Laser scanning system parameter self-calibration method based on holder
By using an automated calibration method aided by environmental characteristics and cooperative goals, the problem of relying on manual operation for the calibration of internal and external parameters of laser scanning systems has been solved, achieving efficient and accurate parameter self-calibration and improving the automation level and practicality of the system.
Patent Information
- Application Number
- CN202511572499.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-10-30
- Publication Date
- 2025-12-30
AI Technical Summary
The calibration process of the internal and external parameters of existing laser scanning systems relies on manual operation, which is time-consuming, has unstable accuracy, and has high maintenance costs, making it difficult to adapt to the recalibration requirements when the equipment is vibrating or replaced.
We employ an internal parameter self-calibration method based on natural environmental characteristics and an external parameter-assisted calibration method based on sparse cooperative objectives. Through nonlinear optimization and point cloud data processing, we automatically decouple the internal and external parameters to achieve high-precision and automated calibration.
It improves calibration efficiency and accuracy consistency, reduces manual intervention, lowers usage costs, enhances the system's environmental adaptability and reliability, and shortens calibration time.
Smart Images
Figure CN121232162A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of three-dimensional measurement technology, and in particular to a method for self-calibrating parameters of a gimbal-based laser scanning system. Background Technology
[0002] Optical scanning systems, especially those combining high-precision LiDAR with flexible multi-degree-of-freedom intelligent gimbals, have become core 3D measurement devices in numerous industrial scenarios, including port bulk material handling automation, steel metallurgy, intelligent warehousing, and environmental perception for unmanned vehicles and robots. This system controls the horizontal and vertical movement of the gimbal to drive the LiDAR to scan a wide area, thereby acquiring high-density 3D point cloud data. This data is used for functions such as environmental modeling, material volume measurement, material position measurement, equipment positioning, and obstacle avoidance.
[0003] In order to accurately stitch together the local point cloud data collected by the LiDAR at different times and under different gimbal postures into a complete and unified 3D scene model, it is necessary to accurately know two sets of core parameters of the scanning system: 1. Internal parameters (internal parameters): also known as internal assembly parameters, specifically refers to the three-dimensional spatial translation vector and rotational attitude of the lidar's measurement coordinate system origin relative to the gimbal's rotation / pitch center axis (i.e., ... (Six degrees of freedom). This set of parameters is the system's inherent geometric structure parameter.
[0004] 2. External parameters: refer to the translation and rotation relationship of the reference coordinate system of the entire laser scanning system (as a whole) relative to a global, fixed world coordinate system (e.g., the coordinate system of the host equipment, the dock geodetic coordinate system, or the warehouse coordinate system).
[0005] In existing technologies, the calibration process for these intrinsic and extrinsic parameters typically relies heavily on manual operation. Technicians need to use auxiliary measuring tools such as tape measures, total stations, theodolites, and levels to perform tedious on-site measurements and alignments of the physical positions of the lidar and pan-tilt unit. The entire process is not only time-consuming (usually requiring several hours or even a whole day), but the calibration accuracy is also highly dependent on the operator's experience and sense of responsibility, resulting in poor consistency and unstable accuracy between different personnel or different calibration attempts.
[0006] More seriously, when the scanning equipment is displaced due to vibration or collision, or when it is reinstalled after maintenance or replacement, the above-mentioned complex manual calibration process must be repeated. This greatly increases the system's maintenance costs and downtime, reduces operational efficiency, and becomes a major bottleneck restricting the stability and deployment efficiency of industrial automation systems.
[0007] Therefore, there is an urgent need for an efficient, accurate, and repeatable automated calibration method that can replace manual operation in order to solve the aforementioned pain points in existing technologies. Summary of the Invention
[0008] To address the aforementioned technical problems, this invention provides a method for self-calibrating parameters of a gimbal-based laser scanning system. This method decouples the calibration process into two core stages: an internal parameter self-calibration stage based on natural environmental characteristics and an external parameter auxiliary calibration stage based on sparse cooperative objectives, thereby achieving high-precision and automated solution of the system's internal and external parameters.
