Laser tracker-based common-field-of-view-free camera external parameter calibration method and system

Abstract

The invention relates to a common-field-of-view-free camera external parameter calibration method and system based on a laser tracker. The method comprises the following steps: firstly, setting a laser tracker and defining a global coordinate system based on calibration of external parameters of a common-field-of-view-free camera of the laser tracker; for each single camera, if no common view exists between the single camera and the tracker, the tracker is transferred to the position with the common view, a temporary calibration plate is arranged, and the pose relation between the temporary calibration plate and the global coordinate system is calculated; when a public view exists, a calibration plate with a target seat is arranged in the public view, a laser tracker is used for measuring coordinates of a target ball, a camera is used for measuring coordinates of a conversion ball, a target ball coordinate system and a conversion ball coordinate system are respectively established and unified, and therefore the pose relation between the camera and the calibration plate and the pose relation between the calibration plate and the global coordinate system are calculated; and finally, obtaining external parameters of the common-field-of-view-free camera by combining all pose relationships. Compared with the prior art, the method has the advantages that the external parameters of the camera can be accurately and efficiently acquired when the camera system without the public view field is processed, the anti-interference capability is strong, the efficiency is high and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of camera extrinsic parameter calibration, and in particular to a laser tracker-based method and system for extrinsic parameter calibration of a camera without a common field of view. Background Art

[0002] Photogrammetry technology is widely used due to its advantages such as non-contact, high precision, and dynamics. Multi-camera systems, compared to single-camera systems, can acquire geometric information about large and complex objects through an extended field of view, thus attracting more attention. However, in practical applications, cameras in a multi-camera system often lack or have a very small common field of view, necessitating global calibration.

[0003] Global calibration is the foundation for multi-camera measurement data fusion. Its purpose is to transform the measurement data from each camera into a unified global coordinate system. The result of global calibration is the camera's extrinsic parameters. These parameters are independent of the camera's intrinsic parameters but are closely related to the spatial layout, significantly impacting the overall measurement accuracy of the system.

[0004] Currently, the global calibration methods for multi-camera systems are mainly divided into two categories: One approach involves adding additional fixed constraints, including the fixed-connected camera method and the fixed-connected calibration plate method. The fixed-connected camera method uses a fixed camera group to capture a fixed-pose calibration plate, leveraging the fixed connection to determine the pose relationship between the cameras. The fixed-connected calibration plate method uses a large calibration plate to cover the fields of view of different cameras, determining the pose relationship between the cameras based on the positional relationships of feature points on the target. While both methods are flexible and convenient, they suffer from the limitations of the small range of combined target pose variations and unstable calibration results.

[0005] Another approach relies on additional equipment to provide additional information, such as the reflector method and the auxiliary camera method. The reflector method uses mirror reflection to expand the camera's field of view and calculates the relationship between the camera and the target to determine the relative position and pose of the cameras. However, in complex environments, it is difficult to ensure that all cameras can observe the target. The auxiliary camera method uses additional cameras to provide pose information and solves the pose relationship between cameras through coordinate transformation, but this limits the flexibility of camera layout.

[0006] Current technologies such as CN104457569A and CN102506758A provide methods based on a visual measurement system, a laser tracker system, and a positioning system.

[0007] For example, patent publication CN104457569A discloses a method for visually measuring the geometric parameters of large composite panels. This method utilizes a binocular vision measurement system and a laser tracker system to achieve rapid on-site 3D reconstruction and measurement. However, this method faces challenges with coordinate system I when dealing with cameras without a common field of view, particularly when the laser tracker's field of view is obstructed. This method is unable to accurately and efficiently obtain camera extrinsic parameters and ensure the accuracy of coordinate system I and 3D data across multiple cameras.

[0008] Patent publication number CN102506758A proposes a multi-sensor flexible dynamic vision measurement system and method for the three-dimensional surface topography of an object. This system uses multiple three-dimensional optical probes and planar targets to measure the three-dimensional topography of the object being measured. This solves the problem of rigid connections between the visual sensors in a multi-vision sensor measurement system and improves dynamic measurement accuracy. However, this method suffers from operational complexity and low efficiency when calibrating external parameters for cameras without a common field of view, especially when laser trackers in a multi-camera system need to frequently relocate to accommodate the field environment.

[0009] In summary, the current problems are: when dealing with a camera system without a common field of view, especially when the field of view of the laser tracker is blocked, the problem of coordinate system one is faced, and the camera extrinsic parameters cannot be obtained accurately and efficiently. In addition, most camera extrinsic parameter calibration methods still have the problems of complex operation and low efficiency. Summary of the Invention

[0010] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and provide a method and system for calibrating the extrinsic parameters of cameras without a common field of view based on a laser tracker. The method aims to solve the problem of camera extrinsic calibration in a multi-camera system without a common field of view and when the field of view of the laser tracker is limited, improve the calibration efficiency and accuracy, and realize the effective application of the multi-camera system in complex environments.

