A system and calibration method for simultaneous automated calibration of binocular cameras.

Abstract

This invention relates to the field of computer vision and image measurement technology, and discloses a system and method for synchronous automated calibration of binocular cameras. The system includes: a base; an inner ring flange fixed on the base for fixing M calibration plates; an outer ring flange disposed on the base and rotating around the tangent of the inner ring flange radius for arranging and driving N sets of binocular cameras; a cross slide assembly disposed on the inner ring flange for moving the calibration plates to different positions; a camera bracket fixed on the outer ring flange for fixing the binocular cameras; and a control and data processing unit connected to the outer ring flange, the cross slide assembly, and the binocular cameras for controlling the rotation of the outer ring flange, the movement of the calibration plates, and completing the calibration after the binocular cameras acquire images. This invention can automatically acquire multiple sets of images with different poses within one rotation cycle, achieving fully automated calibration of multiple sets of binocular cameras and improving production efficiency.

Description

Technical Field

[0001] This invention relates to the field of computer vision and image measurement technology, and in particular to a system and method for synchronous automated calibration of binocular cameras. Background Technology

[0002] In the wave of industrialization, binocular calibration is rapidly penetrating into the capillaries of discrete manufacturing and smart terminals along three main lines: automation, online operation, and miniaturization. For example, in industrial robots, binocular calibration provides millimeter-level three-dimensional grasping coordinates for collaborative arms, synchronizing production line cycle time with calibration cycle time; in autonomous driving, in-vehicle binocular systems compensate for temperature drift and vibration through online self-calibration; in security monitoring, RDID drone detection systems utilize binocular calibration to achieve real-time three-dimensional positioning of drones and operators; and in consumer electronics, AR / VR headsets use miniaturized binocular modules, which undergo 0.1 mm-level spatial tracking calibration at the factory through a fully automated calibration station.

[0003] Binocular calibration is a prerequisite for acquiring 3D geometric information in stereo vision, and its accuracy directly determines the reliability of subsequent ranging, reconstruction, or positioning. Traditional methods rely on manual handheld or robot-assisted placement of planar checkerboard targets. After acquiring images in 10-20 different poses, intrinsic and extrinsic parameters are solved using Zhang Zhengyou's calibration method and stereo calibration procedures. However, this paradigm exposes three prominent problems when facing mass production: First, manual intervention leads to uncontrollable pace, and the acquisition process is time-consuming and labor-intensive, making it difficult to achieve online calibration on mass production lines; Second, factors such as changes in lighting, target contamination, and motion blur cause inconsistent image quality, resulting in reprojection error fluctuations, making it difficult to cover the entire field of view of the binocular camera and different depths and tilt angles in a limited space at once, leading to large extrapolation errors in calibration parameters; Third, offline calibration results lack traceability. Once the camera is affected by temperature drift or mechanical vibration and undergoes slight displacement, rework and retesting are necessary, resulting in high after-sales costs. Furthermore, existing methods are designed for visible light cameras and do not adequately support multimodal systems such as infrared, fisheye, and RGB-D. Distortion models and baseline compensation strategies need to be customized for each model, which further limits production flexibility.

[0004] Therefore, researching a fully automated, high-precision, and scalable dual-target positioning technology has become an urgent need to promote the mass production of stereo vision products. Summary of the Invention

[0005] This invention provides a system and method for synchronous automated calibration of binocular cameras, capable of automatically acquiring multiple sets of images in different poses within a single rotation cycle, achieving fully automated calibration of multiple binocular cameras. It eliminates the efficiency bottlenecks and human errors caused by manual placement of calibration boards and manual triggering of shooting. It fundamentally suppresses reprojection deviations caused by inconsistent image set quality, meeting the efficiency requirements of mass production lines, while significantly improving calibration accuracy and efficiency.

[0006] This invention provides a system for automated synchronous calibration of binocular cameras, comprising:

[0007] Base;

[0008] An inner ring flange, which is fixed on the base, is used to fix M calibration plates;

[0009] An outer ring flange is disposed on the base and rotates around the inner ring flange. The outer ring flange is used to arrange and drive N sets of binocular cameras to rotate around the calibration plate.

[0010] A cross slide assembly is disposed on the inner ring flange, and the calibration plate is fixed thereon. The cross slide assembly is used to drive the calibration plate to change position in the front-back and left-right directions, so as to adjust the optical axis distance between the binocular camera and the calibration plate.

[0011] A camera bracket, which is fixed to the outer ring flange, is used to place and fix the binocular camera;

[0012] A control and data processing unit is fixed inside the base and connected to the outer ring flange, the cross slide assembly, and the binocular camera. The control and data processing unit is used to control the rotation of the outer ring flange, control the position movement of the calibration plate, and control the binocular camera to acquire images. It also calculates the intrinsic and extrinsic parameters of the binocular camera based on the acquired image set, thereby realizing the automatic calibration of the binocular camera.

