A method for calibrating the camera internal parameters of a defocus camera based on a round hole target

Abstract

The application provides a method for calibrating the inner parameters of a defocus camera based on a circular hole target, comprising the following steps: step one: according to a light ray tracing model, an inner parameter matrix model containing unknown image distance, focal length and camera center position coordinates is established; step two: the circular hole target is moved for multiple times, and the size of each element in the matrix is obtained by using Zhang Zhengyou's calibration method; step three: the value of a polynomial containing the image distance and the focal length is calculated by moving the position of a point light source; step four: the CCD camera is moved, and a nonlinear equation set is established according to the difference value information of different image distances and focal lengths and the parameters obtained in step two, and the image distance, the camera focal length, the camera center position coordinates and the position of the point light source are solved. The application can solve the focal length and the coordinates of the camera center point under the condition of unknown image distance, and solves the problem that most calibration methods assume that the image distance is equal to the focal length, and the calibration is not accurate enough in the case that the image plane deviates from the focal plane.

Description

Technical Field

[0001] This invention belongs to the field of optical imaging and photoelectric measurement, and specifically relates to a method for calibrating the intrinsic parameters of a defocused camera based on a circular aperture target. Background Technology

[0002] The geometric correspondence between a point on the surface of a 3D object and a point in a 2D image is determined by the camera's geometric imaging model and its position and orientation. The parameters of the geometric imaging model are the camera's intrinsic parameters, and the intrinsic parameter matrix typically includes focal length, principal point coordinates, and tilt factor. Camera calibration is a crucial step in 3D measurement, aiming to determine the relationship between the world coordinate system and the pixel coordinate system. The calibration accuracy directly affects the accuracy of the vision measurement system.

[0003] Currently, the main methods for front camera calibration can be divided into two categories: self-calibration methods and calibration target methods. Self-calibration methods require rigid features in the measured scene, limiting their application. Calibration target methods, due to their simplicity, are more widely used. Among calibration target methods, the Zhang Zhengyou calibration method is widely used. The Zhang Zhengyou calibration method sets up a world coordinate system, a camera coordinate system, an image coordinate system, and a pixel coordinate system. The extrinsic parameter matrix is ​​from the world coordinate system to the camera coordinate system; the intrinsic parameter matrix is ​​from the camera coordinate system to the pixel coordinate system. This method obtains the intrinsic parameter matrix based on a pinhole imaging model and assumes that the image distance equals the focal length. If the image plane deviates from the focal plane, the calibrated focal length will not be accurate enough. Summary of the Invention

[0004] To address the deficiencies and shortcomings of existing technologies, this invention provides a method for calibrating the intrinsic parameters of a defocused camera based on a circular aperture target, enabling accurate calibration of the imaging intrinsic parameters of a monocular vision system and providing a technical foundation for ensuring the three-dimensional perception accuracy of monocular stereo vision technology.

[0005] The technical solution adopted in this invention is: a method for calibrating the intrinsic parameters of a defocused camera based on a circular aperture target, comprising the following steps:

[0006] Step 1: First, define a coordinate system O with the point light source as the center and the origin. A -X w Y w Z w Define the coordinates of point A (x) A ,y A The coordinates of point C are (x, 0). C ,y C ,z C ), coordinates of point D (x D ,y D ,z D Establish a model containing unknowns: image distance L, focal length f, camera center coordinates (u0, v0), and the distance z of the lens principal point on the optical axis.C A new intrinsic parameter matrix model.

[0007] Step 2: Define the circular target with the vertical Z-axis as the coordinates in the world coordinate system; move the circular target multiple times to obtain the coordinates of the light rays emitted from the circular target on the camera target at this time, save n images (n≥3, usually 4-6) of the light rays emitted from the target reaching the image plane, and use Zhang Zhengyou calibration method to find the size of each element in the intrinsic parameter matrix of Step 1.

[0008] Step 3: Move the position of the point light source to obtain the values ​​of the polynomial containing the image distance and focal length under different point light source positions.