[0011] The purpose of the present invention can be achieved by the following technical solutions: According to one aspect of the present invention, a method for calibrating extrinsic parameters of a camera without a common field of view based on a laser tracker is provided, the method comprising the following steps: S1. Arrange the laser tracker and build and define the global coordinate system based on the laser tracker measurement benchmark; S2. Select a single camera without a common field of view and determine whether the current camera and the laser tracker have a common measurement field of view. If so, jump to step S5; if not, execute step S3; S3. Move the laser tracker to a preset station so that it has a common field of view with the current measurement camera, and set up a temporary calibration plate; place the target ball on the target seat in the temporary calibration plate, use the laser tracker before and after the station transfer to measure the spatial coordinates of the target ball in the temporary calibration plate, and establish a first temporary coordinate system and a second temporary coordinate system; S4, setting the first temporary coordinate system and the second temporary coordinate system to the same coordinate system, and using the spatial coordinates to calculate the posture conversion relationship between the temporary calibration plate and the global coordinate system; S5. Place a calibration plate equipped with a target mount within the common measurement field of view of the camera and the laser tracker; place a target sphere on the target mount of the calibration plate, measure the spatial coordinates of the target sphere using the laser tracker, and establish a target sphere coordinate system; replace the target sphere with a photogrammetry-tracker conversion sphere, place it on the corresponding target mount of the calibration plate, measure the spatial coordinates of the photogrammetry-tracker conversion sphere using the current camera, and establish a conversion sphere coordinate system; S6. Set the target sphere coordinate system and the conversion sphere coordinate system to the same coordinate system, and use the spatial coordinates to calculate the pose conversion relationship between the current camera and the calibration plate and the pose conversion relationship between the calibration plate and the global coordinate system; S7, using all the calculated posture transformation relationships, calculate the current camera extrinsic parameters; S8. Repeat steps S2-S7 until the extrinsic calibration of all single cameras without a common field of view is completed.

[0012] As an optimal technical solution, both the calibration plate and the temporary calibration plate are equipped with three target seats, which are non-collinearly distributed. A permanent magnet is provided at the bottom of the target seat, and a concentric annular coded grating is engraved on the surface of the target seat. When the camera shoots the conversion ball, the grating phase information is extracted with the help of Fourier transform, and the eccentricity of the center of the conversion ball relative to the center of the target seat is further calculated. The correction amount is then calculated based on the eccentricity, thereby correcting the coordinates of the conversion ball in the camera coordinate system in real time.

[0013] As a preferred technical solution, the correction amount consists of a correction term and an integral term, wherein the correction term is obtained by nonlinear correction of the eccentricity. The specific calculation formula of the correction amount is: ; in, is the correction amount; is the eccentricity; is the proportionality coefficient; is the integration coefficient.

[0014] As a preferred technical solution, in S1, multiple reference targets are set within a preset range centered on the laser tracker. The reference targets are used to correct the origin of the global coordinate system in real time as temperature, humidity, and time change. The specific formula is: ; in, is the updated origin coordinate; is the initial origin coordinate of the global coordinate system; The reference target weight is assigned based on the stability of the target material; is the temperature sensitivity coefficient of the i-th reference target; is the temperature difference between time t and the initial time; is the coordinate of the i-th reference target in the initial global coordinate system; is the humidity sensitivity coefficient of the i-th benchmark target; is the humidity difference between time t and the initial time; is the mechanical stress release coefficient; is the attenuation factor, describing the release rate of the residual mechanical stress of the reference target; It is an error compensation term based on wavelet denoising, which achieves dynamic smoothing by decomposing high-frequency noise.

[0015] As a preferred technical solution, when determining whether the current camera and the laser tracker have a common measurement field of view in S2, a virtual light model is constructed, and the camera's intrinsic parameter matrix and the laser tracker's measurement range parameters are used to calculate the intersection of the camera imaging light and the laser tracker's measurement cone; let the camera optical center be O c , the laser tracker measurement center is O t , the spatial light direction vector corresponding to a point on the camera imaging plane is d p , the laser tracker measures the cone half angle as , then determine whether there is a public view by judging whether the following inequality holds: ; If there are multiple points on the camera imaging plane that satisfy the above inequality, it is determined that there is a common measurement field of view.

[0016] As an optimal technical solution, in S4, when calculating the pose conversion relationship between the temporary calibration plate and the global coordinate system, if there is a camera with calibrated external parameters, the coordinates of the conversion sphere in the calibrated camera coordinate system are converted to the global coordinate system to obtain the first set of pose parameters of the temporary calibration plate in the global coordinate system. Based on the measurement data of the laser tracker after the station is transferred, the second set of pose parameters of the temporary calibration plate in the global coordinate system is solved, and the deviation between the first set of pose parameters and the second set of pose parameters is calculated, which is the redundant constraint residual, and the Gauss-Helmert adjustment model is used to iteratively optimize the residual.

[0017] As a preferred technical solution, the current camera extrinsic parameters calculated in S7 include two cases: When the current camera and the laser tracker have a common measurement field of view, the pose transformation relationship includes the pose transformation relationship between the current camera and the calibration plate and the pose transformation relationship between the calibration plate and the global coordinate system; When the current camera and the laser tracker do not have a common measurement field of view, the pose transformation relationship includes the pose transformation relationship between the temporary calibration plate and the global coordinate system, the pose transformation relationship between the current camera and the calibration plate, and the pose transformation relationship between the calibration plate and the global coordinate system.