[0013] Furthermore, the inner ring flange is circular in shape, the outer ring flange is annular in shape, and the inner ring of the outer ring flange surrounds the outer circumference of the inner ring flange;

[0014] The calibration board is a checkerboard calibration board or a center-shaped calibration board, and the number of the calibration board and the binocular camera is greater than or equal to 2.

[0015] Furthermore, the calibration plate and the cross slide assembly are connected by a calibration plate bracket. Multiple cross slide assemblies are provided on the inner ring flange. Each cross slide assembly corresponds to a calibration plate bracket and a calibration plate. The heights of the multiple calibration plate brackets are different, and the height difference of the multiple calibration plate brackets is Δh. Δh satisfies Δh=tanθ*D, where D is the closest distance from the calibration plate to the optical center of the binocular camera, θ is the vertical field of view angle of the binocular camera, and the pitch angle can be independently adjusted to simulate different field of view tilt angles.

[0016] Furthermore, the process of controlling the binocular camera to acquire images in the control and data processing unit is as follows: when the binocular camera detects the calibration board, it controls the binocular camera to synchronously acquire images of the calibration board and records the current calibration board area S. n With the center point coordinates O n Simultaneously, the cross slide assembly is controlled to begin a slow, uniform backward movement. When a new calibration plate area S is detected... n+1 With S n The difference is greater than the corresponding area threshold T s Or the new center point coordinates O n+1 With O n The difference is greater than the corresponding coordinate point deviation threshold O s If the calibration board is considered to have changed its spatial pose, the current image of the calibration board will be acquired; otherwise, the image acquisition signal will not be triggered.

[0017] Furthermore, in the control and data processing unit, after the binocular camera starts acquiring calibration images, the outer ring flange is controlled to rotate at a constant speed within one rotation cycle, while the binocular camera performs real-time detection of the calibration plate; after the outer ring flange completes one rotation cycle, the binocular camera is controlled to stop real-time detection and acquisition of images.

[0018] Furthermore, in the control and data processing unit, when the calibration board is a checkerboard calibration board, the Harris corner detection or Shi-Tomasi algorithm is used to detect potential corners in the image, the effective area is extracted through morphological operations and contour analysis, and the effective corners are screened using quadrilateral fitting and geometric constraints.

[0019] When the calibration board is a center-shaped calibration board, bright / dark circular spots in the LoG or DoG detection image are used as the result of blob extraction. Then, the neighborhood of each blob is binarized and the contour is extracted. The least squares method is used to fit the ellipse, and the center coordinates and radius are calculated. Valid circular points are screened by geometric symmetry and gray-level distribution, and the extracted center is optimized at the sub-pixel level using the gray-level centroid method.

[0020] Furthermore, when calculating intrinsic parameters in dual-objective calibration,

[0021] When the calibration board is a checkerboard calibration board, Zhang Zhengyou's calibration algorithm is called to perform corner detection → homography matrix estimation → closed-form solution of intrinsic and extrinsic parameter matrices → reprojection error minimization on each image, thereby outputting the intrinsic and extrinsic parameter matrices K and distortion coefficients D of the left and right cameras;

[0022] When the calibration board is a circle-centered calibration board, ellipse detection, eccentricity compensation, and circle center positioning are performed on the circle-centered calibration board image to obtain the circle center coordinates of the image; using the correspondence between the circle center and the known world coordinates of the planar circular array, the intrinsic / extrinsic parameters are initially estimated through the homography matrix, and the camera intrinsic parameter matrix K, distortion coefficient D, and extrinsic parameters of each image are globally optimized with the goal of minimizing the circle center reprojection error.

[0023] Furthermore, elliptic matching feature point localization is performed on each left and right image. The world coordinate system coordinates are set to plane Z=0, and a three-dimensional-two-dimensional coordinate correspondence is constructed. Taking the left camera coordinate system as the reference, the right camera extrinsic parameter [R|T] and the two camera intrinsic parameter matrices K1, K2, and distortion coefficients D1, D2 are set as joint optimization variables. The Levenberg-Marquardt method is used to minimize the reprojection error of all circle centers. After iterative convergence, the final binocular extrinsic parameter matrix R and translation vector T are output to complete the calibration.

[0024] This invention also provides a calibration method for simultaneous automatic calibration of binocular cameras. Based on the binocular camera simultaneous automatic calibration system described above, the calibration method specifically includes:

[0025] S1. Fix the binocular camera on the camera bracket, move the calibration plate back and forth to observe the image quality of the calibration plate, and ensure that the calibration plate does not exceed the field of view of the binocular camera. At the same time, adjust the depth of field of the binocular camera lens to ensure that the image is always clear within the target range. Then reset the cross slide assembly that drives the calibration plate to move, and prepare to start image acquisition.

[0026] S2. Within one rotation cycle, the control and data processing unit controls the outer ring flange to rotate at a set speed at a uniform speed. At the same time, the binocular camera performs real-time detection of the calibration plate. When the binocular camera detects the calibration plate, the control and data processing unit controls the binocular camera to acquire images of the calibration plate.

