the Fisher information matrix for the MLE problem (Equation 17), which contains all the information from the measurement. In unmanned vehicle motion, due to the lack of motion excitation and / or environmental constraints, It might be rank deficient. Therefore, collecting information and data is necessary to ensure... Full rank (or observability) is crucial. Therefore, the entire dataset is first divided into multiple segments of constant time length, and then each segment is... Performing singular value decomposition on the matrix, omitting the indices, yields:

[0085] (20)

[0086] in, It is a diagonal matrix composed of singular values ​​arranged in descending order. , is an orthogonal matrix. An information metric (or observability index) based on singular values ​​for a single segment is used, as a larger information metric implies more information segments. The filtering expression is as follows:

[0087] (twenty one)

[0088] in, and These are the thresholds for the maximum and minimum singular values, respectively. Therefore, for a specific collected data sequence, the information metric of each segment is first evaluated, and then the segment with the highest information content is selected for optimization.

[0089] Specifically, embodiments of the present invention also propose an observability-based state update. To ensure the robustness (weak) unobservability of the proposed camera-IMU calibration, an observability constraint is explicitly executed in each iteration, i.e., only states falling within observable directions are updated. To this end, embodiments of the present invention employ TSVD to perform a low-rank approximation of the information matrix, i.e., at the information threshold chosen in the design. Choose singular values ​​above Based on the observability-guaranteed information matrix and the SVD-based matrix pseudoinverse, the following solution is adopted:

[0090] (twenty two)

[0091] Therefore, compared to solving using all singular values ​​of the full information matrix... In contrast (which may incorrectly introduce observable directions due to numerical issues), the solution using Equation (22) explicitly avoids updating unobservable (or weakly observable) parameters, thus conservatively ensuring observability.

[0092] The implementation of the various embodiments of the present invention is based on programmed processing by a device with processor functionality. Therefore, in practical engineering, the technical solutions and functions of the various embodiments of the present invention are encapsulated into various modules. Based on this reality, and building upon the above embodiments, the embodiments of the present invention provide a GNSS-assisted, precise, and robust camera-IMU calibration system, which is used to execute a GNSS-assisted, precise, and robust camera-IMU calibration method from the above method embodiments.

[0093] The system includes a pose acquisition module, which uses visual odometry to acquire camera pose and GNSS / INS loose combination method to acquire IMU pose. The two poses are then used to obtain their respective epoch relative poses through epoch difference.

[0094] The system comprises the following modules: an extrinsic parameter construction module for constructing camera-IMU rotation and translation extrinsic parameter relationships based on their relative poses between epochs; a preliminary solution module for performing preliminary solutions to camera-IMU rotation extrinsic parameters using least squares and inter-epoch iteration based on the camera-IMU rotation extrinsic parameter relationships, and a preliminary solution for camera-IMU translation extrinsic parameters using least squares calculation and iteration, combining the results of the preliminary solutions to the camera-IMU rotation and translation extrinsic parameter relationships; and an extrinsic parameter optimization and calibration module for performing observable extrinsic parameter optimization using data filtering and truncated singular value decomposition methods, based on the preliminary solutions to the camera-IMU rotation and translation extrinsic parameter relationships, to obtain accurate solutions for the camera-IMU rotation and translation extrinsic parameters, thus completing the calibration.

[0095] This invention provides a precise and robust camera-IMU calibration system using GNSS assistance. Addressing the issues of poor observability data and low accuracy and stability in calibration methods, this system employs several modules to divide camera and IMU observations into different segments. A truncated singular value decomposition method is used to filter singular values, effectively eliminating poorly observable segments, suppressing IMU error drift over long-term use, and improving calibration accuracy and robustness.

[0096] Based on the same inventive concept as the foregoing embodiments, this embodiment of the invention also provides an electronic device, including a memory and a processor. The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the GNSS-assisted accurate and robust camera-IMU calibration method proposed in the above embodiments.

[0097] This invention also provides a computer-readable storage medium storing a computer program. When executed by a processor, this program filters out poorly observable segments, improving overall observation quality and calibration accuracy and robustness. The storage medium can be any non-volatile storage device such as a hard disk, solid-state drive, flash drive, or optical disk, and is used to store computer program code and necessary data files. The stored computer program includes: a pose acquisition module, an extrinsic parameter relationship construction module, a preliminary solution module, and an extrinsic parameter optimization and calibration module.

[0098] Finally, it should be noted that the above specific embodiments are merely representative examples of the present invention. Obviously, the present invention is not limited to the above specific embodiments and many variations are possible. Any simple modifications, equivalent changes, and alterations made to the above specific embodiments based on the technical essence of the present invention should be considered within the protection scope of the present invention.