[0018] As a preferred technical solution, when the current camera and the laser tracker have a common measurement field of view, the specific formula for calculating the current camera extrinsic parameters is: ; ; ; in, For single camera and calibration plate Serial number; 、 and Respectively represent The homogeneous pose transformation matrix, The rotation matrix and The translation vector of 、 and Subscripted variables ci 、 ti and w Represent the definition symbols of each single camera, calibration plate and laser tracker, that is, the global coordinate system; 、 and Represents single camera Relative to the global coordinate system The homogeneous pose transformation matrix, rotation matrix and translation vector of .

[0019] As a preferred technical solution, when the current camera and the laser tracker do not have a common measurement field of view, the specific formula for calculating the current camera extrinsic parameters is: ; ; ; in, For single camera and calibration plate Serial number, 、 and Respectively represent The homogeneous pose transformation matrix, The rotation matrix and The translation vector of 、 and Subscripted variables cj 、 tj 、 temp 、 w and wt They represent the definition symbols of each single camera, calibration plate, temporary calibration plate, laser tracker and laser tracker after transfer, such as 、 、 Represents single camera Relative to the global coordinate system The homogeneous pose transformation matrix, rotation matrix and translation vector of .

[0020] According to another aspect of the present invention, a laser tracker-based extrinsic calibration system for cameras without a common field of view is provided, the system comprising a laser tracker, a camera group to be calibrated, a calibration plate group, a ball group, and a data processing unit; The laser tracker is used as a global coordinate system reference device; The camera group to be calibrated consists of multiple single cameras without a common field of view; The calibration plate set includes a calibration plate and a temporary calibration plate. The calibration plate is used to associate the position of the camera and laser tracker in the public field of view, and the temporary calibration plate is used to connect the coordinate systems before and after the laser tracker is transferred. The ball set includes target balls and transition balls; The data processing unit is equipped with a calibration algorithm and an error compensation model. This unit is used to complete the pose conversion relationship solution and camera extrinsic parameter calculation, and output the calibration results.

[0021] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention uses a laser tracker to define a global coordinate system, incorporating a calibration plate as a medium. Leveraging the laser tracker's high-precision, non-contact, and wide-range measurement advantages, this method achieves extrinsic parameter calibration between cameras without a shared field of view, without relying on a common field of view, adding auxiliary cameras, or employing low-precision calibration methods. Furthermore, when the laser tracker's field of view is obstructed, a mechanism using a transfer station and a temporary calibration plate overcomes field of view obstruction and ensures uninterrupted calibration in obstructed scenarios. Multi-camera calibration can be completed simply by looping this method, and its standardized process significantly improves efficiency. When processing camera systems without a shared field of view, this method can accurately and efficiently obtain camera extrinsic parameters, exhibits strong anti-interference capabilities, and exhibits high efficiency.

[0022] 2. In the present invention, both the calibration plate and the temporary calibration plate are equipped with three target seats. A permanent magnet is provided at the bottom of the target seat, which improves the placement efficiency of the target ball and the conversion ball. A concentric annular coded grating is engraved on the surface of the target seat. When the camera shoots the conversion ball, the grating phase information is extracted with the help of Fourier transform, and the eccentricity of the center of the conversion ball relative to the center of the target seat is further calculated. The correction amount is then calculated based on the eccentricity, thereby correcting the coordinates of the conversion ball in the camera coordinate system in real time. This provides higher-precision basic data for the subsequent solution of the posture relationship, so that the overall camera extrinsic calibration error is small.

[0023] 3. In the present invention, multiple reference targets are set within a preset range centered on the laser tracker. The reference targets are used to establish a multi-parameter coupling model, taking into account the effects of temperature, humidity, and stress release, and incorporating wavelet noise reduction. The origin of the global coordinate system is corrected in real time as temperature, humidity, and time change, ensuring the long-term consistency of multi-camera extrinsic parameter calibration and solving the problem of "large deviation in multi-batch calibration" caused by reference drift in traditional methods.

[0024] 4. In the present invention, when calculating the pose conversion relationship between the temporary calibration plate and the global coordinate system, if there is a camera with calibrated extrinsic parameters, the coordinates of the conversion sphere in the calibrated camera coordinate system are converted to the global coordinate system to obtain a first set of pose parameters of the temporary calibration plate in the global coordinate system. Based on the measurement data of the laser tracker after the transfer station, the second set of pose parameters of the temporary calibration plate in the global coordinate system is solved, and the deviation between the first set of pose parameters and the second set of pose parameters is calculated, which is the redundant constraint residual. The Gauss-Helmert adjustment model is used to iteratively optimize the residual. The redundant constraint residual is constructed through the dual data sources of the calibrated camera and the tracker after the transfer station. The Gauss-Helmert adjustment is then used for iterative optimization, so that the pose solution accuracy of the temporary calibration plate is improved, and high-precision coordinate system connection is provided for the transfer calibration of cameras without a common field of view, thereby ensuring the global consistency of multi-station calibration.

[0025] 5. In the present invention, a virtual ray model and a cone intersection judgment method are used to quantitatively determine whether the current camera and the laser tracker have a common measurement field of view, thereby improving the degree of automation and reducing misjudgment, so as to realize the automatic distinction between the two working conditions of "solving camera extrinsics with a common measurement field of view and solving camera extrinsics without a common measurement field of view". The method also clarifies the corresponding pose relationships required for the two working conditions, as well as the specific methods for solving camera extrinsics based on the pose relationships corresponding to the two working conditions. This makes the calibration method adaptable to all scenarios, and the working conditions can be automatically identified through the common pose relationship chain, and the corresponding settlement logic can be called, which is suitable for complex scenarios, improving the automation rate of the overall calibration process and reducing errors caused by manual intervention.