[0009] Step 4: Move the CCD camera along the optical axis to obtain the moving distance, and solve the system of equations to obtain the system focal length and principal point coordinates: Based on the parametric equations in Step 2 and the camera moving distance information, solve the system of equations to obtain the system image distance, focal length, camera center position coordinates, and point light source position.

[0010] Further, step one: First, define the camera coordinate system O with the camera center as the origin. A -X w Y w Z w Define the coordinates of point A (x) A ,y A The coordinates of point C are (x, 0). C ,y C ,z C ), coordinates of point D (x D ,y D ,z D Based on the ray tracing model, an unknown quantity is established: image distance L, focal length f, camera center position coordinates (u0, v0), and point light source position (x). A ,y A And the distance z of the principal point of the lens on the optical axis C A new intrinsic parameter matrix model.

[0011] Further, in step two: move the circular target multiple times to obtain images of the light rays emitted from the circular target reaching the image plane and save n images, n≥3, usually 4-6. Use Zhang Zhengyou calibration method to obtain the size of each element in the new intrinsic parameter matrix.

[0012] Further, step three: move the position of the point light source and calculate the value of the polynomial that includes the image distance and focal length.

[0013] Further, step four: Move the camera target along the optical axis and establish a system of equations based on the moving distance: Move the CCD camera multiple times and solve the system of equations to obtain the system image distance, focal length, camera center position coordinates, and point light source position.

[0014] The principle of this invention is as follows: A light ray emitted from a point light source passes through the center of a circular hole on a calibration target. The camera's main lens is considered a thin lens. Based on the imaging principle of a thin lens, the relationship between a point on the lens and a point received by the camera is obtained. Based on ray tracing, an intrinsic parameter matrix containing unknown quantities—image distance, focal length, and the coordinates of the camera's center position—is derived. The magnitudes of the unknown elements in the intrinsic parameter matrix are solved using Zhang Zhengyou's calibration method. Then, by changing the position information of the point light source and using the distance information between the point light sources, a system of nonlinear equations sufficient to solve for all parameters is obtained.

[0015] The beneficial effects of this invention are as follows:

[0016] (1) The present invention can achieve camera internal parameter calibration when the camera image plane is not in the focal plane, which is not possible with the prior art; the present method can calculate the focal length based on the image distance.

[0017] (2) The calibration technology based on Zhang Zhengyou’s traditional calibration method has good robustness, practicality and high accuracy, and can obtain higher accuracy calibration results. Attached Figure Description

[0018] Figure 1 This is a schematic flowchart of a method for calibrating the intrinsic parameters of a defocused camera based on a circular aperture target according to the present invention.

[0019] Figure 2 This is a schematic diagram for calibrating a defocused camera.

[0020] Among them, 1 is a point light source, 2 is a circular target, 3 is the main lens, 4 is the CCD target surface, and 5 is the virtual object surface. Detailed Implementation

[0021] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0022] like Figure 1 As shown, the present invention provides a method for calibrating the intrinsic parameters of a defocused camera based on a circular aperture target, comprising the following steps:

[0023] Step 1: First, define a coordinate system O with the camera center as the origin. A -X w Y w Z w Define the coordinates of point A (x) A ,y A The coordinates of point C are (x, 0). C ,y C ,z C ), coordinates of point D (x D ,y D ,z D From point C to point D, we can obtain the following by ray tracing:

[0024]

[0025] Where: θ', θ、 Points D and X are respectively w axis, Y w The angle between the axes and point C and X w axis, Y w The angle between the axes, L is the distance from the camera principal point to the CCD target surface, and f is the camera focal length.

[0026] Simplifying the matrix yields:

[0027]

[0028] Furthermore, based on similar triangles, we can obtain:

[0029]

[0030] Further simplification yields:

[0031]

[0032] From point E to point C on the virtual surface, satisfying the pinhole imaging model, point A can be considered as the origin in Zhang Zhengyou's calibration intrinsic parameters. Therefore, the distance from point E to point C, expressed in matrix form, is:

[0033]

[0034] Among them, (x E ,y E ,z E () represents the coordinates of point E.