[0027] S3, the control and data processing unit reads the preset area threshold T. s Deviation threshold O from coordinate points s Record the calibration plate area S0 and center coordinate point O0 in the current calibration plate image, and simultaneously control the cross slide assembly to drive the calibration plate to move back and forth, and calculate the calibration plate area S in the subsequent calibration plate image. n and center point coordinates O n The calibration plate area S of the subsequent calibration plate image n and center point coordinates On Compared to the calibration plate area S in the previous image n-1 and center point coordinates O n-1 If the absolute value deviation between the two is greater than the preset area threshold T, then... s Deviation threshold O from coordinate points s If the current calibration board image is selected, it will be added to the image dataset; otherwise, the current calibration board image will not be selected.

[0028] S4. After acquiring the calibration plate area S0 and center coordinate point O0 in the first calibration plate image, the control and data processing unit starts to control the cross slide assembly to drive the calibration plate to move back and forth, which is used to simulate the process of acquiring calibration images when the camera module is at different distances from the calibration plate; when the calibration plate image is not detected, the cross slide assembly is controlled to stop driving the calibration plate to move until the binocular camera detects the calibration plate again.

[0029] S5. After completing the acquisition of the binocular image set, the control and data processing unit uses the binocular calibration algorithm for calibration.

[0030] Furthermore, S5 specifically includes:

[0031] S501, Calibration Board Key Point Inspection:

[0032] If the calibration board is a checkerboard calibration board, then the Harris corner detection or Shi-Tomasi algorithm is used to detect potential corners in the calibration board image, and the effective area is extracted through morphological operations and contour analysis. The effective corners are then selected using quadrilateral fitting and geometric constraints.

[0033] If the calibration board is a center-shaped calibration board, the bright / dark circular spots in the LoG or DoG detection image are used as the result of blob extraction. Then, the neighborhood of each blob is binarized and the contour is extracted. The least squares method is used to fit the ellipse, the center coordinates and radius are calculated, and the effective circular points are screened by geometric symmetry and gray-level distribution. The extracted center is optimized at the sub-pixel level using methods such as the gray-level centroid method.

[0034] S502. Calculate intrinsic parameters in dual-target calibration:

[0035] If the calibration board is a checkerboard calibration board, then the Zhang Zhengyou calibration algorithm is called to perform corner detection → homography matrix estimation → closed-form solution of intrinsic and extrinsic parameter matrices → reprojection error minimization on each image, thereby outputting the intrinsic and extrinsic parameter matrices K and distortion coefficients D of the left and right cameras;

[0036] If the calibration board is a circular calibration board, perform ellipse detection → eccentricity compensation → circle center localization on the image of the circular calibration board to obtain the coordinates of the circle center of the image; use the correspondence between it and the world coordinates of the known planar circular array to initially estimate the intrinsic / extrinsic parameters through the homography matrix, and globally optimize the camera intrinsic parameter matrix K, distortion coefficient D and the extrinsic parameters of each image with the goal of minimizing the circle center reprojection error;

[0037] S503. Perform stereo correction and calibrate binocular extrinsic parameters:

[0038] Ellipse matching feature point localization is performed on each left and right image. The world coordinate system coordinates are set to plane Z=0, and a three-dimensional-two-dimensional coordinate correspondence is constructed. Taking the left camera coordinate system as the reference, the right camera extrinsic parameter [R|T] and the two camera intrinsic parameter matrices K1, K2, and distortion coefficients D1, D2 are set as joint optimization variables. The Levenberg-Marquardt method is used to minimize the reprojection error of all circle centers. After iterative convergence, the final binocular extrinsic parameter matrix R and translation vector T are output to complete the calibration.

[0039] S504. Calculate the calibration accuracy error for evaluation:

[0040] The reprojection error is calculated based on the calibration results. If the reprojection error is less than 0.4 pixels, the calibration error is considered acceptable, and the calibration results of the binocular camera are obtained. Otherwise, the binocular camera calibration should be performed again.

[0041] The beneficial effects of this invention are as follows:

[0042] 1. Achieve full automation of the calibration process: Through automatic shooting control, online image quality evaluation and adaptive screening, completely eliminate the efficiency bottlenecks and human errors caused by manual placement of calibration boards and manual triggering of shooting, and meet the cycle time requirements of mass production lines.

[0043] 2. Unify and improve calibration accuracy: During the image acquisition stage, the illumination, blur, and occlusion are scored in real time, and only high-scoring images are retained for calculation, thereby suppressing the reprojection deviation caused by inconsistent image set quality from the root.

[0044] 3. Build a closed-loop quality traceability system: Generate a unique calibration parameter file and corresponding image fingerprint for each piece of equipment leaving the factory, supporting subsequent traceability and rapid online recalibration, reducing after-sales maintenance costs.

[0045] 4. Compatible with multiple camera and lens models: The algorithm is universal for multi-modal imaging systems such as visible light, infrared, and fisheye. Through adaptive distortion model and baseline compensation mechanism, it ensures consistent accuracy across different batches of hardware. Attached Figure Description

[0046] Figure 1This is a schematic diagram of the system for automated synchronous calibration of binocular cameras in this invention.