[0026] 6. The system in this invention integrates a laser tracker, a camera assembly to be calibrated, a calibration plate assembly, a ball assembly, and a data processing unit, automating the entire process of measurement, solution, calibration, and output. Compared with traditional solutions, this system reduces manual data transmission and processing time, shortens the calibration cycle, and improves calibration efficiency. Furthermore, the system's built-in error compensation model enhances the accuracy of calibration results, making it suitable for large-scale camera extrinsic calibration. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 Schematic diagram of the steps of a laser tracker-based extrinsic parameter calibration method for a camera without a common field of view in the present invention; Figure 2 This is a flow chart of the extrinsic parameter calibration of a camera without a common field of view based on a laser tracker in the present invention; Figure 3 Schematic diagram of extrinsic parameter calibration of a camera without a common field of view based on a laser tracker in the present invention; Figure 4 Schematic diagram of the laser tracker transfer station to achieve camera extrinsic calibration in the present invention. DETAILED DESCRIPTION

[0028] 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 part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0029] Photogrammetry technology is widely used due to its advantages such as non-contact, high precision, and dynamics. Multi-camera systems, compared to single-camera systems, can acquire geometric information about large and complex objects through an extended field of view, thus attracting more attention. However, in practical applications, cameras in a multi-camera system often lack or have a very small common field of view, necessitating global calibration.

[0030] Therefore, this proposal proposes a laser tracker-based extrinsic parameter calibration method and system for cameras without a common field of view, which flexibly and accurately realizes the global calibration of multi-camera systems, reduces the hardware cost of auxiliary cameras, and helps multi-camera systems achieve full-field, high-precision, and dynamic monitoring tasks. It has important theoretical research significance and broad application prospects.

[0031] Example 1 In this embodiment, a method for calibrating the external parameters of a camera without a common field of view based on a laser tracker is adopted. The basic principle of this method for calibrating the external parameters of a camera without a common field of view based on a laser tracker is as follows: First, consider the cost, site environment and other factors to select the number of single cameras and their positions. If the multi-camera system can use the common field of view between the cameras to unify the coordinate systems of the multiple cameras into the global coordinate system, then the internal and external parameters of the cameras can be calibrated directly using the reference ruler. Otherwise, place the laser tracker at the appropriate position. , with Single camera: , and a calibration plate is placed in sequence The measurement field of view of the camera and laser tracker is: The calibration plate consists of three target bases and is equipped with a target ball and a photogrammetry-tracker conversion ball. If each calibration plate is within the measurement field of view of the camera and tracker, the target ball is placed on the target base of the calibration plate, and the spatial coordinates of the target ball are measured by the tracker to establish Target sphere coordinate system ( ), remove the target ball, place the photogrammetry-tracker conversion ball on the target seat of the calibration plate, use the camera to measure the spatial coordinates of the conversion ball, and establish Transformed spherical coordinate system ( ), the corresponding target sphere coordinate system and the transformation sphere coordinate system are the same coordinate system, then the single camera can be solved Coordinate system to global coordinate system The rotation matrix and translation vector are: ; ; ; in, For single camera and calibration plate Serial number, 、 and Respectively represent The homogeneous pose transformation matrix, The rotation matrix and The translation vector of 、 and The subscripted variables represent the definition symbols of each single camera, calibration plate and laser tracker, such as 、 、 Represents single camera Relative to the global coordinate system The homogeneous pose transformation matrix, rotation matrix and translation vector of .

[0032] Then, if the laser tracker is used to calibrate the external parameters of the camera without a public field of view, and the field of view of the tracker is blocked by obstacles, the tracker needs to be transferred to the selected transfer station, and then the external parameters of the camera are calibrated. First, the laser tracker is moved from the initial position to the selected transfer station and the temporary calibration plate required for the transfer station is arranged. , place the target ball on the target seat in the calibration plate and use trackers at different positions to measure the spatial coordinates of the target ball and establish the target ball coordinate system and , these two coordinate systems are the same coordinate system, then the single camera can be solved when the tracker's field of view is blocked To the global coordinate system The rotation matrix and translation vector are: ; ; ; in, For single camera and calibration plate Serial number, 、 and Respectively represent The homogeneous pose transformation matrix, The rotation matrix and The translation vector of 、 and The subscripted variables represent the definition symbols of each single camera, calibration plate and laser tracker, such as 、 、 Represents single camera Relative to the global coordinate system The homogeneous pose transformation matrix, rotation matrix and translation vector of .

[0033] After completing the external parameter calibration between cameras without a common field of view, the external parameters of the cameras with a common field of view and the intrinsic parameters of each camera are calibrated using a reference ruler, so that the single camera in the multi-camera system can be ( ) measurement results are unified into the global coordinate system middle.

[0034] In this embodiment, this method is applied to calibrate the multi-camera extrinsic parameters of a large wind turbine tower. The specific implementation process is as follows: The specific scenario requirements are as follows: 6 cameras need to be deployed around the wind turbine tower to monitor tower deformation. Some cameras have no public field of view due to obstruction, and laser trackers are required to achieve external parameter calibration.

