A method and device for calibrating planar CT system parameters under non-ideal conditions

Abstract

The present invention discloses a method and device for calibrating parameters of a planar CT system under non-ideal conditions, the method comprising: establishing a physical model of a CT shooting platform; setting a calibration plate at any position between a light source and a detector, obtaining a first calibration image of the calibration plate through the detector, and obtaining the physical coordinates of the first sphere center of multiple spheres in the calibration plate coordinate system, and calculating an internal reference matrix in a relationship matrix according to the first calibration image and the first sphere center physical coordinates; setting the calibration plate on a rotating platform and rotating it for one circle, obtaining a second calibration plate image at multiple angles through the detector, and obtaining the second sphere center physical coordinates at corresponding angles, and calculating a rotation transformation matrix and a translation transformation matrix in a relationship matrix according to the second calibration plate image, the second sphere center physical coordinates and the internal reference matrix; and calculating the position parameters of the rotation axis in the rotating platform. The method can overcome the influence of mechanical errors of a planar CT system under non-ideal conditions.

Description

Technical Field

[0001] The present invention relates to the technical field of computer tomography, and in particular to a method and device for calibrating parameters of a planar CT system under non-ideal conditions. Background Art

[0002] In industrial production, defect detection is of great significance. It is not only an important means to ensure product quality, but also plays a key role in the entire production process. First, it can directly identify various flaws and defects on the production line to ensure high product quality; second, detecting product defects in the early stages of production can avoid entering the subsequent processing and assembly links, thereby reducing material waste and rework costs; in addition, the automated defect detection system can achieve high-speed and high-precision detection and improve production efficiency. However, conventional machine vision inspection can only detect damage information on the surface of an object, and it is difficult to detect defects inside the object, while CT inspection can clearly detect internal information of the object.

[0003] Computed Tomography (CT) is a commonly used non-destructive testing technology. It only needs to scan around an object at a certain angle to obtain the three-dimensional information inside the object. It is an efficient technology for detecting internal damage of parts. However, conventional cone-beam CT is easily limited by the volume and shape of the object during detection. For example, for larger plate-like structures, when the X-ray beam is close to parallel to the component, the transmission intensity is reduced, making the projection information at a certain angle invalid, resulting in artifacts in the reconstruction result and loss of reconstruction accuracy.

[0004] To this end, planar CT (Computed Laminography, CL) solves the above problem by tilting the CT rotation axis at a certain angle so that the entire flat panel is always within the irradiation range of X-rays during the shooting process, thereby obtaining complete projection information. However, traditional planar CT imaging technology requires precise mechanical accuracy during application to ensure that the X-ray source, rotation axis and detector centerline are all in the same plane. The current mainstream three-dimensional reconstruction algorithms are all based on the situation where the device is under ideal conditions, that is, it is assumed that the light source, the rotation axis of the rotating platform, and the vertical centerline of the detector are coplanar. However, it is difficult to meet this condition in actual situations. Due to the influence of mechanical manufacturing and assembly errors, there is often a certain offset between the rotation axis and the detector. This brings systematic errors to the three-dimensional imaging results, and also requires fine adjustments to the CT system during the production process, which greatly increases the manufacturing cost of the product.

[0005] Patent CN102488528B proposes a cone-beam CT parameter calibration method based on global optimization. It reconstructs the original projection image and the detector's horizontal and vertical rotation images around the receiving plate, respectively. It is believed that when the rotation angle is accurate, the two reconstruction effects are the same. In this way, the optimization objective function is constructed to solve the tilt angle and roll angle of the detector, the distance from the rotation axis to the detector, and the distance from the ray source to the detector. However, this method has some defects. First, the perspective problem of cone-beam CT is not taken into account: when the detector is rotated, the distance from the pixel on it to the source will change, which will cause imaging distortion; secondly, this method uses a simulated annealing algorithm for solution. The algorithm is prone to local convergence when there are too many parameters, and the solution process is slow.

[0006] Patent CN118298019A proposes a planar CT parameter calibration method based on feature point projection. This method uses the feature points extracted from each frame of the projection image when it rotates a circle to calculate the model parameters, including the inclination angle of the rotation axis, the rotation angle in the receiver plane, the center of the detector, the distance from the light source to the receiver, and the distance from the light source to the rotation axis. Specifically, the trajectory of the extracted projection points in rotation is calculated, and the parameters to be calibrated are used as parameters to fit the captured projection trajectory and the ideal projection trajectory, and the solution is obtained by iterative optimization. However, in actual calculations, this method needs to construct an ideal projection trajectory, that is, it has high requirements on the position and accuracy of the object; and it includes all parameters in the optimization range, and there are problems such as slow convergence and poor calculation accuracy; on the other hand, the physical model it constructs does not take into account the offset of the inclination angle of the rotation axis, which is inconsistent with the actual situation. Summary of the invention

[0007] The present invention provides a method and device for calibrating parameters of a planar CT system under non-ideal conditions, which can overcome the influence of mechanical errors of the planar CT system under non-ideal conditions, improve the accuracy of three-dimensional imaging results, and reduce production costs.

[0008] A method for calibrating parameters of a planar CT system under non-ideal conditions, wherein the planar CT system includes a light source, a rotating platform, and a detector; the method includes:

[0009] Establishing a physical model of a CT imaging platform according to the positional relationship among the light source, the rotating platform and the detector;

[0010] Providing a calibration plate, wherein the calibration plate includes a plurality of spheres arranged in an array;

[0011] Establishing a relationship matrix according to the physical model of the CT imaging platform;

[0012] The calibration plate is set at any position between the light source and the detector, a first calibration image of the calibration plate is obtained through the detector, and the first sphere center physical coordinates of the plurality of spheres in the calibration plate coordinate system are obtained, an intrinsic parameter matrix in the relationship matrix is ​​calculated according to the first calibration image and the first sphere center physical coordinates, and position parameters of the light source and the detector are obtained;

[0013] The calibration plate is placed on the rotating platform and rotated for one circle, and the second calibration plate image at multiple angles is obtained by the detector, and the physical coordinates of the second sphere center at the corresponding angles are obtained, and the rotation transformation matrix and the translation transformation matrix in the relationship matrix are calculated according to the second calibration plate image, the second sphere center physical coordinates and the intrinsic parameter matrix;

[0014] The position parameters of the rotation axis in the rotating platform are calculated according to the rotation transformation matrix and the translation transformation matrix.

[0015] Furthermore, the calibration plate includes a first plate body and a second plate body arranged in parallel, and the first plate body and the second plate body are respectively provided with hemispherical grooves in an array for accommodating the sphere, the sphere is a metal sphere, and the first plate body and the second plate body are made of plastic.