[0047] Figure 2 This is a flowchart illustrating the automated calibration method for binocular cameras in this invention.

[0048] Figure 3 This is a flowchart illustrating the automated control algorithm for binocular cameras in this invention.

[0049] In the attached diagram, there are: camera bracket 1, inner ring flange 2, outer ring flange 3, cross slide assembly 4, calibration plate bracket 5, and control and data processing unit 6.

[0050] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0051] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0052] Binocular calibration originates from the epipolar constraint theory in stereo vision geometry. Its core task is to solve for the extrinsic parameter matrices (rotation R and translation T) between the two cameras, thereby establishing the mapping relationship between 3D spatial points and the coordinates of the left and right images. The traditional process follows a two-step approach: monocular first, then binocular. First, Zhang's calibration method is used to estimate the intrinsic parameters (focal length, principal point, distortion coefficients) of the left and right cameras respectively. Then, the extrinsic parameters are calculated using corresponding feature points, and nonlinear optimization is performed with reprojection error as the objective function to further reduce the calibration error.

[0053] Existing binocular camera calibration methods typically involve handheld or static support, placing the planar calibration plate in different poses and manually acquiring images. As a result, it is difficult to cover the entire field of view of the binocular camera and different depths and tilt angles in a limited space at one time. This can easily lead to poor repeatability of the calibration plate pose, resulting in calibration accuracy being greatly affected by human factors. Furthermore, this acquisition process is time-consuming and labor-intensive, making it difficult to achieve online calibration on mass production lines.

[0054] This invention provides a compact and highly automated binocular camera calibration device capable of automatically acquiring multiple sets of images with different poses within a single rotation cycle, and calibrating multiple sets of binocular cameras. This significantly improves calibration accuracy and efficiency, while ensuring that the calibration parameters of binocular cameras produced on a production line are essentially the same. This invention is used to accurately obtain the relative pose parameters between cameras and can be further applied to scenarios such as 3D reconstruction, robot navigation, autonomous driving, or industrial inspection. Specifically, the calibration method is applicable to combined systems of visible light cameras, infrared cameras, fisheye cameras, or multispectral cameras.

[0055] like Figure 1As shown, the present invention provides a system for automated synchronous calibration of binocular cameras, comprising:

[0056] Base;

[0057] Inner ring flange 2, the inner ring flange 2 is fixed on the base, the inner ring flange 2 is circular in shape, and is used to fix M (M≥2) calibration plates, the calibration plates are checkerboard calibration plates or center-shaped calibration plates;

[0058] The outer ring flange 3 is annular in shape and is mounted on the base. The inner ring of the outer ring flange 3 surrounds the outer circumference of the inner ring flange 2 and rotates around the outer circumference of the inner ring flange 2. The outer ring flange 3 is arranged along the inner ring radius and drives N sets (N≥2) of binocular cameras to rotate around the calibration plate.

[0059] A cross slide assembly 4 is disposed on the inner ring flange 2, and the calibration plate is fixed thereon. The cross slide assembly 4 is used to drive the calibration plate to change position in the front-back and left-right directions in order to adjust the optical axis distance between the binocular camera and the calibration plate.

[0060] Camera bracket 1, which is fixed on the outer ring flange 3, is used to place and fix the binocular camera;

[0061] The control and data processing unit 6 is fixed inside the base and connected to the outer ring flange 3, the cross slide assembly 4, and the binocular camera. It is used to control the rotation of the outer ring flange 3, control the position movement of the calibration plate, and control the binocular camera to acquire images, i.e., stepper motor control and synchronous control of image acquisition. Then, it calculates the intrinsic and extrinsic parameters of the binocular camera based on the acquired image set to realize the automatic calibration of the binocular camera.

[0062] In one embodiment, the calibration plate and the cross slide assembly 4 are connected by a calibration plate bracket 5. Multiple cross slide assemblies 4 are mounted on the inner ring flange 2, each corresponding to a calibration plate bracket 5 and a calibration plate. The height of the calibration plate brackets 5 is adjusted so that the heights of the multiple calibration plate brackets 5 are all different, resulting in unequal heights of the calibration plates and a certain height difference between them. The height difference between the multiple calibration plate brackets 5 and the calibration plates is Δh, and Δh satisfies Δh=tanθ*D, where D is the closest distance from the calibration plate to the optical center of the binocular camera, θ is the vertical field of view angle of the binocular camera, and the pitch angle can be independently adjusted to simulate different field of view tilt angles.

[0063] In one embodiment, the process of controlling the binocular camera to acquire images in the control and data processing unit 6 is as follows: when the binocular camera detects the calibration board, it controls the binocular camera to synchronously acquire images of the calibration board and records the current calibration board area S. n With the center point coordinates O n Simultaneously, the cross slide assembly 4 is controlled to begin moving slowly and uniformly backward. When a new calibration plate area S is detected... n+1 With S n The difference is greater than the corresponding area threshold T s Or the new center point coordinates O n+1 With O n The difference is greater than the corresponding coordinate point deviation threshold O s If the calibration board is considered to have changed its spatial pose, the current image of the calibration board will be acquired; otherwise, the image acquisition signal will not be triggered.