[0035] First, build and initialize the global coordinate system according to this method. The steps are as follows Figure 1 As shown, specifically including: S1. Arrange the laser tracker and build and define the global coordinate system based on the laser tracker measurement benchmark; S2. Select a single camera without a common field of view and determine whether the current camera and the laser tracker have a common measurement field of view. If so, jump to step S5; if not, execute step S3; S3. Move the laser tracker to a preset station so that it has a common field of view with the current measurement camera, and set up a temporary calibration plate; place the target ball on the target seat in the temporary calibration plate, use the laser tracker before and after the station transfer to measure the spatial coordinates of the target ball in the temporary calibration plate, and establish a first temporary coordinate system and a second temporary coordinate system; S4, setting the first temporary coordinate system and the second temporary coordinate system to the same coordinate system, and using the spatial coordinates to calculate the posture conversion relationship between the temporary calibration plate and the global coordinate system; S5. Place a calibration plate equipped with a target mount within the common measurement field of view of the camera and the laser tracker; place a target sphere on the target mount of the calibration plate, measure the spatial coordinates of the target sphere using the laser tracker, and establish a target sphere coordinate system; replace the target sphere with a photogrammetry-tracker conversion sphere, place it on the corresponding target mount of the calibration plate, measure the spatial coordinates of the photogrammetry-tracker conversion sphere using the current camera, and establish a conversion sphere coordinate system; S6. Set the target sphere coordinate system and the conversion sphere coordinate system to the same coordinate system, and use the spatial coordinates to calculate the pose conversion relationship between the current camera and the calibration plate and the pose conversion relationship between the calibration plate and the global coordinate system; S7, using all the calculated posture transformation relationships, calculate the current camera extrinsic parameters; S8. Repeat steps S2-S7 until the extrinsic calibration of all single cameras without a common field of view is completed.

[0036] First, arrange the equipment: A laser tracker is installed at the base of the tower, and the global coordinate system G is defined with the tracker's measurement center as the reference. Three ceramic reference targets and two metal reference targets are placed within a 3-meter radius around the tracker to correct for global origin drift in real time. Using the reference targets, the origin of the global coordinate system is corrected in real time as temperature, humidity, and time change. Then perform public field of view judgment and camera grouping: Six cameras are arranged at different heights of the tower (20m, 40m, and 60m). Some cameras have no common field of view due to the obstruction of the tower flange. A virtual light model is constructed to calculate the intersection of the camera imaging light and the tracker measurement cone (half angle). When judging whether the current camera and the laser tracker have a common measurement field of view, the virtual light model is constructed, and the camera's intrinsic parameter matrix and the laser tracker's measurement range parameters are used to calculate the intersection of the camera imaging light and the laser tracker measurement cone; let the camera optical center be O c , the laser tracker measurement center is O t , the spatial light direction vector corresponding to a point on the camera imaging plane is d p , the laser tracker measures the cone half angle as , then determine whether there is a public view by judging whether the following inequality holds: ; If there are multiple points on the camera imaging plane that satisfy the above inequality, it is determined that there is a common measurement field of view.

[0037] Then perform camera calibration without common field of view: Both the calibration plate and the temporary calibration plate are equipped with three target holders, which are non-collinearly distributed. A permanent magnet is provided at the bottom of the target holder, and a concentric annular coded grating is engraved on the surface of the target holder. When the camera shoots the conversion ball, the grating phase information is extracted by Fourier transform, and the eccentricity of the conversion ball center relative to the target holder center is further calculated. The correction value is then calculated based on the eccentricity, thereby correcting the coordinates of the conversion ball in the camera coordinate system in real time. The correction value consists of a correction term and an integral term, among which the correction term is obtained by nonlinear correction of the eccentricity. The specific calculation formula of the correction value is: ; in, is the correction amount; is the eccentricity; is the proportionality coefficient; is the integration coefficient.

[0038] During the station transfer, the laser tracker was moved to the 20-meter platform on the tower (pre-set station location) and a temporary calibration plate was placed. The target sphere coordinates were measured with the tracker before the transfer to establish a first temporary coordinate system. After the transfer, the target sphere coordinates were measured to establish a second temporary coordinate system. The first and second temporary coordinate systems were unified using a Gauss-Helmert adjustment (residual ≤ 0.5 μm), and the position of the laser tracker relative to the global coordinate system was calculated after the station transfer.

[0039] For cameras without a public field of view, a calibration plate is placed in the public field of view. The target sphere is measured to establish the target sphere coordinate system, and the conversion sphere is measured to establish the conversion sphere coordinate system. The eccentricity is corrected by the coded grating, and the pose of the camera without a public field of view and the target sphere coordinate system is calculated. The pose of the target sphere coordinate system and the global coordinate system are combined to finally complete the calibration of the current camera without a public field of view.

[0040] In S4, when calculating the pose conversion relationship between the temporary calibration plate and the global coordinate system, if there is a camera with calibrated extrinsic parameters, the coordinates of the conversion sphere in the calibrated camera coordinate system are converted to the global coordinate system to obtain the first set of pose parameters of the temporary calibration plate in the global coordinate system. Based on the measurement data of the laser tracker after the station is transferred, the second set of pose parameters of the temporary calibration plate in the global coordinate system is solved, and the deviation between the first set of pose parameters and the second set of pose parameters is calculated, which is the redundant constraint residual, and the Gauss-Helmert adjustment model is used to iteratively optimize the residual.