[0035] Furthermore, the matrix from point E to point D, from the camera coordinate system to the image coordinate system, can be represented as:

[0036]

[0037] For an infinitely distant target z E Approaching infinity, we can therefore obtain:

[0038]

[0039] From the image coordinate system to the pixel coordinate system, neglecting the tilt factor, we obtain the intrinsic parameter matrix:

[0040]

[0041] Where: M is defined as the intrinsic parameter matrix, containing unknowns: L is the distance from the center of the principal lens to the camera target surface, f is the camera focal length, and z is the distance from the center of the principal lens to the target surface of the camera. CThe direction of distance from the lens center to the origin is (x) A ,y A Let (u0, v0) be the coordinates of the point light source, and (u0, v0) be the coordinates of the camera center. dx and dy represent the physical lengths of a pixel on the camera target surface in the X and Y directions, respectively, and are generally considered to be known quantities.

[0042] Step Two: As Figure 2 As shown, the system includes a point light source 1, a circular aperture target 2, a main lens 3, a CCD target surface 4, and a virtual object surface 5. Light emitted from point light source 1 passes through the circular aperture target 2, then through the main lens 3, and reaches the CCD target surface 4. The virtual object surface 5 is conjugate to the CCD target surface 4. First, for the light emitted from a point light source A passing through the circular aperture target 2, move the circular aperture target 2 and change the angle between the plane of the circular aperture target 2 and the camera detection surface to obtain images of the light emitted from the circular aperture target 2 reaching the image plane, and save n images (n≥3, usually 4-6). Record the image point D of each circular aperture center point B on the calibration plate image and the coordinates of the circular aperture center point B in the world coordinate system. Based on Zhang Zhengyou's camera calibration method, solve for the intrinsic parameter matrix at this time, and then calculate the size of each element in the intrinsic parameter matrix of step one.

[0043]

[0044] Where: f x ', f y ', u0', and v0' are all known quantities obtained by Zhang Zhengyou's calibration method. Therefore:

[0045]

[0046] Step 3: Move the point light source to change the coordinates of point A, obtain the distances between multiple points A, and perform the calculation: Assume the previous point A was A... i (x Ai ,y Ai ,0), the next time point A is A j (x Aj ,y Aj Point A was measured (0, 0). i and A j distance d ij :

[0047]

[0048] Among them, f xi ′、u 0i ′、v 0i ′;f xj ′、u 0j ′、v 0j ′ represents the elements in the two intrinsic parameter matrices calculated by Zhang Zhengyou's calibration method after the two light source movements.

[0049] Therefore, the value of △k=1-L / f can be calculated.

[0050] Step 4: Move the camera CCD to obtain the distance the CCD moves along the optical axis, and calculate other parameters. Assume the previous distance of the CCD from the lens center was L. m The previous distance between the CCD and the center of the lens was L. n The distance L that the CCD moved along the optical axis was measured. m -L n :

[0051] Therefore, it can be calculated that when the point light source is stationary:

[0052]

[0053] Among them, f xm ′、f xn ′ are f obtained after two CCD movements, respectively. x ';△k m , △k m The values ​​of Δk = 1 - L / f are obtained after the CCDs move in two separate cycles.

[0054] Furthermore, we can determine: L m L n And by substituting f back into the equation, we can obtain:

[0055]

[0056] Among them, u 0m ′、v 0m ′;u 0n ′、v 0n ′ represents u'0v'0 obtained after two CCD movements, respectively.

[0057] It is easy to see that x can be calculated. A ,y A ,u0,v0.