[0064] Once the binocular camera starts acquiring calibration images, the outer ring flange 3 is controlled to rotate at a constant speed within one rotation cycle, while the binocular camera performs real-time detection of the calibration plate. After the outer ring flange 3 completes one rotation cycle, the binocular camera is controlled to stop real-time detection and acquisition of images.

[0065] That is, the process of controlling the binocular camera to acquire images in the control and data processing unit is as follows: after the system equipment is started, the outer ring flange is controlled to rotate at a constant speed within one rotation cycle, while the binocular camera performs real-time detection of the calibration plate; when the binocular camera detects the calibration plate, it is controlled to synchronously acquire the calibration plate image and record the current calibration plate area S. n With the center point coordinates O n Simultaneously, the cross slide assembly is controlled to begin a slow, uniform backward movement. When a new calibration plate area S is detected... n+1 With S n The difference is greater than the corresponding area threshold T s Or the new center point coordinates O n+1 With O n The difference is greater than the corresponding coordinate point deviation threshold O s When the calibration plate is considered to have changed its spatial pose, the current image of the calibration plate is acquired; otherwise, the image acquisition signal is not triggered. After the outer ring flange completes one rotation cycle, the binocular camera is controlled to stop real-time detection and acquisition of images.

[0066] In one embodiment, the detection algorithm used by the control and data processing unit includes, but is not limited to, corner detection using Harris corner detection or Shi-Tomasi algorithm and related optimization algorithms, and center detection using LoG or DoG detection and related center fitting algorithms. The process of dual-target calibration in the control and data processing unit 6 is as follows: when the calibration board is a checkerboard calibration board, Harris corner detection or Shi-Tomasi algorithm is used to detect potential corners in the image; effective regions are extracted through morphological operations and contour analysis; and effective corners are selected using quadrilateral fitting and geometric constraints. When calculating intrinsic parameters in dual-target calibration, Zhang Zhengyou calibration algorithm is called to perform corner detection → homography matrix estimation → closed-form solution of intrinsic and extrinsic parameter matrices → minimization of reprojection error for each image, thereby outputting the intrinsic and extrinsic parameter matrices K and distortion coefficients D of the left and right cameras.

[0067] When the calibration board is a center-type calibration board, bright / dark circular spots in the image are detected using LoG or DoG methods as the result of blob extraction. Then, the neighborhood of each blob is binarized and its contour extracted. An ellipse is fitted using the least squares method, and the center coordinates and radius are calculated. Valid circular points are selected based on geometric symmetry and grayscale distribution, and the extracted center is optimized at the sub-pixel level using the grayscale centroid method. When calculating intrinsic parameters in dual-target calibration, ellipse detection → eccentricity compensation → center localization are performed on the center-type calibration board image to obtain the image center coordinates. Using the correspondence between these coordinates and the known world coordinates of a planar circular array, the intrinsic / extrinsic parameters are initially estimated using the homography matrix. Finally, the camera intrinsic parameter matrix K, distortion coefficient D, and extrinsic parameters for each image are globally optimized with the goal of minimizing the center reprojection error.

[0068] Elliptic matching feature point localization is performed on each left and right image. The world coordinate system coordinates are set to plane Z=0, and a three-dimensional-two-dimensional coordinate correspondence is constructed. Taking the left camera coordinate system as the reference, the right camera extrinsic parameter [R|T] and the two camera intrinsic parameter matrices K1, K2, and distortion coefficients D1, D2 are set as joint optimization variables. The Levenberg-Marquardt method is used to minimize the reprojection error of all circle centers. After iterative convergence, the final binocular extrinsic parameter matrix R and translation vector T are output to complete the calibration.

[0069] like Figure 2 , Figure 3 As shown, the present invention also provides a calibration method for simultaneous automatic calibration of binocular cameras. Based on the binocular camera simultaneous automatic calibration system described above, the calibration method specifically includes:

[0070] S1. Equipment debugging and initialization

[0071] First, fix the binocular camera on the camera bracket 1, move the calibration plate back and forth to observe the image quality of the calibration plate, and ensure that the calibration plate does not exceed the field of view of the binocular camera. At the same time, adjust the depth of field of the binocular camera lens to ensure that the image is always clear within the target range. Then, reset the cross slide assembly 4 that drives the calibration plate to move, and prepare to start acquiring multiple sets of binocular images.

[0072] S2, Equipment control outer ring rotation

[0073] After the binocular camera begins acquiring calibration images, within one rotation cycle, the control and data processing unit 6 controls the outer ring flange 3 to rotate at a set speed at a constant speed. Simultaneously, the binocular camera performs real-time detection of the calibration plate. Once the binocular camera detects the calibration plate, the control and data processing unit 6 controls the binocular camera to acquire images of the calibration plate. If the outer ring completes one rotation cycle, the binocular camera stops real-time image detection and acquisition, and all modules of the device reset to their original positions.