[0041] There are two cases in which the current camera extrinsics are calculated in S7: When the current camera and laser tracker have a common measurement field of view, the calibration diagram is as follows: Figure 3 As shown, the pose conversion relationship at this time includes the pose conversion relationship between the current camera and the calibration plate and the pose conversion relationship between the calibration plate and the global coordinate system; When the current camera and laser tracker do not have a common measurement field of view, the calibration diagram is as follows: Figure 4 As shown, at this time, the pose conversion relationship includes the pose conversion relationship between the temporary calibration plate and the global coordinate system, the pose conversion relationship between the current camera and the calibration plate, and the pose conversion relationship between the calibration plate and the global coordinate system.

[0042] When the current camera and the laser tracker have a common measurement field of view, the specific formula for calculating the current camera extrinsic parameters is: ; ; ; in, For single camera and calibration plate Serial number; 、 and Respectively represent The homogeneous pose transformation matrix, The rotation matrix and The translation vector of 、 and Subscripted variables ci 、 ti and wRepresent the definition symbols of each single camera, calibration plate and laser tracker, that is, the global coordinate system; 、 and Represents single camera Relative to the global coordinate system The homogeneous pose transformation matrix, rotation matrix and translation vector of .

[0043] When the current camera and the laser tracker do not have a common measurement field of view, the specific formula for calculating the current camera extrinsic parameters is: ; ; ; in, For single camera and calibration plate Serial number, 、 and Respectively represent The homogeneous pose transformation matrix, The rotation matrix and The translation vector of 、 and Subscripted variables cj 、 tj 、 temp 、 w and wt They represent the definition symbols of each single camera, calibration plate, temporary calibration plate, laser tracker and laser tracker after transfer, such as 、 、 Represents single camera Relative to the global coordinate system The homogeneous pose transformation matrix, rotation matrix and translation vector of .

[0044] Finally, perform full system calibration and verification: For cameras with a public field of view, the camera's internal and external parameters are calibrated directly using a reference ruler. For cameras without a public field of view, after completing the extrinsic parameter calibration based on the above steps, the camera's intrinsic parameters are calibrated using a reference ruler, completing the full system calibration. Finally, a global verification is performed, using the reference ruler to measure the tower's feature points and verify the consistency of the multi-camera extrinsic parameters (coordinate deviation ≤ 2μm), completing the camera parameter calibration.

[0045] In summary, this method leverages the high-precision, non-contact, and large-scale measurement advantages of laser trackers, eliminating the need for a common field of view, additional auxiliary cameras, or low-precision calibration methods to achieve extrinsic calibration between cameras without a common field of view. Laser trackers can accurately acquire the spatial position information of measurement points over a large range. Through a specific calibration process and algorithm, the coordinate systems of different cameras are linked to the coordinate system of the laser tracker, thus avoiding the traditional methods' reliance on a common field of view and the need for additional hardware or low-precision calibration methods. This allows for flexible and high-precision extrinsic calibration, thereby achieving global calibration of multi-camera systems, reducing the cost of hardware such as cameras, and helping multi-camera systems achieve full-field, high-precision, dynamic monitoring tasks.

[0046] Example 2 In this embodiment, a laser tracker-based non-common field camera extrinsic parameter calibration system is adopted, and the system includes a laser tracker, a camera group to be calibrated, a calibration plate group, a ball group, and a data processing unit; The laser tracker is used as the global coordinate system reference device; the camera group to be calibrated consists of multiple single cameras without a common field of view; the calibration plate group includes a calibration plate and a temporary calibration plate, among which the calibration plate is used to associate the posture of the camera and the laser tracker in the common field of view, and the temporary calibration plate is used to connect the coordinate systems before and after the laser tracker is transferred; the ball group includes a target ball and a conversion ball; the data processing unit is equipped with a calibration algorithm to complete the solution of the posture transformation relationship and the calculation of the camera external parameters, and output the calibration results.

[0047] The specific process of using this system to calibrate the parameters of a multi-camera system for monitoring the synchronous lifting of large structures is as follows: Figure 2 As shown, the details are as follows: (1) Use finite element analysis and other technical means to conduct dynamic stress analysis on large structures to simulate their stress conditions under actual operating conditions. Accurately identify stress concentration areas, areas with drastic stress changes, and areas that are subjected to high stress for a long time, and identify these areas as key areas to be monitored; (2) Considering the cost, on-site construction environment, and the distribution of the area to be monitored, select the appropriate number of cameras and determine the camera locations; (3) Find the cameras that have no common field of view or a small common field of view with other cameras in the multi-camera system, and select the position of the laser tracker based on the position of these cameras. Then, place the calibration plate within the measurement range of these cameras and the laser tracker; (4) Using the camera and laser tracker at the selected position, measure the spatial coordinates of the photogrammetry-tracker conversion ball and the target ball in the calibration plate respectively; (5) Based on the acquired spatial coordinate data, the rotation matrix and translation vector from the camera coordinate system to the global coordinate system are solved to determine the position and posture of the camera in the global coordinate system; (6) If there is an obstacle blocking the field of view of the laser tracker, it is necessary to move the tracker and set up a common calibration plate to perform the tracker transfer operation, and at the same time calibrate the rotation matrix and translation vector between the trackers; (7) Repeat the steps of spatial coordinate measurement (similar to step 4) and camera position and attitude calculation (similar to step 5) to further calculate the rotation matrix and translation vector from the camera coordinate system to the global coordinate system to improve accuracy or adapt to environmental changes; (8) After completing the external parameter calibration of the camera without a common field of view in the multi-camera system, use the reference ruler to calibrate the external parameters of other cameras and the intrinsic parameters of all cameras, and finally achieve the global calibration of the multi-camera system; (9) The multi-camera system performs full-field, high-precision, and dynamic measurement of the dynamic data of all areas to be measured in the global coordinate system during the synchronous lifting of large structures.