Claims

1. A method for calibrating the intrinsic parameters of a defocused camera based on a circular aperture target, characterized in that: Includes the following steps: Step 1: First, define a coordinate system O with the point light source as the center and the origin. A -X w Y w Z w Define the coordinates of point A ( , ， 0), coordinates of point C ( , , ), coordinates of point D ( , , Based on the ray tracing model, a new intrinsic parameter matrix model is established that includes unknowns: image distance, camera focal length, camera center position coordinates, and point light source position. The coordinate system O is defined with the camera center as the origin. A -X w Y w Z w Define the coordinates of point A ( , , 0), coordinates of point C ( , , ), coordinates of point D ( , , From point C to point D, we can obtain the following by ray tracing: in: , , , Points D and X are respectively w axis, Y w The angle between the axes and point C and X w axis, Y w The included angle of the axis, The distance from the camera's principal point to the CCD target surface. The focal length of the camera; From point E to point C on the virtual surface, satisfying the pinhole imaging model, point A can be considered as the origin in Zhang Zhengyou's calibration intrinsic parameters. Therefore, the distance from point E to point C, expressed in matrix form, is: in,( , , ( ) represents the coordinates of point E; Furthermore, the matrix from point E to point D, from the camera coordinate system to the image coordinate system, can be represented as: For infinity target Approaching infinity, we can therefore obtain: From the image coordinate system to the pixel coordinate system, neglecting the tilt factor, we obtain the intrinsic parameter matrix: Where: Definition This is the intrinsic parameter matrix, containing unknowns: The distance from the center of the main lens to the camera target surface. For camera focal length, The direction of distance from the lens center to the origin, ( , ) are the coordinates of the point light source, ( , ( ) represents the coordinates of the camera center. They represent The physical length of a pixel in a certain direction on the camera target surface. The quantity is known; Step 2: Move the circular target multiple times and use Zhang Zhengyou's calibration method to obtain the size of each element in the new intrinsic parameter matrix; By solving the intrinsic parameter matrix using Zhang Zhengyou's camera calibration method, the size of each element in the intrinsic parameter matrix from step one can be determined: in: , , , All of these are known quantities obtained by Zhang Zhengyou's calibration method, from which we can obtain: Step 3: Move the point light source and solve the polynomial containing the image distance and camera focal length based on the distance between the point light sources at different positions; Step 4: Move the CCD camera, and based on the difference between the image distance and the camera focal length, and the size of each element in the new intrinsic parameter matrix obtained in Step 2 using the Zhang Zhengyou calibration method, establish a nonlinear equation system and solve for the image distance, camera focal length, camera center position coordinates, and point light source position.

2. The method for calibrating the intrinsic parameters of a defocused camera based on a circular aperture target according to claim 1, characterized in that... Step 1: First, define the camera coordinate system O with the camera center as the origin. A -X w Y w Z w Define the coordinates of point A ( , , 0), coordinates of point C ( , , ), coordinates of point D ( , , Based on the ray tracing model, a model is established that includes unknowns: image distance. Camera focal length Camera center position coordinates ( ), point light source position ( , and the distance of the principal point of the lens on the optical axis A new intrinsic parameter matrix model.

3. The method for calibrating the intrinsic parameters of a defocused camera based on a circular aperture target according to claim 2, characterized in that... Step 2: Move the circular target multiple times to obtain images of the light rays emitted from the circular target reaching the image plane and save n images, n≥3. Use Zhang Zhengyou's calibration method to obtain the size of each element in the new intrinsic parameter matrix.

4. The method for calibrating the intrinsic parameters of a defocused camera based on a circular aperture target according to claim 3, characterized in that... Step 3: Move the point light source to calculate the value of the polynomial that includes the image distance and the camera focal length.

5. The method for calibrating the intrinsic parameters of a defocused camera based on a circular aperture target according to claim 4, characterized in that... Step 4: Move the camera target along the optical axis and establish a system of equations based on the moving distance: Move the CCD camera multiple times and solve the system of equations to obtain the system image distance, camera focal length, camera center position coordinates, and point light source position.

Images