[0074] S3, Calibration Board Image Detection and Acquisition

[0075] Control and data processing unit 6 reads the preset area threshold T s Deviation threshold O from coordinate points s Record the calibration plate area S0 and center coordinate point O0 in the current calibration plate image, and simultaneously control the cross slide assembly 4 to drive the calibration plate to start moving back and forth, and calculate the calibration plate area S in the subsequent calibration plate image. n and center point coordinates O n The calibration plate area S of the subsequent calibration plate image n and center point coordinates O n Compared to the calibration plate area S in the previous image n-1 and center point coordinates O n-1 If the absolute value deviation between the two is greater than the preset area threshold T, then... s Deviation threshold O from coordinate points s If the calibration board has changed its spatial pose, then this calibration board image should be collected as the calibration image dataset; otherwise, the current calibration board image should not be collected.

[0076] S4, Motion control of the calibration board

[0077] After acquiring the calibration plate area S0 and center coordinate point O0 in the first calibration plate image, the control and data processing unit 6 starts to control the cross slide assembly 4 to drive the calibration plate to move back and forth, which is used to simulate the process of acquiring calibration images under different distances from the camera module to the calibration plate; when the calibration plate image is not detected, the cross slide assembly 4 is controlled to stop driving the calibration plate to move until the binocular camera detects the calibration plate again.

[0078] S5. After the outer ring has rotated one revolution, stop rotating the outer ring to complete the acquisition of the binocular image set. At the same time, the calibration board support 5 and the cross slide assembly 4 are reset. The control and data processing unit 6 uses a binocular calibration algorithm to perform binocular calibration and burns the generated binocular calibration results into the binocular camera module. The binocular camera module can then be removed and used. The binocular calibration algorithm specifically includes:

[0079] S501, Calibration Board Key Point Inspection:

[0080] Key points in the calibration board are detected. If the calibration board is a checkerboard calibration board, the Harris corner detection or Shi-Tomasi algorithm is used to detect potential corner points in the calibration board image. The effective area is extracted by morphological operations (such as dilation and erosion) and contour analysis. The effective corner points are screened by quadrilateral fitting and geometric constraints (such as collinear corner points and spacing rules).

[0081] If the calibration board is a center-shaped calibration board, bright / dark circular spots in the image are detected using LoG (Laplacian of Gaussian) or DoG (Difference of Gaussians) as the results of blob extraction. Then, the neighborhood of each blob is binarized and its contour extracted. An ellipse is fitted using the least squares method, and the center coordinates and radius are calculated. Valid circles are selected based on geometric symmetry (such as dot spacing and row / column alignment) and grayscale distribution (contrast between the dot center and the background). Subpixel-level optimization of the extracted center is performed using methods such as the grayscale centroid method.

[0082] S502. Calculate intrinsic parameters in dual-target calibration:

[0083] If the calibration board is a checkerboard calibration board, then the Zhang Zhengyou calibration algorithm is called to perform corner detection → homography matrix estimation → closed-form solution of intrinsic and extrinsic parameter matrices → reprojection error minimization on each image, thereby outputting the intrinsic and extrinsic parameter matrices K and distortion coefficients D of the left and right cameras.

[0084] If the calibration board is a circle-centered calibration board, perform ellipse detection → eccentricity compensation → circle center positioning on the circle-centered calibration board image to obtain the circle center coordinates of the image; use the correspondence between it and the world coordinates of the known planar circular array to initially estimate the intrinsic / extrinsic parameters through the homography matrix, and globally optimize the camera intrinsic parameter matrix K, distortion coefficient D and the extrinsic parameters of each image with the goal of minimizing the circle center reprojection error.

[0085] S503. Perform stereo correction and calibrate binocular extrinsic parameters:

[0086] Elliptic matching feature point localization is performed on each left and right image. The world coordinate system coordinates are set to plane Z=0, and a three-dimensional-two-dimensional coordinate correspondence is constructed. Taking the left camera coordinate system as the reference, the right camera extrinsic parameter [R|T] and the two camera intrinsic parameter matrices K1, K2, and distortion coefficients D1, D2 are set as joint optimization variables. The Levenberg-Marquardt method is used to minimize the reprojection error of all circle centers. After iterative convergence, the final binocular extrinsic parameter matrix R and translation vector T are output to complete the calibration.

[0087] S504. Calculate the calibration accuracy error for evaluation:

[0088] The reprojection error is calculated based on the calibration results. If the reprojection error is less than 0.4 pixels, the calibration error is considered acceptable, and the calibration results of the binocular camera are obtained. Otherwise, the binocular camera calibration should be performed again.

[0089] In this invention, the control and data processing unit 6 outputs a synchronous trigger signal, causing each group of binocular cameras to acquire images at the same time. Simultaneously, it drives the outer ring of the concentric flange to rotate around its central axis, driving the calibration plate located on the inner ring to generate a two-dimensional translation along the cross guide rail. This causes the calibration plate to form a multi-pose, full-coverage motion trajectory within the common field of view of each binocular camera. Under each pose of the trajectory, each group of binocular cameras synchronously captures images of the calibration plate, obtaining a multi-binocular, multi-view calibration image dataset. Finally, the dataset is subjected to center detection, eccentricity compensation, and joint optimization, and the intrinsic parameters, distortion coefficients, and extrinsic parameter matrices between all binocular cameras are calculated at once.