[0016] Furthermore, according to the positional relationship among the light source, the rotating platform and the detector, a physical model of the CT imaging platform is established, including:

[0017] With the light source as the origin, a light source coordinate system is established, so that the X-axis of the light source coordinate system is parallel to the horizontal pixel vector of the detector, the Y-axis is parallel to the vertical pixel vector of the detector, and the Z-axis is the perpendicular line from the light source to the receiving plane of the detector, and the position parameters of the rotation axis in the rotating platform are represented in the light source coordinate system.

[0018] Furthermore, the relationship matrix is ​​the relationship between the detector pixel coordinates, the three-dimensional coordinates of any point in space in the object coordinate system, the rotation transformation matrix from the object coordinate system to the light source coordinate system, the translation transformation matrix from the object coordinate system to the light source coordinate system, and the intrinsic parameter matrix.

[0019] Further, obtaining the first sphere center physical coordinates of the plurality of spheres in the calibration plate coordinate system, and calculating the internal parameter matrix in the relationship matrix according to the first calibration image and the first sphere center physical coordinates, comprises:

[0020] The calibration plate coordinate system is constructed by taking the center of the sphere at the upper left vertex of the calibration plate as the origin and making the X-axis and Y-axis, wherein the X-axis is parallel to one side of the calibration plate and the Y-axis is parallel to the other side of the calibration plate;

[0021] According to the spherical center distances in the calibration plate, obtaining the physical coordinates of the first spherical center of the plurality of spherical centers in the coordinate system of the calibration plate;

[0022] According to the first calibration image, first spherical center pixel coordinates of a plurality of spherical centers in the first calibration image are acquired and normalized to obtain first pixel normalized coordinates;

[0023] At least four groups of the first sphere center physical coordinates and the first pixel normalized coordinates are obtained at different positions of the calibration plate between the light source and the detector, and the coordinates are substituted into the relationship matrix to calculate and obtain the closed solution of the intrinsic parameter matrix.

[0024] Further, obtaining the physical coordinates of the second sphere center at the corresponding angle, and calculating the rotation transformation matrix and the translation transformation matrix in the relationship matrix according to the second calibration plate image, the physical coordinates of the second sphere center and the intrinsic parameter matrix, including:

[0025] For the second calibration image at each angle, obtaining the second spherical center physical coordinates of multiple spherical centers in the calibration plate coordinate system according to the spherical center distances in the calibration plate;

[0026] According to the second calibration image, obtaining second spherical center pixel coordinates of a plurality of spherical centers in the second calibration image and performing normalization processing to obtain second pixel normalized coordinates;

[0027] Substituting the internal parameter matrix, the physical coordinates of the second sphere center obtained at multiple angles, and the second pixel normalized coordinates into the relationship matrix and performing equation transformation and vector form conversion to obtain a vector equation group;

[0028] The vector equations are solved based on the singular value decomposition method to obtain the rotation transformation matrix and the translation transformation matrix at different angles.

[0029] Further, calculating the position parameters of the rotation axis in the rotating platform according to the rotation transformation matrix and the translation transformation matrix includes:

[0030] According to the rotation transformation matrices at two different angles, the relative rotation matrices of the two angles are calculated;

[0031] According to the relative rotation matrix, the rotation vector of the calibration plate relative to the rotation axis at two angles is calculated based on the Rodriguez formula;

[0032] Calculate the actual rotation axis vector representation according to the rotation vector;

[0033] Selecting feature points that are located on both the calibration plate and the rotation axis;

[0034] Calculate the linear equation of the rotation axis in the light source coordinate system according to the feature points, the actual rotation axis vector representation, the rotation transformation matrix and the translation transformation matrix;

[0035] According to the linear equation of the rotation axis, the position parameters of the rotation axis are calculated.

[0036] Furthermore, according to the feature points, the actual rotation axis vector representation, the rotation transformation matrix and the translation transformation matrix, the rotation axis straight line equation in the light source coordinate system is calculated, including:

[0037] According to the rotation transformation matrix and translation transformation matrix at two different angles, the angle transformation and translation relationship equation is established;

[0038] Establish multiple sets of angle transformation and translation relationship equations at different angles;

[0039] The singular value decomposition method is used to solve the angle transformation and translation relationship equations at multiple different angles to calculate the coordinates of the feature points in the calibration plate coordinate system;

[0040] According to the relationship matrix, the coordinates of the feature points in the light source coordinate system are calculated;

[0041] According to the coordinates of the feature point in the light source coordinate system, the translation transformation matrix and the actual rotation axis vector representation, the rotation axis straight line equation in the light source coordinate system is obtained.

[0042] Furthermore, the position parameters of the rotation axis include the Z coordinate of the intersection of the XZ plane and the rotation axis when y=0 in the light source coordinate system, the distance from the intersection of the XZ plane and the rotation axis to the Z axis, the angle between the rotation axis and the Z axis, and the angle between the rotation axis and the X axis. 。

[0043] A CT system parameter calibration device applied to non-ideal conditions as in the above method, comprising:

[0044] A model building module, used to build a physical model of the CT imaging platform according to the positional relationship among the light source, the rotating platform and the detector;

[0045] A relationship matrix determination module, used to establish a relationship matrix according to the physical model of the CT imaging platform;

[0046] A first image acquisition module is used to set the calibration plate at any position between the light source and the detector, acquire a first calibration image of the calibration plate through the detector, and acquire the first spherical center physical coordinates of the plurality of spheres in the calibration plate coordinate system, calculate the intrinsic parameter matrix in the relationship matrix according to the first calibration image and the first spherical center physical coordinates, and obtain the position parameters of the light source and the detector;

[0047] A second image acquisition module is used to place the calibration plate on the rotating platform and rotate it for one circle, obtain the second calibration plate image at multiple angles through the detector, and obtain the physical coordinates of the second sphere center at the corresponding angles, and calculate the rotation transformation matrix and the translation transformation matrix in the relationship matrix according to the second calibration plate image, the second sphere center physical coordinates and the intrinsic parameter matrix;

[0048] A calculation module, used for calculating the position parameters of the rotation axis in the rotating platform according to the rotation transformation matrix and the translation transformation matrix;

[0049] Wherein, the calibration plate includes a plurality of spheres arranged in an array.

[0050] Furthermore, the calibration plate includes a first plate body and a second plate body arranged in parallel, and the first plate body and the second plate body are respectively provided with hemispherical grooves in an array for accommodating the sphere, the sphere is a metal sphere, and the first plate body and the second plate body are made of plastic.

[0051] Furthermore, the model building module builds a physical model of the CT imaging platform according to the positional relationship among the light source, the rotating platform and the detector, including:

[0052] With the light source as the origin, a light source coordinate system is established, so that the X-axis of the light source coordinate system is parallel to the horizontal pixel vector of the detector, the Y-axis is parallel to the vertical pixel vector of the detector, and the Z-axis is the perpendicular line from the light source to the receiving plane of the detector, and the position parameters of the rotation axis in the rotating platform are represented in the light source coordinate system.