Claims

1. A precise and robust camera-IMU calibration method using GNSS assistance, characterized in that, include: The camera pose is obtained by visual odometry and the IMU pose is obtained by GNSS / INS loose combination method. The two poses are then obtained by inter-epoch difference to obtain their respective inter-epoch relative poses. Based on their relative poses between epochs, the camera-IMU rotation extrinsic parameter relationship and the camera-IMU translation extrinsic parameter relationship are constructed. Based on the camera-IMU rotation extrinsic parameter relationship, the camera-IMU rotation extrinsic parameter is initially solved by least squares and interepoch iteration. Combining the results of the initial solution of the camera-IMU rotation extrinsic parameter and the camera-IMU translation extrinsic parameter relationship, the camera-IMU translation extrinsic parameter is initially solved by least squares and iteration. Based on the preliminary solutions of the camera-IMU rotation extrinsic parameters and the camera-IMU translation extrinsic parameters, observable extrinsic parameters are optimized using data filtering and truncated singular value decomposition methods to obtain accurate solutions for the camera-IMU rotation and translation extrinsic parameters, thus completing the calibration. Observable extrinsic parameters are optimized using data filtering and truncated singular value decomposition methods, including: Based on the preliminary solutions of the camera-IMU rotation extrinsic parameters and the camera-IMU translation extrinsic parameters, normal equations and Fisher information matrices are constructed, and data filtering is performed. Based on the filtered data segments, the truncated singular value decomposition method is used to obtain singular values ​​that meet the threshold limit for solving, resulting in optimized camera-IMU rotation extrinsic parameters and camera-IMU translation extrinsic parameters.

2. The precise and robust camera-IMU calibration method using GNSS assistance according to claim 1, characterized in that, The initial solution for the camera-IMU rotation extrinsic parameters is obtained using least squares and inter-epoch iteration, including: The rotation matrix in the camera-IMU rotation extrinsic parameter relationship is converted into a quaternion, and the quaternion is solved by least squares. A system of rotation equations is constructed based on K time points. The system of rotation equations is solved through inter-epoch iterations to obtain the preliminary solution of the camera-IMU rotation extrinsic parameters.

3. The precise and robust camera-IMU calibration method using GNSS assistance according to claim 2, characterized in that, The initial solution for the camera-IMU translation extrinsic parameters is obtained using least squares and iteration, including: The camera-IMU translation extrinsic parameter relationship is transformed into a system of translation equations. The least squares method is used to calculate the formula. The formula is then iteratively solved based on the preliminary solution of the camera-IMU rotation extrinsic parameter to obtain the preliminary solution of the camera-IMU translation extrinsic parameter.

4. The precise and robust camera-IMU calibration method using GNSS assistance according to claim 1, characterized in that, Data filtering includes: Calculate the minimum singular value of the Fisher information matrix for different data segments, and filter out the data segments with the most information.

5. A precise and robust camera-IMU calibration system using GNSS assistance, characterized in that, include: The pose acquisition module is used to acquire camera pose using visual odometry and IMU pose using GNSS / INS loose combination method. The two poses are then used to obtain their respective epoch relative poses through epoch difference. The module for constructing extrinsic parameter relationships is used to construct camera-IMU rotation extrinsic parameter relationships and camera-IMU translation extrinsic parameter relationships based on their respective relative poses between epochs. The preliminary solution module is used to perform preliminary solution of camera-IMU rotation extrinsic parameters based on the camera-IMU rotation extrinsic parameter relationship, using least squares and interepoch iteration. Combining the results of the preliminary solution of camera-IMU rotation extrinsic parameters with the camera-IMU translation extrinsic parameter relationship, the module uses least squares calculation and iteration to perform preliminary solution of camera-IMU translation extrinsic parameters. The extrinsic parameter optimization and calibration module is used to optimize the observable extrinsic parameters based on the preliminary solutions of the camera-IMU rotation extrinsic parameters and the preliminary solutions of the camera-IMU translation extrinsic parameters after the initial solution of the camera-IMU rotation extrinsic parameters. It uses data filtering methods and truncated singular value decomposition methods to obtain the exact solutions of the camera-IMU rotation extrinsic parameters and the camera-IMU translation extrinsic parameters, and completes the calibration. Observable extrinsic parameters are optimized using data filtering and truncated singular value decomposition methods, including: Based on the preliminary solutions of the camera-IMU rotation extrinsic parameters and the camera-IMU translation extrinsic parameters, normal equations and Fisher information matrices are constructed, and data filtering is performed. Based on the filtered data segments, the truncated singular value decomposition method is used to obtain singular values ​​that meet the threshold limit for solving, resulting in optimized camera-IMU rotation extrinsic parameters and camera-IMU translation extrinsic parameters.

6. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the precise and robust camera-IMU calibration method using GNSS assistance as described in any one of claims 1 to 4.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the precise and robust camera-IMU calibration method using GNSS assistance as described in any one of claims 1 to 4.

Images