[0009] The technical means employed in this invention are as follows: A self-calibration method for parameters of a gimbal-based laser scanning system includes: scanning in a natural scene with multiple non-parallel, non-coplanar planes to collect multi-pose data; establishing a nonlinear optimization model with the internal parameters to be calibrated as optimization variables, and iteratively optimizing based on plane consistency; calculating the increment of the internal parameters using the nonlinear optimization model, terminating the optimization after satisfying a preset convergence condition, and outputting the calibration values of the internal parameters; deploying cooperative targets with easily identifiable geometric features, and scanning the area containing the cooperative targets based on the calibration values of the internal parameters; processing the scanned point cloud and calculating the three-dimensional coordinates of the feature points in the reference coordinate system; establishing the correspondence between the reference coordinate system and the world coordinate system, solving the external parameter transformation matrix, and realizing parameter self-calibration of the scanning system.
[0010] Furthermore, the multi-pose data is a series of raw three-dimensional point cloud data, which are scanned in a natural scene with multiple non-parallel and non-coplanar plane features using a laser scanning system to be calibrated, and the gimbal attitude angle of each set of point cloud data is recorded.
[0011] Furthermore, during the iterative optimization process, the nonlinear optimization model transforms the collected multi-pose data from the lidar coordinate system to the reference coordinate system based on the current internal parameter assumptions, forming a fused global point cloud. In the fused global point cloud, multiple planar feature patches are automatically identified and extracted, and the flatness of each planar patch is calculated. The flatness represents the root mean square error (RMSE) or standard deviation of all points constituting the plane to the fitted plane. The sum of the weighted flatness errors of all planar patches is used to construct a cost function to evaluate the merits of the current internal parameter assumptions.
[0012] Furthermore, the internal parameter increment is calculated using the nonlinear optimization model, specifically including: using a nonlinear optimization algorithm, combined with the cost function and its gradient, to calculate the internal parameter increment, driving the cost function to converge in the minimization direction; when the iteration satisfies that the change in the cost function is less than a threshold or the maximum number of iterations is reached, the iterative optimization process is terminated, and the optimization result obtained is the calibration value of the internal parameter.
[0013] Furthermore, the cooperative objective is a target with easily identifiable geometric features placed in a natural scene with multiple non-parallel, non-coplanar planes, and the three-dimensional coordinates of the feature points of the cooperative objective in the world coordinate system are known values.
[0014] Furthermore, the processing of the scanned point cloud specifically includes: automatically extracting the point cloud clusters of each cooperative target from the background point cloud through point cloud segmentation, clustering and geometric feature analysis algorithms, performing accurate geometric fitting on each point cloud cluster, and calculating the three-dimensional coordinates of the feature points in the reference coordinate system of the scanning system.
[0015] Furthermore, the establishment of the correspondence between the reference coordinate system and the world coordinate system, and the solution of the external parameter transformation matrix, specifically includes: performing point cloud registration by solving the absolute orientation problem, using the singular value decomposition method or the iterative nearest point algorithm to calculate the rigid body transformation matrix from the reference coordinate system to the world coordinate system, which includes the rotation matrix and translation vector. The solved transformation matrix is the external parameter to be obtained.
[0016] Compared with the prior art, the present invention has the following advantages: This invention provides a self-calibration method for parameters of a gimbal-based laser scanning system. This fully automated calibration method utilizes automatic detection and matching algorithms to minimize manual intervention, improving calibration efficiency and accuracy consistency, and ensuring that calibration results are independent of operator subjective experience. The method employs simple, low-cost, and easily deployable calibration reference markers, eliminating reliance on dedicated calibration tools, reducing usage costs, and enhancing the method's applicability in various scenarios. By designing universal geometric constraints, the calibration method can adapt to multiple types of calibration markers, enhancing the system's environmental adaptability.