[0053] Furthermore, the relationship matrix is ​​the relationship between the detector pixel coordinates, the three-dimensional coordinates of any point in space in the object coordinate system, the rotation transformation matrix from the object coordinate system to the light source coordinate system, the translation transformation matrix from the object coordinate system to the light source coordinate system, and the intrinsic parameter matrix.

[0054] Furthermore, the first image acquisition module acquires the physical coordinates of the first sphere centers of the plurality of spheres in the calibration plate coordinate system, and calculates the internal parameter matrix in the relationship matrix according to the first calibration image and the first sphere center physical coordinates, including:

[0055] The calibration plate coordinate system is constructed by taking the center of the sphere at the upper left vertex of the calibration plate as the origin and making the X-axis and Y-axis, wherein the X-axis is parallel to one side of the calibration plate and the Y-axis is parallel to the other side of the calibration plate;

[0056] According to the spherical center distances in the calibration plate, obtaining the physical coordinates of the first spherical center of the plurality of spherical centers in the coordinate system of the calibration plate;

[0057] According to the first calibration image, first spherical center pixel coordinates of a plurality of spherical centers in the first calibration image are acquired and normalized to obtain first pixel normalized coordinates;

[0058] At least four groups of the first sphere center physical coordinates and the first pixel normalized coordinates are obtained at different positions of the calibration plate between the light source and the detector, and the coordinates are substituted into the relationship matrix to calculate and obtain the closed solution of the intrinsic parameter matrix.

[0059] Furthermore, the second image acquisition module acquires the physical coordinates of the second sphere center at the corresponding angle, and calculates the rotation transformation matrix and the translation transformation matrix in the relationship matrix according to the second calibration plate image, the physical coordinates of the second sphere center and the intrinsic parameter matrix, including:

[0060] According to the spherical center distances in the calibration plate, obtaining the second spherical center physical coordinates of the plurality of spherical centers in the calibration plate coordinate system;

[0061] According to the second calibration image, obtaining second spherical center pixel coordinates of a plurality of spherical centers in the second calibration image and performing normalization processing to obtain second pixel normalized coordinates;

[0062] Substituting the internal parameter matrix, the physical coordinates of the second sphere center obtained at multiple angles, and the second pixel normalized coordinates into the relationship matrix and performing equation transformation and vector form conversion to obtain a vector equation group;

[0063] The vector equations are solved based on the singular value decomposition method to obtain the rotation transformation matrix and the translation transformation matrix at different angles.

[0064] Furthermore, the calculation module calculates the position parameters of the rotation axis in the rotating platform according to the rotation transformation matrix and the translation transformation matrix, including:

[0065] According to the rotation transformation matrices at two different angles, the relative rotation matrices of the two angles are calculated;

[0066] According to the relative rotation matrix, the rotation vector of the calibration plate relative to the rotation axis at two angles is calculated based on the Rodriguez formula;

[0067] Calculate the actual rotation axis vector representation according to the rotation vector;

[0068] Selecting feature points that are located on both the calibration plate and the rotation axis;

[0069] Calculate the linear equation of the rotation axis in the light source coordinate system according to the feature points, the actual rotation axis vector representation, the rotation transformation matrix and the translation transformation matrix;

[0070] According to the linear equation of the rotation axis, the position parameters of the rotation axis are calculated.

[0071] Furthermore, the calculation module calculates the linear equation of the rotation axis in the light source coordinate system according to the feature point, the actual rotation axis vector representation, the rotation transformation matrix and the translation transformation matrix, including:

[0072] According to the rotation transformation matrix and translation transformation matrix at two different angles, the angle transformation and translation relationship equation is established;

[0073] Establish multiple sets of angle transformation and translation relationship equations at different angles;

[0074] The singular value decomposition method is used to solve the angle transformation and translation relationship equations at multiple different angles to calculate the coordinates of the feature points in the calibration plate coordinate system;

[0075] According to the relationship matrix, the coordinates of the feature points in the light source coordinate system are calculated;

[0076] According to the coordinates of the feature point in the light source coordinate system, the translation transformation matrix and the actual rotation axis vector representation, the rotation axis straight line equation in the light source coordinate system is obtained.

[0077] Furthermore, the position parameters of the rotation axis include, in the light source coordinate system, when y=0, the Z coordinate of the intersection of the XZ plane and the rotation axis, the distance from the intersection of the XZ plane and the rotation axis to the Z axis, the angle between the rotation axis and the Z axis, and the angle between the rotation axis and the X axis. 。

[0078] The method and device for calibrating planar CT system parameters under non-ideal conditions provided by the present invention have at least the following beneficial effects:

[0079] (1) A physical model of planar CT under non-ideal conditions was constructed, and a special calibration plate was designed. The vertical offset and horizontal offset of the detection plate and the position parameters of the rotation axis were accurately calculated to overcome the influence of mechanical errors of the planar CT system under non-ideal conditions, improve the accuracy of three-dimensional imaging results, and reduce production costs.

[0080] (2) During the calculation process, the light source is used as the origin of the coordinate system, the direction of the detector is used to construct the axis direction of the coordinate system, and the rotation axis is a straight line in the coordinate system. Compared with the conventional method of establishing coordinates on the rotation axis, the coordinate system in this method does not move, is easy to solve, and reduces the amount of calculation. BRIEF DESCRIPTION OF THE DRAWINGS

[0081] Figure 1 The present invention provides a flow chart of an embodiment of a method for calibrating planar CT system parameters under non-ideal conditions.

[0082] Figure 2The present invention is a schematic structural diagram of an embodiment of a physical model of a CT imaging platform in a method for calibrating planar CT system parameters under non-ideal conditions provided by the present invention.

[0083] Figure 3 The present invention is a schematic structural diagram of an embodiment of a calibration plate in the planar CT system parameter calibration method under non-ideal conditions provided by the present invention.

[0084] Figure 4 The present invention provides a flowchart of an embodiment of calculating the internal parameter matrix in the planar CT system parameter calibration method under non-ideal conditions provided by the present invention.

[0085] Figure 5 The present invention provides a flowchart of an embodiment of calculating the rotation transformation matrix and the translation transformation matrix in the planar CT system parameter calibration method under non-ideal conditions provided by the present invention.

[0086] Figure 6 The present invention provides a flowchart of an embodiment of calculating position parameters in a planar CT system parameter calibration method under non-ideal conditions provided by the present invention.