[0090] This invention directly addresses the four major pain points of traditional dual-target calibration: manual setup, inconsistent frame quality, cycle time bottlenecks, and loss of consistency control. It establishes a calibration system with "zero manual labor, zero defects, and zero line changes," employing adaptive visual quality gatekeeping, millisecond-level closed-loop calculation, and multi-model co-line adaptation to achieve automated, standardized, and large-scale output of 3D visual benchmarks in mass production scenarios. Its purpose is:

[0091] 1. Achieve full automation of the calibration process: Through automatic shooting control, online image quality evaluation and adaptive screening, completely eliminate the efficiency bottlenecks and human errors caused by manual placement of calibration boards and manual triggering of shooting, and meet the cycle time requirements of mass production lines.

[0092] 2. Unify and improve calibration accuracy: During the image acquisition stage, the illumination, blur, and occlusion are scored in real time, and only high-scoring images are retained for calculation, thereby suppressing the reprojection deviation caused by inconsistent image set quality from the root.

[0093] 3. Build a closed-loop quality traceability system: Generate a unique calibration parameter file and corresponding image fingerprint for each piece of equipment leaving the factory, supporting subsequent traceability and rapid online recalibration, reducing after-sales maintenance costs.

[0094] 4. Compatible with multiple camera and lens models: The algorithm is universal for multi-modal imaging systems such as visible light, infrared, and fisheye. Through adaptive distortion model and baseline compensation mechanism, it ensures consistent accuracy across different batches of hardware.

[0095] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, apparatus, article, or method. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or method that includes that element.

[0096] The above description is merely a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A system for automated synchronous calibration of binocular cameras, characterized in that, include: Base; An inner ring flange, which is fixed on the base, is used to fix M calibration plates; An outer ring flange is disposed on the base and rotates around the inner ring flange. The outer ring flange is used to arrange and drive N sets of binocular cameras to rotate around the calibration plate. A cross slide assembly is disposed on the inner ring flange, and the calibration plate is fixed thereon. The cross slide assembly is used to drive the calibration plate to change position in the front-back and left-right directions, so as to adjust the optical axis distance between the binocular camera and the calibration plate. A camera bracket, which is fixed to the outer ring flange, is used to place and fix the binocular camera; A control and data processing unit is fixed inside the base and connected to the outer ring flange, the cross slide assembly, and the binocular camera. The control and data processing unit is used to control the rotation of the outer ring flange, control the position movement of the calibration plate, and control the binocular camera to acquire images. It also calculates the intrinsic and extrinsic parameters of the binocular camera based on the acquired image set, thereby realizing the automatic calibration of the binocular camera.

2. The system for automated synchronous calibration of binocular cameras according to claim 1, characterized in that, The inner ring flange is circular in shape, the outer ring flange is annular in shape, and the inner ring of the outer ring flange surrounds the outer circumference of the inner ring flange. The calibration board is a checkerboard calibration board or a center-shaped calibration board, and the number of the calibration board and the binocular camera is greater than or equal to 2.

3. The system for automated synchronous calibration of binocular cameras according to claim 2, characterized in that, The calibration plate and the cross slide assembly are connected by a calibration plate bracket. Multiple cross slide assemblies are provided on the inner ring flange. Each cross slide assembly corresponds to a calibration plate bracket and a calibration plate. The heights of the multiple calibration plate brackets are different, and the height difference of the multiple calibration plate brackets is Δh, and Δh satisfies Δh=tanθ*D, where D is the shortest distance from the calibration plate to the optical center of the binocular camera, θ is the vertical field of view angle of the binocular camera, and the pitch angle can be adjusted independently to simulate different field of view tilt angles.

4. The system for automated synchronous calibration of binocular cameras according to claim 2, characterized in that, The process of controlling the binocular camera to acquire images in the control and data processing unit is as follows: After the system equipment is started, the outer ring flange is controlled to rotate at a constant speed within one rotation cycle, while the binocular camera performs real-time detection of the calibration plate; when the binocular camera detects the calibration plate, it is controlled to synchronously acquire the image of the calibration plate and record the current calibration plate area S. n With the center point coordinates O n Simultaneously, the cross slide assembly is controlled to begin a slow, uniform backward movement. When a new calibration plate area S is detected... n+1 With S n The difference is greater than the corresponding area threshold T s Or the new center point coordinates O n+1 With O n The difference is greater than the corresponding coordinate point deviation threshold O s When the calibration plate is considered to have changed its spatial pose, the current image of the calibration plate is acquired; otherwise, the image acquisition signal is not triggered. After the outer ring flange completes one rotation cycle, the binocular camera is controlled to stop real-time detection and acquisition of images.