[0017] The gimbal-based laser scanning system parameter self-calibration method provided by this invention achieves integrated calibration of internal and external parameters through a unified internal and external parameter calibration scheme. This simplifies the calibration process, improves overall calibration accuracy, and maintains high calibration accuracy even in the presence of noise and partial obstruction, thereby enhancing the system's reliability in real-world industrial environments. This invention can significantly improve the automation, accuracy, and practicality of laser scanning system calibration, providing a more efficient and reliable calibration solution for applications such as industrial automation and robot navigation. Attached Figure Description
[0018] 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 some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0019] Figure 1 This is a flowchart of the method for outputting internal parameters in this invention.
[0020] Figure 2 This is a flowchart of the method for outputting external parameters in this invention.
[0021] Figure 3 This is a schematic diagram of the laser scanning system structure in this invention.
[0022] Figure 4 This is a flowchart of the internal parameter coordinate transformation in this invention.
[0023] Figure 5 This is a flowchart illustrating the external parameter calibration principle in this invention. Detailed Implementation
[0024] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.
[0025] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit the present invention or its application or use. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0026] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.
[0027] Unless otherwise specifically stated, the relative arrangement, numerical expressions, and values of the components and steps described in these embodiments do not limit the scope of the invention. It should also be understood that, for ease of description, the dimensions of the various parts shown in the drawings are not drawn to actual scale. Techniques, methods, and devices known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and devices should be considered part of the specification. In all examples shown and discussed herein, any specific values should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values. It should be noted that similar reference numerals and letters in the following figures denote similar items; therefore, once an item is defined in one figure, it need not be further discussed in subsequent figures.
[0028] This invention provides a method for self-calibrating parameters of a gimbal-based laser scanning system, comprising: in the internal parameter self-calibration stage, using inherent geometric features in the environment (such as planes) as natural calibration references, and solving for the optimal internal parameters by establishing and optimizing a geometric consistency model of point cloud data.
[0029] like Figure 1 As shown, scanning is performed in a natural scene with multiple non-parallel, non-coplanar plane features (e.g., walls, floors, and ceilings of an indoor room) to collect multi-pose data. In a preferred embodiment of the present invention, the multi-pose data is a series of original three-dimensional point cloud data. The laser scanning system to be calibrated is used to scan in a natural scene with multiple non-parallel, non-coplanar plane features, and the gimbal attitude angle (e.g., horizontal angle, pitch angle) of each set of point cloud data is recorded.
[0030] A nonlinear optimization model is established with the internal parameters to be calibrated as optimization variables, and iterative optimization is performed based on planar consistency. The internal parameters to be calibrated are the translation vectors and rotation parameters, such as Euler angles, that describe the transformation from the LiDAR coordinate system to the gimbal reference coordinate system. In a preferred embodiment of this invention, during the iterative optimization process, the nonlinear optimization model transforms the collected multi-pose data from the LiDAR coordinate system to the reference coordinate system based on the current assumed values of the internal parameters, forming a fused global point cloud. In the fused global point cloud, multiple planar feature patches are automatically identified and extracted using algorithms such as RANSAC, and the flatness of each planar patch is calculated. The flatness represents the root mean square error (RMSE) or standard deviation of all points constituting the plane to the fitted plane. The sum of the weighted flatness errors of all planar patches is used to construct a cost function to evaluate the merits of the current assumed values of the internal parameters. This cost function introduces a distance weight factor, that is, assigns higher weights to planar patches that are farther away from the scanner, in order to enhance the constraint of the optimization process on the rotation parameters.
[0031] like Figure 1 As shown, the internal parameter increment is calculated using a nonlinear optimization model. After satisfying the preset convergence condition, the optimization is terminated, and the calibration value of the internal parameter is output. In a specific implementation, as a preferred embodiment of the present invention, a nonlinear optimization algorithm, such as the Levenberg-Marquardt method, the Gauss-Newton method, or the L-BFGS-B algorithm, is used in conjunction with the cost function and its gradient to calculate the internal parameter increment, driving the cost function to converge in the minimization direction. When the iteration satisfies that the change in the cost function is less than a threshold or the maximum number of iterations is reached, the iterative optimization process is terminated, and the optimization result obtained is the calibration value of the internal parameter.