[0087] Figure 7 The present invention is a schematic structural diagram of an embodiment of a planar CT system parameter calibration device under non-ideal conditions provided by the present invention. DETAILED DESCRIPTION

[0088] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0089] refer to Figure 1 In some embodiments, a method for calibrating parameters of a planar CT system under non-ideal conditions is provided, wherein the planar CT system includes a light source, a rotating platform, and a detector; the method includes:

[0090] S1. Establishing a physical model of a CT imaging platform according to the positional relationship among the light source, the rotating platform and the detector;

[0091] S2. Setting a calibration plate, wherein the calibration plate includes a plurality of spheres arranged in an array;

[0092] S3, establishing a relationship matrix according to the physical model of the CT imaging platform;

[0093] S4, setting the calibration plate at any position between the light source and the detector, obtaining a first calibration image of the calibration plate through the detector, and obtaining the first sphere center physical coordinates of the plurality of spheres in the calibration plate coordinate system, calculating the intrinsic parameter matrix in the relationship matrix according to the first calibration image and the first sphere center physical coordinates, and obtaining the position parameters of the light source and the detector;

[0094] S5, placing the calibration plate on the rotating platform and rotating it for one circle, obtaining the second calibration plate images at multiple angles through the detector, and obtaining the second sphere center physical coordinates of the multiple spheres in the calibration plate coordinate system at corresponding angles, and calculating the rotation transformation matrix and the translation transformation matrix in the relationship matrix according to the second calibration plate image, the second sphere center physical coordinates and the intrinsic parameter matrix;

[0095] S6. Calculate the position parameters of the rotation axis in the rotating platform according to the rotation transformation matrix and the translation transformation matrix.

[0096] Specifically, in step S1, the planar CT system mainly includes a light source, a rotating platform, and a detector. When shooting, the object to be detected is placed on the rotating platform, and the detector is used to capture images at each rotation angle.

[0097] In this embodiment, the following Figure 2 The physical model of the CT imaging platform shown in the figure, where 1 is the light source, 2 is the rotation axis of the rotating platform, and 3 is the detector. In the physical model of the CT imaging platform, there is a certain offset between the midpoint of the detector 3 and the vertical line from the light source 1 to the detector 3. The light source 1 is taken as the origin S, the plane of the detector 3 is I, and the straight line where the rotation axis 2 is located is O. The light source is used to construct a light source coordinate system S-XYZ, where the X-axis is parallel to the horizontal pixel vector U of the detector 3, the Y-axis is parallel to the vertical pixel vector V of the detector 3, and the Z-axis is the vertical line from the origin S to the receiving plane I of the detector 3.

[0098] The position parameters of the rotation axis of the rotating platform in the light source coordinate system, including:

[0099] The distance between the intersection of the perpendicular line from the origin S to the receiving plane I of the detector 3 and the plane I and the perpendicular midline of the receiving plane I of the detector 3 is d x , the distance between the intersection point and the horizontal midline of the receiving plane I of the detector 3 is d y ; SID (Source Image Distacne) is the length of the perpendicular line from the origin S to the receiving plane I of the detector 3; considering the offset of the rotation axis, the position of the rotation axis 2 is expressed in the light source coordinate system. When Y=0, SOD (Source Image Distacne) is the Z coordinate of the intersection of the XZ plane and the rotation axis. When , the distance from the intersection of the XZ plane and the rotation axis 2 to the Z axis is recorded as d a , the angle between the rotation axis 2 and the Z axis is recorded as , the angle between the rotation axis and the X axis is .

[0100] Further, in step S2, the calibration plate includes a first plate body and a second plate body arranged in parallel, and the first plate body and the second plate body are respectively provided with hemispherical grooves in an array for accommodating the sphere, the sphere is a metal sphere, and the first plate body and the second plate body are made of plastic.

[0101] Specifically, unlike camera photography, the planar CT system cannot capture the color information of the object surface. The image captured by the planar CT system is only affected by the material, density and shape of the object. Therefore, the conventional planar calibration plate cannot be used. In this embodiment, a planar calibration plate is designed as follows: Figure 3 The calibration plate shown includes a first plate body and a second plate body arranged in parallel, and the first plate body and the second plate body are respectively provided with hemispherical grooves for accommodating the spheres in an array, and the material thereof is low-density engineering plastic. In some embodiments, hemispherical holes are neatly arranged in 8 rows and 10 columns on the first plate body and the second plate body. When used, metal balls of equal diameter are placed in the holes, and the second calibration plate is covered on the surface of the first plate body. When photographed using a CT system, the ball will appear round on the detector due to the density difference between the metal material and the plastic material.

[0102] Further, in step S3, the relationship matrix is ​​the relationship between the detector pixel coordinates, the three-dimensional coordinates of any point in the space in the object coordinate system, the rotation transformation matrix from the object coordinate system to the light source coordinate system, the translation transformation matrix between the object coordinate system and the light source coordinate system, and the intrinsic parameter matrix.

[0103] Specifically, to solve the position parameters in the above model, for the three-dimensional coordinates of any point in space Since the planar CT system is consistent with the perspective model, its coordinates are consistent with the detector pixel coordinates. The following corresponding relationships exist:

[0104] ; (1)

[0105] in, , represents the three-dimensional coordinates of any point in the object coordinate system, , represents the detector pixel coordinates, R is the rotation transformation matrix from the object coordinate system to the light source coordinate system, represents the rotation transformation from the object coordinate system to the light source coordinate system; t is the translation transformation matrix from the object coordinate system to the light source coordinate system, f It represents the focal length from the light source to the detector, that is, the vertical distance from the light source to the detector; dX Indicates the actual spatial distance corresponding to the pixel in the horizontal direction; d Y Indicates the actual spatial distance corresponding to the pixel in the vertical direction; c x Indicates the principal point offset of the detector in the horizontal direction; cy Indicates the principal point offset of the detector in the vertical direction; Represents the normalization coefficient in the light source coordinate system.

[0106] Then the relationship matrix is ​​solved. In step S4, the calibration plate is set at any position between the light source 1 and the detector 3 so that the detector 3 can capture a complete image of the calibration plate. Multiple first calibration images of the calibration plate are obtained through the detector 3. Figure 4 , obtaining the first sphere center physical coordinates of the plurality of spheres in the calibration plate coordinate system, and calculating the internal parameter matrix in the relationship matrix according to the first calibration image and the first sphere center physical coordinates, specifically comprising:

[0107] S41, taking the center of the sphere at the upper left vertex of the calibration plate as the origin and making the X-axis and the Y-axis, wherein the X-axis is parallel to one side of the calibration plate and the Y-axis is parallel to the other side of the calibration plate, to construct a calibration plate coordinate system;

[0108] S42, obtaining the physical coordinates of the first sphere center of multiple sphere centers in the calibration plate coordinate system according to the sphere center distance in the calibration plate;

[0109] S43, according to the first calibration image, obtaining first sphere center pixel coordinates of multiple sphere centers in the first calibration image and performing normalization processing to obtain first pixel normalized coordinates;

[0110] S44, obtaining at least four groups of the first sphere center physical coordinates and the first pixel normalized coordinates at different positions of the calibration plate between the light source and the detector, substituting them into the relationship matrix, and calculating to obtain the closed solution of the intrinsic parameter matrix.