[0048] In this practical example, the calibration system is applied to calibrate the parameters of a multi-camera system in a smart factory. The specific implementation process is as follows: In this embodiment, 12 cameras are deployed on the smart factory production line to monitor robot motion. Some cameras have no common field of view due to obstruction by production line equipment, necessitating real-time online calibration of external parameters. A laser tracker is deployed at the center of the production line, serving as the global coordinate system reference and enabling rapid station rotation. The camera group to be calibrated consists of 12 industrial cameras distributed around the production line, some of which have no common field of view. The sphere group consists of a ceramic target sphere and a conversion sphere with dual-band surface features. The data processing unit uses an industrial computer equipped with a calibration algorithm that supports real-time solution.

[0049] The system workflow is as follows: First, the global coordinate system is initialized. Using the tracker and five reference targets, the global origin is corrected in real time to ensure long-term stability. Next, a dynamic judgment of the shared field of view is performed, calculating the intersection of the camera and tracker's fields of view in real time and automatically grouping cameras with and without a shared field of view. For cameras with a shared field of view, extrinsic parameters are calculated in real time using a calibration plate. For cameras without a shared field of view, the tracker is triggered to relocate, the coordinate system is connected using a temporary calibration plate, and extrinsic parameters are calculated and updated synchronously. Finally, accuracy verification and feedback are provided using feature points at the end of the production line robot to verify the consistency of extrinsic parameters across multiple cameras, achieving closed-loop calibration.

[0050] This system has strong dynamic adaptability, good real-time performance and strong robustness. It supports "field of view changes" caused by the movement of production line equipment (such as robot handling), automatically re-evaluates the public field of view and triggers calibration; and the posture solution delay of the data processing unit is ≤20ms, which meets the needs of online monitoring. Integrating the coded grating eccentricity correction and Gauss-Helmert adjustment, the calibration accuracy can still be maintained at ≤5μm in the vibration environment of the production line.

[0051] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and such modifications or substitutions are intended to be within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be subject to the scope of protection of the claims.

Claims

1. A method for calibrating external parameters of a camera without a common field of view based on a laser tracker, characterized in that: The method steps include: S1. Arrange the laser tracker and build and define the global coordinate system based on the laser tracker measurement benchmark; S2. Select a single camera without a common field of view and determine whether the current camera and the laser tracker have a common measurement field of view. If so, jump to step S5; if not, execute step S3; S3. Move the laser tracker to a preset station so that it has a common field of view with the current measurement camera, and set up a temporary calibration plate; place the target ball on the target seat in the temporary calibration plate, use the laser tracker before and after the station transfer to measure the spatial coordinates of the target ball in the temporary calibration plate, and establish a first temporary coordinate system and a second temporary coordinate system; S4, setting the first temporary coordinate system and the second temporary coordinate system to the same coordinate system, and using the spatial coordinates to calculate the posture conversion relationship between the temporary calibration plate and the global coordinate system; S5. Place a calibration plate equipped with a target mount within the common measurement field of view of the camera and the laser tracker; place a target sphere on the target mount of the calibration plate, measure the spatial coordinates of the target sphere using the laser tracker, and establish a target sphere coordinate system; replace the target sphere with a photogrammetry-tracker conversion sphere, place it on the corresponding target mount of the calibration plate, measure the spatial coordinates of the photogrammetry-tracker conversion sphere using the current camera, and establish a conversion sphere coordinate system; S6. Set the target sphere coordinate system and the conversion sphere coordinate system to the same coordinate system, and use the spatial coordinates to calculate the pose conversion relationship between the current camera and the calibration plate and the pose conversion relationship between the calibration plate and the global coordinate system; S7, using all the calculated posture transformation relationships, calculate the current camera extrinsic parameters; S8. Repeat steps S2-S7 until the extrinsic calibration of all single cameras without a common field of view is completed.

2. The method for calibrating external parameters of a camera without a common field of view based on a laser tracker according to claim 1, wherein: Both the calibration plate and the temporary calibration plate are equipped with three target seats, which are distributed non-collinearly. A permanent magnet is provided at the bottom of the target seat, and a concentric annular coded grating is engraved on the surface of the target seat. When the camera shoots the conversion ball, the grating phase information is extracted with the help of Fourier transform, and the eccentricity of the center of the conversion ball relative to the center of the target seat is further calculated. The correction amount is then calculated based on the eccentricity, thereby correcting the coordinates of the conversion ball in the camera coordinate system in real time.

3. The method for calibrating external parameters of a camera without a common field of view based on a laser tracker according to claim 2, wherein: The correction amount is composed of a correction term and an integral term, wherein the correction term is obtained by nonlinear correction of the eccentricity. The specific calculation formula of the correction amount is: ； in, is the correction amount; is the eccentricity; is the proportionality coefficient; is the integration coefficient.