[0032] This invention innovatively utilizes natural planar features within the scanning environment as a benchmark for internal parameter calibration, transforming the calibration problem from a complex physical operation reliant on external precision measuring tools (such as total stations and theodolites) into a purely mathematical optimization problem based on the internal geometric consistency of point cloud data. This fundamentally eliminates the dependence on expensive external equipment and tedious manual measurements.
[0033] This invention employs a nonlinear optimization algorithm to iteratively solve for the optimal parameters, with the global plane "flatness" as the objective function. This method, based on massive data statistics and mathematical optimal solutions, replaces the uncertainty of manual operation with the determinism of machine calculation, thus ensuring high precision, high repeatability, and objectivity of the calibration results in principle, and eliminating human error caused by differences in personnel skills and states.
[0034] like Figure 2 As shown, based on the completed intrinsic parameter calibration, the extrinsic parameters are determined, and the precise pose of the entire scanning system (represented by its reference coordinate system) relative to the external world coordinate system is automatically determined. Its core idea is to solve for the extrinsic parameters by accurately registering the positions of a small number of cooperative targets in the scene across the two coordinate systems.
[0035] The invention deploys cooperative targets with easily identifiable geometric features and scans the area containing these targets based on calibration values of internal parameters. In a preferred embodiment, the cooperative targets are targets with easily identifiable geometric features placed in a natural scene with multiple non-parallel, non-coplanar planes. The three-dimensional coordinates of the feature points of the cooperative targets (such as building corner points or bottom surface marking corner points) in the world coordinate system are known values. The cooperative targets can be highly reflective measuring spheres, corner prisms, or reflective patches of specific shapes.
[0036] The scanned point cloud is processed to calculate the three-dimensional coordinates of the feature points in the reference coordinate system. In a preferred embodiment of the present invention, point cloud segmentation, clustering and geometric feature analysis algorithms are used to automatically extract the point cloud clusters of each cooperative target from the background point cloud, perform accurate geometric fitting on each point cloud cluster, and calculate the three-dimensional coordinates of the feature points in the reference coordinate system of the scanning system.
[0037] like Figure 4 and 5 As shown, the correspondence between the reference coordinate system and the world coordinate system is established, and the external parameter transformation matrix is solved to achieve parameter self-calibration in the scanning system. In a preferred embodiment of the invention, point cloud registration is performed by solving the absolute orientation problem. The singular value decomposition method or the iterative nearest point algorithm is used to calculate the rigid body transformation matrix from the reference coordinate system to the world coordinate system, which includes the rotation matrix and translation vector. The solved transformation matrix is the desired external parameter.
[0038] For extrinsic parameter calibration, this invention cleverly transforms the complex global pose alignment problem into a classic absolute orientation problem with a deterministic solution by deploying a small number (e.g., three or more) of easily identifiable cooperative targets (such as reflective spheres or reflective stickers) in the scene. This method balances high accuracy (utilizing the high signal-to-noise ratio of cooperative targets) and high efficiency (requiring only the scanning and identification of a small number of target points), making it an extremely efficient and robust extrinsic parameter solution strategy.
[0039] By sequentially performing the two stages of internal parameter self-calibration and external parameter auxiliary calibration, the present invention can complete the calibration of all internal and external parameters of the laser scanning system in a highly automated and precise manner.
[0040] Example like Figure 3 As shown, this invention acquires point cloud data using a corresponding laser scanning system. By deploying a small number (e.g., three or more) of easily identifiable cooperative targets (such as reflective spheres or reflective stickers) in the scene, the complex global pose alignment problem is cleverly transformed into a classic absolute orientation problem with a deterministic solution. This method balances high accuracy (utilizing the high signal-to-noise ratio of cooperative targets) and high efficiency (requiring only the scanning and identification of a small number of target points), making it an extremely efficient and robust extrinsic parameter solving strategy.
[0041] The entire automated calibration process (including internal and external parameters) takes less than 5 minutes, which is more than 90% more efficient than the traditional manual calibration method (which usually takes 2-8 hours), greatly reducing equipment downtime.