[0111] Specifically, in step S41, the calibration plate is placed at any position between the light source and the detector so that the detector can capture the complete calibration plate. The center of the metal sphere is extracted from the captured image, and the pixel coordinates are sorted by rows and columns. Since the calibration plate is a standard array, a calibration plate coordinate system is constructed using the calibration plate as a coordinate system, with the center of the sphere in the upper left corner as the coordinate origin, one side as the X-axis, the other side as the Y-axis, and the Z-axis perpendicular to the X-axis and the Y-axis. In some embodiments, the array is 8 rows and 10 columns, and the long side of the calibration plate can be the X-axis and the short side as the Y-axis.

[0112] In step S42, the ball is located at the first m Row, No. n The distance between the sphere centers is l , then the three-dimensional coordinates of the center of the sphere are After obtaining the three-dimensional coordinates of each sphere center, the physical coordinates of the first sphere center of multiple sphere centers in the calibration plate coordinate system can be obtained.

[0113] In step S43, the first spherical center pixel coordinates of each spherical center are obtained according to the first calibration image acquired by the detector, and the corresponding pixel normalized coordinates are , and then obtain the normalized coordinates of the first pixel.

[0114] Furthermore, in step S44, since the Z coordinate in the physical coordinates of the first sphere center is always 0, equation (1) is transformed into:

[0115] ; (2)

[0116] Among them, the internal parameter matrix , r 1 and r 2 Represents the rotation transformation matrix The first column vector and the second column vector of .

[0117] Then the matrix H is as follows:

[0118] ; (3)

[0119] is a A matrix with one element being a homogeneous coordinate and a constraint modulus of 1, that is, the matrix There are 8 unknown elements, and due to the rotation transformation matrix is a unit orthogonal matrix, then:

[0120] ; (4)

[0121] ; (5)

[0122] in, and are the rotation transformation matrices The transpose of the first and second column vectors of .

[0123] Substituting it into the matrix The following relationship can be obtained:

[0124] ; (6)

[0125] ; (7)

[0126] Among them, h 1 and h 2 denote the first column vector and the second column vector of the matrix H respectively, and are the transposes of the first and second column vectors of the matrix H, respectively. represents the inverse of the internal parameter matrix K, represents the inverse of the transpose of the intrinsic parameter matrix K.

[0127] Internal parameter matrix Contains 5 elements, so when more than 3 sets of first sphere center physical coordinates and first pixel normalized coordinates are taken, they are substituted into the above relationship. The first sphere center physical coordinates are the coordinates in the object coordinate system, and the first pixel normalized coordinates are the detector pixel coordinates. The intrinsic parameter matrix can be obtained. The closed solution of .

[0128] Specifically, the position parameters of the light source and the detector are further obtained: the internal parameter matrix In and is the actual spatial distance corresponding to the detector pixel in the horizontal and vertical directions, so the focal length can be calculated ,Right now Figure 2 The length SID of the perpendicular line from the origin S in the detector to the receiving plane I of the detector 3, and the internal parameter matrix In and is the pixel coordinate of the vertical projection point of the light source on the detector. Its horizontal and vertical distances to the center of the detector can be calculated by multiplying the pixel difference by the actual spatial size of the pixel.

[0129] Furthermore, in step S5, after obtaining the internal reference matrix in the relationship matrix, the calibration plate is placed on the rotating platform and rotated for one circle, and the second calibration plate image at multiple angles is obtained through the detector, and the physical coordinates of the second sphere center at the corresponding angles are obtained.

[0130] refer to Figure 5 , obtaining the physical coordinates of the second sphere center at the corresponding angle, and calculating the rotation transformation matrix and the translation transformation matrix in the relationship matrix according to the second calibration plate image, the physical coordinates of the second sphere center and the intrinsic parameter matrix, including:

[0131] S51, for each second calibration image at each angle, obtaining the second spherical center physical coordinates of multiple spherical centers in the calibration plate coordinate system according to the spherical center distance in the calibration plate;

[0132] S52, according to the second calibration image, obtaining second sphere center pixel coordinates of multiple sphere centers in the second calibration image and performing normalization processing to obtain second pixel normalized coordinates;

[0133] S53, substituting the internal parameter matrix, the physical coordinates of the second sphere center obtained at multiple angles, and the second pixel normalized coordinates into the relationship matrix and performing equation transformation and vector form conversion to obtain a vector equation group;

[0134] S54. Solve the vector equation group based on the singular value decomposition method to obtain the rotation transformation matrix and the translation transformation matrix at different angles.

[0135] Specifically, after calculating the internal parameter matrix After that, the coordinates of the rotation axis in the light source coordinate system are calculated. The rotation axis in the light source coordinate system is the rotation axis of the calibration plate relative to the light source coordinate system. Therefore, in the embodiment, the rotation transformation matrix and translation transformation matrix of the calibration plate in the light source coordinate system are first obtained, and then the relative rotation between them is calculated. The calibration plate is placed on the rotating platform and rotated for one circle. The detector is used to obtain the calibration plate images at multiple angles. Still in the calibration plate coordinate system, the first m Row, No. n The three-dimensional coordinates of each sphere center on the calibration plate are , the physical coordinates of the second sphere center are .

[0136] In step S52, according to the second calibration image, the pixel coordinates of the second sphere centers of the plurality of sphere centers in the second calibration image are obtained and normalized to obtain , and then obtain the normalized coordinates of the second pixel.