4. The method for calibrating external parameters of a camera without a common field of view based on a laser tracker according to claim 1, wherein: In S1, multiple reference targets are set within a preset range centered on the laser tracker. The reference targets are used to correct the origin of the global coordinate system in real time as temperature, humidity, and time change. The specific formula is: ； in, is the updated origin coordinate; is the initial origin coordinate of the global coordinate system; The reference target weight is assigned based on the stability of the target material; is the temperature sensitivity coefficient of the i-th reference target; is the temperature difference between time t and the initial time; is the coordinate of the i-th reference target in the initial global coordinate system; is the humidity sensitivity coefficient of the i-th benchmark target; is the humidity difference between time t and the initial time; is the mechanical stress release coefficient; is the attenuation factor, describing the release rate of the residual mechanical stress of the reference target; It is an error compensation term based on wavelet denoising, which achieves dynamic smoothing by decomposing high-frequency noise.

5. The method for calibrating external parameters of a camera without a common field of view based on a laser tracker according to claim 1, wherein: When determining whether the current camera and the laser tracker have a common measurement field of view in S2, a virtual light model is constructed, and the intersection of the camera imaging light and the laser tracker measurement cone is calculated using the camera's intrinsic parameter matrix and the laser tracker's measurement range parameters; let the camera optical center be O c , the laser tracker measurement center is O t , the spatial light direction vector corresponding to a point on the camera imaging plane is d p , the laser tracker measures the cone half angle as , then determine whether there is a public view by judging whether the following inequality holds: ； If there are multiple points on the camera imaging plane that satisfy the above inequality, it is determined that there is a common measurement field of view.

6. The method for calibrating external parameters of a camera without a common field of view based on a laser tracker according to claim 1, wherein: In the S4, when calculating the pose conversion relationship between the temporary calibration plate and the global coordinate system, if there is a camera with calibrated extrinsic parameters, the coordinates of the conversion sphere in the calibrated camera coordinate system are converted to the global coordinate system to obtain a first set of pose parameters of the temporary calibration plate in the global coordinate system. Based on the measurement data of the laser tracker after the station is transferred, the second set of pose parameters of the temporary calibration plate in the global coordinate system is solved, and the deviation between the first set of pose parameters and the second set of pose parameters is calculated, which is the redundant constraint residual, and the Gauss-Helmert adjustment model is used to iteratively optimize the residual.

7. The method for calibrating external parameters of a camera without a common field of view based on a laser tracker according to claim 1, wherein: The calculation of the current camera extrinsic parameters in S7 includes two cases: When the current camera and the laser tracker have a common measurement field of view, the pose conversion relationship includes the pose conversion relationship between the current camera and the calibration plate and the pose conversion relationship between the calibration plate and the global coordinate system; When the current camera and the laser tracker do not have a common measurement field of view, the pose conversion relationship includes the pose conversion relationship between the temporary calibration plate and the global coordinate system, the pose conversion relationship between the current camera and the calibration plate, and the pose conversion relationship between the calibration plate and the global coordinate system.

8. The method for calibrating external parameters of a camera without a common field of view based on a laser tracker according to claim 7, wherein: When the current camera and the laser tracker have a common measurement field of view, the specific formula for calculating the current camera extrinsic parameters is: ； ； ； in, For single camera and calibration plate Serial number; 、 and Respectively represent The homogeneous pose transformation matrix, The rotation matrix and The translation vector of 、 and Subscripted variables ci 、 ti and w Represent the definition symbols of each single camera, calibration plate and laser tracker, that is, the global coordinate system; 、 and Represents single camera Relative to the global coordinate system The homogeneous pose transformation matrix, rotation matrix and translation vector of .

9. The method for calibrating external parameters of a camera without a common field of view based on a laser tracker according to claim 7, wherein: When the current camera and the laser tracker do not have a common measurement field of view, the specific formula for calculating the current camera extrinsic parameters is: ； ； ； in, For single camera and calibration plate Serial number, 、 and Respectively represent The homogeneous pose transformation matrix, The rotation matrix and The translation vector of 、 and Subscripted variables cj 、 tj 、 temp 、 w and wt They represent the definition symbols of each single camera, calibration plate, temporary calibration plate, laser tracker and laser tracker after transfer, such as 、 、 Represents single camera Relative to the global coordinate system The homogeneous pose transformation matrix, rotation matrix and translation vector of .

10. A laser tracker-based extrinsic calibration system for cameras without a common field of view, characterized in that: The system applies a laser tracker-based non-common field-of-view camera extrinsic parameter calibration method as described in any one of claims 1 to 9, and the system includes a laser tracker, a camera group to be calibrated, a calibration plate group, a ball group, and a data processing unit; The laser tracker is used as a global coordinate system reference device; The camera group to be calibrated consists of multiple single cameras without a common field of view; The calibration plate set includes a calibration plate and a temporary calibration plate, wherein the calibration plate is used to associate the position of the camera and the laser tracker in a common field of view, and the temporary calibration plate is used to connect the coordinate systems before and after the laser tracker is transferred; The ball set includes a target ball and a conversion ball; The data processing unit is equipped with a calibration algorithm and an error compensation model. The unit is used to complete the posture conversion relationship solution and camera extrinsic parameter calculation, and output the calibration result.

Images