[0042] The overall positioning accuracy of the system (i.e., the root mean square error (RMSE) of point cloud registration to the world coordinate system) is better than ±30mm. It can be widely applied to any indoor or outdoor structured scene with basic planar features such as walls and ground. There is no need to build a dedicated calibration field. During the application process, there is no need to use expensive measuring instruments such as total stations. Only low-cost cooperative targets (such as industrial-grade reflective stickers with a single cost of less than 50 yuan) need to be deployed. The overall implementation cost is reduced by more than 80%.
[0043] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.
Claims
1. A method for self-calibrating parameters of a gimbal-based laser scanning system, characterized in that, The application relates to a method for calibrating a laser scanning system. Scanning in a natural scene with multiple non-parallel, non-coplanar planar features, collecting multi-pose data; Establishing a nonlinear optimization model with internal parameters to be calibrated as optimization variables, iteratively optimizing based on planar consistency; Using the nonlinear optimization model to calculate the internal parameter increment, and terminating the optimization when a preset convergence condition is met, and outputting the calibrated value of the internal parameter; Laying out a cooperative target with easily identifiable geometric features, and scanning the region containing the cooperative target based on the calibrated value of the internal parameter; Processing the scanning point cloud, and calculating the three-dimensional coordinates of the feature points in the reference coordinate system; Establishing the corresponding relationship between the reference coordinate system and the world coordinate system, and solving the external parameter transformation matrix to realize the self-calibration of the scanning system.
2. The cloud-based gimbal-based laser scanning system parameter self-calibration method of claim 1, wherein, The multi-pose data is a series of original three-dimensional point cloud data, and a laser scanning system to be calibrated is used to scan a natural scene with multiple non-parallel, non-coplanar planar features, and the pan-tilt pose angle of each group of point cloud data is recorded.
3. The cloud-invariant-based self-calibration method of laser scanning system parameters according to claim 1, characterized in that, In the iterative optimization process of the nonlinear optimization model, the multi-pose data collected is converted from the laser radar coordinate system to the reference coordinate system according to the current internal parameter assumption value, and a fused global point cloud is formed; In the fused global point cloud, multiple planar feature pieces are automatically identified and extracted, the flatness of each planar piece is calculated, the flatness represents the root mean square error (RMSE) or standard deviation of all points constituting the plane to the fitting plane, and the sum of the weighted flatness errors of all planar pieces is constructed as a cost function for evaluating the advantages and disadvantages of the current internal parameter assumption value.
4. The cloud-invariant-based self-calibration method of laser scanning system parameters according to claim 1, characterized in that, The nonlinear optimization model is used to calculate the internal parameter increment, and the calculation process specifically includes: Using a nonlinear optimization algorithm, combining the cost function and its gradient, calculating the increment of the internal parameter, driving the cost function to converge in the minimization direction, and terminating the iterative optimization process when the iteration meets the condition that the change of the cost function is less than a threshold value or the maximum iteration number is reached, and the obtained optimization result is the calibrated value of the internal parameter.
5. The cloud-invariant-based self-calibration method of laser scanning system parameters according to claim 1, characterized in that, The cooperative target is a target with easily identifiable geometric features placed in a natural scene with multiple non-parallel, non-coplanar planar features, and the three-dimensional coordinates of the feature points of the cooperative target in the world coordinate system are known values.
6. The cloud-invariant-based self-calibration method of laser scanning system parameters according to claim 1, characterized in that, The processing of the scanning point cloud specifically includes: Through point cloud segmentation, clustering and geometric feature analysis algorithm, each point cloud cluster of the cooperative target is automatically extracted from the background point cloud, and the three-dimensional coordinates of the feature points in the scanning system reference coordinate system are calculated through accurate geometric fitting of each point cloud cluster.
7. The cloud-invariant-based self-calibration method of laser scanning system parameters according to claim 1, characterized in that, The corresponding relationship between the reference coordinate system and the world coordinate system is established, and the external parameter transformation matrix is solved, and the process specifically includes: The point cloud registration is performed by solving the absolute orientation problem, the singular value decomposition method or the iterative closest point algorithm is used to calculate the rigid transformation matrix from the reference coordinate system to the world coordinate system, the rigid transformation matrix contains a rotation matrix and a translation vector, and the solved transformation matrix is the external parameter.