[0137] In step S53, the second sphere center physical coordinate matrix and the second pixel normalized coordinate matrix are substituted into formula (2) to obtain:

[0138] ; (8)

[0139] Among them, the internal parameter matrix It has been found that the rotation transformation matrix R and the translation transformation matrix t are the quantities to be solved. Expand the above formula, where for The matrix, each element of which is f nm Indicates that we get:

[0140] ; (9)

[0141] Transformed into a system of equations, we get:

[0142] ; (10)

[0143] Convert the quantity to be solved into vector form to obtain the vector equation:

[0144] ; (11)

[0145] When there are more than six sets of the above vector equations, the singular value decomposition (SVD) method can be used to directly solve the least squares solution of the vector equations to obtain:

[0146] ; (12)

[0147] Right now: ; (13)

[0148] So the calculation expression of the rotation transformation matrix R is:

[0149] ; (14)

[0150] The calculation expression of the translation transformation matrix is:

[0151] ; (15)

[0152] Specifically, refer to Figure 6 In step S6, the position parameters of the rotation axis in the rotating platform are calculated according to the rotation transformation matrix and the translation transformation matrix, including:

[0153] S61, calculating and obtaining relative rotation matrices of two angles according to the rotation transformation matrices at two different angles;

[0154] S62, calculating the rotation vector of the calibration plate relative to the rotation axis at two angles based on the relative rotation matrix and the Rodriguez formula;

[0155] S63, calculating the actual rotation axis vector representation according to the rotation vector;

[0156] S64, selecting feature points located on both the calibration plate and the rotation axis;

[0157] S65, calculating the linear equation of the rotation axis in the light source coordinate system according to the feature point, the actual rotation axis vector representation, the rotation transformation matrix and the translation transformation matrix;

[0158] S66. Calculate the position parameters of the rotation axis according to the linear equation of the rotation axis.

[0159] Specifically, in step S61, the rotation and translation matrix in the light source coordinate system corresponding to the two-dimensional image captured at each angle is calculated, and the first i The rotation transformation matrix under angle is R i , the translation transformation matrix is ​​t i Then there is the i The calibration plate at the angle is relative to the j The relative rotation matrix R of the calibration plate at angles ij for:

[0160] : (16)

[0161] in, For thei The rotation transformation matrix under angle is The transpose of .

[0162] Further, in step S62, according to the plurality of sets of relative rotation matrices, the rotation vector of the calibration plate relative to the rotation axis at two angles is calculated based on the Rodriguez formula, specifically:

[0163] ; (17)

[0164] ; (18)

[0165] Among them, tr (R ij ) represents the relative rotation matrix R ij The trace of the identity matrix, tr(I), ij represents the rotation vector, θ Indicates the rotation angle of the calibration plate relative to the rotation axis at two angles. Represents the trace of the product of a rotation vector and its transpose.

[0166] Furthermore, the rotation vector is expressed as ,satisfy:

[0167] ; (19)

[0168] According to the rotation angles relative to the rotation axis at two angles, the rotation vector can be calculated r ij , the rotation vector is expressed as the vector representation of the rotation axis in the light source coordinate system.

[0169] In step S63, after calculating and obtaining a number of rotation vectors, in order to improve the solution accuracy, the least square solution is solved to obtain the actual rotation axis vector representation, and the actual rotation axis vector representation r satisfies the linear equation group:

[0170] ; (20)

[0171] Rotate vector when cross-producting The antisymmetric matrix can be used instead, and the linear equations can be expressed as:

[0172] ;(twenty one)

[0173] The actual rotation axis vector r can be calculated using the singular value decomposition (SVD) method.

[0174] Next, calculate the coordinates of a point on the rotation axis in the light source coordinate system.

[0175] Specifically, in step S64 and step S65, there is a point on the calibration plate, and its coordinates after transformation to the light source coordinate system are still the same, because this point is located on the rotation axis, and its coordinates in the calibration plate coordinate system are also the same. Then this point is recorded as a feature point, and the coordinates of the feature point in the calibration plate coordinate system are , the corresponding coordinates of the feature point in the light source coordinate system are , satisfying the rotation transformation matrix and translation transformation matrix calculated in equation (8), the established angle transformation and translation relationship equation is as follows:

[0176] ;(twenty two)

[0177] Among them, R i and R j Respectively represent the rotation angle i and j The rotation transformation matrix is: is the coordinate of the feature point in the calibration plate coordinate system, t i and t j Respectively represent the rotation angle i and j The translation transformation matrix below.

[0178] Establish multiple sets of angle transformation and translation relationship equations at different angles. The singular value decomposition method can also be used to solve the coordinate transformation and translation relationship equation to obtain the coordinates of the feature points in the calibration plate coordinate system. , and then get the coordinates O of the feature point in the light source coordinate system c Then the linear equation of the rotation axis O in the light source coordinate system can be expressed as:

[0179] ;(twenty three)

[0180] Among them, O represents the linear equation of the rotation axis, t represents the translation transformation matrix, and O c Represents the coordinates of the feature point in the light source coordinate system, and r is the actual rotation axis vector.

[0181] Further, in step S66, the position parameters of the rotation axis are calculated according to the linear equation of the rotation axis, referring to Figure 2 The position parameters include the Z coordinate SOD of the intersection of the XZ plane and the rotation axis when Y=0. When the XZ plane intersects the rotation axis 2, the distance from the Z axis is d a , the angle between the rotation axis 2 and the Z axis , the angle between the rotation axis and the X axis .

[0182] Specifically, when y=0, the distance from the intersection of the XZ plane and the rotation axis 2 to the Z axis isd a is the equation of the straight line of the rotation axis Intersection with XZ plane Coordinates, SOD is the Z coordinate of the intersection of the XZ plane and the linear equation of the rotation axis O, and the angle between the rotation axis 2 and the Z axis , the angle between the rotation axis and the X axis .

[0183] In some embodiments, reference Figure 7 , a CT system parameter calibration device under non-ideal conditions applied to the above method is provided, comprising:

[0184] A model building module 201 is used to build a physical model of the CT imaging platform according to the positional relationship among the light source, the rotating platform and the detector;

[0185] A relationship matrix determination module 202 is used to establish a relationship matrix according to the physical model of the CT imaging platform;

[0186] The first image acquisition module 203 is used to set the calibration plate at any position between the light source and the detector, obtain the first calibration image of the calibration plate through the detector, and obtain the first sphere center physical coordinates of the plurality of spheres in the calibration plate coordinate system, calculate the internal parameter matrix in the relationship matrix according to the first calibration image and the first sphere center physical coordinates, and obtain the position parameters of the light source and the detector;

[0187] The second image acquisition module 204 is used to place the calibration plate on the rotating platform and rotate it for one circle, obtain the second calibration plate images at multiple angles through the detector, and obtain the physical coordinates of the second sphere center at the corresponding angles, and calculate the rotation transformation matrix and the translation transformation matrix in the relationship matrix according to the second calibration plate image, the second sphere center physical coordinates and the intrinsic parameter matrix;

[0188] A calculation module 205, configured to calculate position parameters of a rotation axis in the rotating platform according to the rotation transformation matrix and the translation transformation matrix;

[0189] Wherein, the calibration plate includes a plurality of spheres arranged in an array.

[0190] Furthermore, the calibration plate includes a first plate body and a second plate body arranged in parallel, and the first plate body and the second plate body are respectively provided with hemispherical grooves in an array for accommodating the sphere, the sphere is a metal sphere, and the first plate body and the second plate body are made of plastic.