5. The system for automated synchronous calibration of binocular cameras according to claim 4, characterized in that, When the stereo camera to be calibrated detects the calibration board, the feature points of the calibration board are identified through corner detection and related optimization methods, and the intrinsic and extrinsic parameters of the stereo camera are solved through the feature points.

6. A calibration method for automated synchronous calibration of a binocular camera, characterized in that, The system for synchronous automated calibration of a binocular camera based on any one of claims 1-5, wherein the calibration method specifically includes: S1. Fix the binocular camera on the camera bracket, move the calibration plate back and forth to observe the image quality of the calibration plate, and ensure that the calibration plate does not exceed the field of view of the binocular camera. At the same time, adjust the depth of field of the binocular camera lens to ensure that the image is always clear within the target range. Then reset the cross slide assembly that drives the calibration plate to move, and prepare to start image acquisition. S2. Within one rotation cycle, the control and data processing unit controls the outer ring flange to rotate at a set speed at a uniform speed. At the same time, the binocular camera performs real-time detection of the calibration plate. When the binocular camera detects the calibration plate, the control and data processing unit controls the binocular camera to acquire images of the calibration plate. S3, the control and data processing unit reads the pre-set area threshold T. s Deviation threshold O from coordinate points s Record the calibration plate area S0 and center coordinate point O0 in the current calibration plate image, and simultaneously control the cross slide assembly to drive the calibration plate to move back and forth, and calculate the calibration plate area S in the subsequent calibration plate image. n and center point coordinates O n The calibration plate area S of the subsequent calibration plate image n and center point coordinates O n Compared to the calibration plate area S in the previous image n-1 and center point coordinates O n-1 If the absolute value deviation between the two is greater than the preset area threshold T, then... s Deviation threshold O from coordinate points s If the current calibration board image is selected, it will be added to the image dataset; otherwise, the current calibration board image will not be selected. S4. After acquiring the calibration plate area S0 and center coordinate point O0 in the first calibration plate image, the control and data processing unit starts to control the cross slide assembly to drive the calibration plate to move back and forth, which is used to simulate the process of acquiring calibration images when the camera module is at different distances from the calibration plate; when the calibration plate image is not detected, the cross slide assembly is controlled to stop driving the calibration plate to move until the binocular camera detects the calibration plate again. S5. After completing the acquisition of the binocular image set, the control and data processing unit uses the binocular calibration algorithm for calibration.

7. The calibration method for synchronous automated calibration of binocular cameras according to claim 6, characterized in that, S5 specifically includes: S501, Calibration Board Key Point Inspection: If the calibration board is a checkerboard calibration board, then the Harris corner detection or Shi-Tomasi algorithm is used to detect potential corners in the calibration board image, and the effective area is extracted through morphological operations and contour analysis. The effective corners are then selected using quadrilateral fitting and geometric constraints. If the calibration board is a center-shaped calibration board, the bright / dark circular spots in the LoG or DoG detection image are used as the result of blob extraction. Then, the neighborhood of each blob is binarized and the contour is extracted. The least squares method is used to fit the ellipse, the center coordinates and radius are calculated, and the effective circular points are screened by geometric symmetry and gray-level distribution. The gray-level centroid method is used to optimize the extracted center at the sub-pixel level. S502. Calculate intrinsic parameters in dual-target calibration: If the calibration board is a checkerboard calibration board, then the Zhang Zhengyou calibration algorithm is called to perform corner detection → homography matrix estimation → closed-form solution of intrinsic and extrinsic parameter matrices → reprojection error minimization on each image, thereby outputting the intrinsic and extrinsic parameter matrices K and distortion coefficients D of the left and right cameras; If the calibration board is a circular calibration board, perform ellipse detection → eccentricity compensation → circle center localization on the image of the circular calibration board to obtain the coordinates of the circle center of the image; use the correspondence between it and the world coordinates of the known planar circular array to initially estimate the intrinsic / extrinsic parameters through the homography matrix, and globally optimize the camera intrinsic parameter matrix K, distortion coefficient D and the extrinsic parameters of each image with the goal of minimizing the circle center reprojection error; S503. Perform stereo correction and calibrate binocular extrinsic parameters: Ellipse matching feature point localization is performed on each left and right image. The world coordinate system coordinates are set to plane Z=0, and a three-dimensional-two-dimensional coordinate correspondence is constructed. Taking the left camera coordinate system as the reference, the right camera extrinsic parameter [R|T] and the two camera intrinsic parameter matrices K1, K2, and distortion coefficients D1, D2 are set as joint optimization variables. The Levenberg-Marquardt method is used to minimize the reprojection error of all circle centers. After iterative convergence, the final binocular extrinsic parameter matrix R and translation vector T are output to complete the calibration. S504. Calculate the calibration accuracy error for evaluation: The reprojection error is calculated based on the calibration results. If the reprojection error is less than 0.4 pixels, the calibration error is considered acceptable, and the calibration results of the binocular camera are obtained. Otherwise, the binocular camera calibration should be performed again.

Images