[0191] Furthermore, the model building module 201 builds a physical model of the CT imaging platform according to the positional relationship among the light source, the rotating platform and the detector, including:

[0192] With the light source as the origin, a light source coordinate system is established, so that the X-axis of the light source coordinate system is parallel to the horizontal pixel vector of the detector, the Y-axis is parallel to the vertical pixel vector of the detector, and the Z-axis is the perpendicular line from the light source to the receiving plane of the detector, and the position parameters of the rotation axis in the rotating platform are represented in the light source coordinate system.

[0193] Furthermore, the relationship matrix is ​​the relationship between the detector pixel coordinates, the three-dimensional coordinates of any point in space in the object coordinate system, the rotation transformation matrix from the object coordinate system to the light source coordinate system, the translation transformation matrix from the object coordinate system to the light source coordinate system, and the intrinsic parameter matrix.

[0194] Furthermore, the first image acquisition module 203 acquires the physical coordinates of the first sphere centers of the plurality of spheres in the calibration plate coordinate system, and calculates the internal parameter matrix in the relationship matrix according to the first calibration image and the first sphere center physical coordinates, including:

[0195] The calibration plate coordinate system is constructed by taking the center of the sphere at the upper left vertex of the calibration plate as the origin and making the X-axis and Y-axis, wherein the X-axis is parallel to one side of the calibration plate and the Y-axis is parallel to the other side of the calibration plate;

[0196] According to the spherical center distances in the calibration plate, obtaining the physical coordinates of the first spherical center of the plurality of spherical centers in the coordinate system of the calibration plate;

[0197] According to the first calibration image, first spherical center pixel coordinates of a plurality of spherical centers in the first calibration image are acquired and normalized to obtain first pixel normalized coordinates;

[0198] At least four groups of the first sphere center physical coordinates and the first pixel normalized coordinates are obtained at different positions of the calibration plate between the light source and the detector, and the coordinates are substituted into the relationship matrix to calculate and obtain the closed solution of the intrinsic parameter matrix.

[0199] Furthermore, the second image acquisition module 204 acquires the physical coordinates of the second sphere center at the corresponding angle, and calculates the rotation transformation matrix and the translation transformation matrix in the relationship matrix according to the second calibration plate image, the physical coordinates of the second sphere center and the intrinsic parameter matrix, including:

[0200] For the second calibration image at each angle, obtaining the second spherical center physical coordinates of multiple spherical centers in the calibration plate coordinate system according to the spherical center distances in the calibration plate;

[0201] According to the second calibration image, obtaining second spherical center pixel coordinates of a plurality of spherical centers in the second calibration image and performing normalization processing to obtain second pixel normalized coordinates;

[0202] Substituting the internal parameter matrix, the physical coordinates of the second sphere center obtained at multiple angles, and the second pixel normalized coordinates into the relationship matrix and performing equation transformation and vector form conversion to obtain a vector equation group;

[0203] The vector equations are solved based on the singular value decomposition method to obtain the rotation transformation matrix and the translation transformation matrix at different angles.

[0204] Furthermore, the calculation module 205 calculates the position parameters of the rotation axis in the rotating platform according to the rotation transformation matrix and the translation transformation matrix, including:

[0205] According to the rotation transformation matrices at two different angles, the relative rotation matrices of the two angles are calculated;

[0206] According to the relative rotation matrix, the rotation vector of the calibration plate relative to the rotation axis at two angles is calculated based on the Rodriguez formula;

[0207] Calculate the actual rotation axis vector representation according to the rotation vector;

[0208] Selecting feature points that are located on both the calibration plate and the rotation axis;

[0209] Calculate the linear equation of the rotation axis in the light source coordinate system according to the feature points, the actual rotation axis vector representation, the rotation transformation matrix and the translation transformation matrix;

[0210] According to the linear equation of the rotation axis, the position parameters of the rotation axis are calculated.

[0211] Furthermore, the calculation module 205 calculates the linear equation of the rotation axis in the light source coordinate system according to the feature points, the actual rotation axis vector representation, the rotation transformation matrix and the translation transformation matrix, including:

[0212] According to the rotation transformation matrix and translation transformation matrix at two different angles, the angle transformation and translation relationship equation is established;

[0213] Establish multiple sets of angle transformation and translation relationship equations at different angles;

[0214] The singular value decomposition method is used to solve the angle transformation and translation relationship equations at multiple different angles to calculate the coordinates of the feature points in the calibration plate coordinate system;

[0215] According to the relationship matrix, the coordinates of the feature points in the light source coordinate system are calculated;

[0216] According to the coordinates of the feature point in the light source coordinate system, the translation transformation matrix and the actual rotation axis vector representation, the rotation axis straight line equation in the light source coordinate system is obtained.

[0217] Furthermore, the position parameters of the rotation axis include, in the light source coordinate system, when y=0, the Z coordinate of the intersection of the XZ plane and the rotation axis, the distance from the intersection of the XZ plane and the rotation axis to the Z axis, the angle between the rotation axis and the Z axis, and the angle between the rotation axis and the X axis. 。

[0218] The method and device for calibrating planar CT system parameters under non-ideal conditions provided in the above embodiments have at least the following beneficial effects:

[0219] (1) A physical model of planar CT under non-ideal conditions was constructed, and a special calibration plate was designed. The vertical offset and horizontal offset of the detection plate and the position parameters of the rotation axis were accurately calculated to overcome the influence of mechanical errors of the planar CT system under non-ideal conditions, improve the accuracy of three-dimensional imaging results, and reduce production costs.

[0220] (2) During the calculation process, the light source is used as the origin of the coordinate system, the direction of the detector is used to construct the axis direction of the coordinate system, and the rotation axis is a straight line in the coordinate system. Compared with the conventional method of establishing coordinates on the rotation axis, the coordinate system in this method does not move, is easy to solve, and reduces the amount of calculation.

[0221] Although preferred embodiments of the present invention have been described, additional changes and modifications may be made to these embodiments by those skilled in the art once the basic inventive concepts are known. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention. Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for calibrating planar CT system parameters under non-ideal conditions, characterized in that: The planar CT system includes a light source, a rotating platform and a detector; the method includes: Establishing a physical model of a CT imaging platform according to the positional relationship among the light source, the rotating platform and the detector; Providing a calibration plate, wherein the calibration plate includes a plurality of spheres arranged in an array; Establishing a relationship matrix according to the physical model of the CT imaging platform; The calibration plate is set at any position between the light source and the detector, a first calibration image of the calibration plate is obtained through the detector, and the first sphere center physical coordinates of the plurality of spheres in the calibration plate coordinate system are obtained, an intrinsic parameter matrix in the relationship matrix is ​​calculated according to the first calibration image and the first sphere center physical coordinates, and position parameters of the light source and the detector are obtained; The calibration plate is placed on the rotating platform and rotated for one circle, and the second calibration plate image at multiple angles is obtained by the detector, and the physical coordinates of the second sphere center at the corresponding angles are obtained, and the rotation transformation matrix and the translation transformation matrix in the relationship matrix are calculated according to the second calibration plate image, the second sphere center physical coordinates and the intrinsic parameter matrix; Calculate the position parameters of the rotation axis in the rotating platform according to the rotation transformation matrix and the translation transformation matrix; According to the positional relationship among the light source, the rotating platform and the detector, a physical model of the CT imaging platform is established, including: Taking the light source as the origin, a light source coordinate system is established, so that the X axis of the light source coordinate system is parallel to the horizontal pixel vector of the detector, the Y axis is parallel to the vertical pixel vector of the detector, the Z axis is the perpendicular line from the light source to the receiving plane of the detector, and the position parameters of the rotation axis in the rotating platform are expressed in the light source coordinate system; Calculating the position parameters of the rotation axis in the rotating platform according to the rotation transformation matrix and the translation transformation matrix includes: According to the rotation transformation matrices at two different angles, the relative rotation matrices of the two angles are calculated; According to the relative rotation matrix, the rotation vector of the calibration plate relative to the rotation axis at two angles is calculated based on the Rodriguez formula; Calculate the actual rotation axis vector representation according to the rotation vector; Selecting feature points that are located on both the calibration plate and the rotation axis; Calculate the linear equation of the rotation axis in the light source coordinate system according to the feature points, the actual rotation axis vector representation, the rotation transformation matrix and the translation transformation matrix; Calculate the position parameters of the rotation axis according to the rotation axis straight line equation; According to the feature points, the actual rotation axis vector representation, the rotation transformation matrix and the translation transformation matrix, the rotation axis straight line equation in the light source coordinate system is calculated, including: According to the rotation transformation matrix and translation transformation matrix at two different angles, the angle transformation and translation relationship equation is established; Establish multiple sets of angle transformation and translation relationship equations at different angles; The singular value decomposition method is used to solve the angle transformation and translation relationship equations at multiple different angles to calculate the coordinates of the feature points in the calibration plate coordinate system; According to the relationship matrix, the coordinates of the feature points in the light source coordinate system are calculated; According to the coordinates of the feature point in the light source coordinate system, the translation transformation matrix and the actual rotation axis vector representation, the rotation axis straight line equation in the light source coordinate system is obtained.

2. The method according to claim 1, characterized in that The calibration plate includes a first plate body and a second plate body arranged in parallel, and the first plate body and the second plate body are respectively provided with hemispherical grooves in an array for accommodating the sphere. The sphere is a metal sphere, and the first plate body and the second plate body are made of plastic.

3. The method according to claim 1, characterized in that: The relationship matrix is ​​the relationship between the detector pixel coordinates, the three-dimensional coordinates of any point in space in the object coordinate system, the rotation transformation matrix from the object coordinate system to the light source coordinate system, the translation transformation matrix from the object coordinate system to the light source coordinate system, and the intrinsic parameter matrix.

4. The method according to claim 3, characterized in that Acquiring the physical coordinates of the first sphere centers of the plurality of spheres in the calibration plate coordinate system, and calculating the internal parameter matrix in the relationship matrix according to the first calibration image and the first sphere center physical coordinates, including: The calibration plate coordinate system is constructed by taking the center of the sphere at the upper left vertex of the calibration plate as the origin and making the X-axis and Y-axis, wherein the X-axis is parallel to one side of the calibration plate and the Y-axis is parallel to the other side of the calibration plate; According to the spherical center distances in the calibration plate, obtaining the physical coordinates of the first spherical center of the plurality of spherical centers in the coordinate system of the calibration plate; According to the first calibration image, first spherical center pixel coordinates of a plurality of spherical centers in the first calibration image are acquired and normalized to obtain first pixel normalized coordinates; At least four groups of the first sphere center physical coordinates and the first pixel normalized coordinates are obtained at different positions of the calibration plate between the light source and the detector, and the coordinates are substituted into the relationship matrix to calculate and obtain the closed solution of the intrinsic parameter matrix.

5. The method according to claim 4, characterized in that Obtaining the physical coordinates of the second sphere center at the corresponding angle, and calculating the rotation transformation matrix and the translation transformation matrix in the relationship matrix according to the second calibration plate image, the physical coordinates of the second sphere center and the intrinsic parameter matrix, including: For the second calibration image at each angle, obtaining the second spherical center physical coordinates of multiple spherical centers in the calibration plate coordinate system according to the spherical center distances in the calibration plate; According to the second calibration image, obtaining second spherical center pixel coordinates of a plurality of spherical centers in the second calibration image and performing normalization processing to obtain second pixel normalized coordinates; Substituting the internal parameter matrix, the physical coordinates of the second sphere center obtained at multiple angles, and the second pixel normalized coordinates into the relationship matrix and performing equation transformation and vector form conversion to obtain a vector equation group; The vector equations are solved based on the singular value decomposition method to obtain the rotation transformation matrix and the translation transformation matrix at different angles.

6. The method according to claim 1, characterized in that The position parameters of the rotation axis include the Z coordinate of the intersection of the XZ plane and the rotation axis, the distance from the intersection of the XZ plane and the rotation axis to the Z axis, the angle between the rotation axis and the Z axis, and the angle between the rotation axis and the X axis when y=0 in the light source coordinate system. 。 7. A CT system parameter calibration device applied to the method according to any one of claims 1 to 6 under non-ideal conditions, characterized in that: include: A model building module, used to build a physical model of the CT imaging platform according to the positional relationship among the light source, the rotating platform and the detector; A relationship matrix determination module, used to establish a relationship matrix according to the physical model of the CT imaging platform; A first image acquisition module is used to set the calibration plate at any position between the light source and the detector, acquire a first calibration image of the calibration plate through the detector, and acquire the first spherical center physical coordinates of the plurality of spheres in the calibration plate coordinate system, calculate the intrinsic parameter matrix in the relationship matrix according to the first calibration image and the first spherical center physical coordinates, and obtain the position parameters of the light source and the detector; A second image acquisition module is used to place the calibration plate on the rotating platform and rotate it for one circle, obtain the second calibration plate image at multiple angles through the detector, and obtain the physical coordinates of the second sphere center at the corresponding angles, and calculate the rotation transformation matrix and the translation transformation matrix in the relationship matrix according to the second calibration plate image, the second sphere center physical coordinates and the intrinsic parameter matrix; A calculation module, used for calculating the position parameters of the rotation axis in the rotating platform according to the rotation transformation matrix and the translation transformation matrix; Wherein, the calibration plate includes a plurality of spheres arranged in an array.

Images