A method for tilt calibration of a projected image

Abstract

The present invention discloses a method for tilt calibration of a projected image, comprising the steps of: (1) obtaining a projected image and obtaining, through calibration, the rotation angle of the actual plane where the current detector is located relative to the ideal plane where its ideal position is located; (2) calculating the rotation transformation matrix between the normal vector of the actual plane and the normal vector of the ideal plane according to the current rotation angle, and calculating and obtaining the plane equations of the ideal plane and the actual plane; (3) obtaining the coordinates of the corresponding feature points on the ideal plane for several feature points selected on the actual plane according to the perspective projection principle; (4) solving by using a linear equation to obtain the transformation relationship between the actual plane and the ideal plane, and calibrating the image accordingly to obtain a calibrated image. The present invention solves the problem of deformation of the projected image caused by the rotation and tilt of the actual detector relative to the ideal position, completes image calibration before reconstructing the projected image, ensures the imaging effect, and effectively eliminates artifacts.

Description

Technical Field

[0001] The present invention relates to the field of image processing, and particularly to a method for tilting calibration of a projection image. Background Art

[0002] Computed Tomography (CT) scans a specific part of the human body at a certain thickness layer by layer with X-rays. Since different human tissues have different absorption abilities for X-rays, the cross-sectional images can be reconstructed by a computer. CT is an important imaging method for obtaining the internal structure information of an object in a non-destructive manner, with many advantages such as high resolution, high sensitivity, and multiple layers, and is widely used in various medical clinical examination fields.

[0003] CT reconstruction technology is one of the core technologies of a CT imaging system. First, the calibration parameters of the CT system are obtained, then the projection image is re-projected onto an ideal detector plane using the calibration parameters for correction, and finally, the corrected projection image is used for reconstruction to obtain a three-dimensional CT image. Image correction is an important step to ensure the quality of the three-dimensional CT image.

[0004] The errors of the projection image mainly come from the offset and rotation between the actual detector and the ideal detector. Through the parameter calibration of the CT system, the offset (usually represented by the puncture point) and the rotation angle (usually represented by the rotation angles of the three coordinate axes relative to the coordinates of the ideal detector) of the actual detector relative to the ideal detector can be obtained. Before performing three-dimensional CT reconstruction, each error angle needs to be corrected separately to eliminate the image errors introduced by the deviation between the actual detector and the ideal detector. Usually, it can be corrected by the rotation transformation of the image. However, due to the tilt of the detector in the CT system, the "near-big and far-small" effect will be generated on the projection image, and the existing image rotation correction methods cannot eliminate this effect. Summary of the Invention

[0005] Object of the Invention: Aiming at the above deficiencies, the present invention proposes a method for tilting calibration of a projection image, which completes image calibration before reconstructing the projection image, ensures the imaging effect of the projection image, and effectively eliminates artifacts.

[0006] Technical Solution: A method for tilting calibration of a projection image, comprising the steps of:

[0007] (1) Obtain a projection image, and obtain the rotation angle of the actual plane where the current detector is located relative to the ideal plane where its ideal position is located through calibration;

[0008] (2) Calculate the rotation transformation matrix between the normal vector of the actual plane of the detector and the normal vector of the ideal plane according to the current rotation angle, and calculate the plane equations of the ideal plane and the actual plane;

[0009] (3) Obtain the coordinates of the feature points corresponding to the feature points selected on several actual planes on the ideal plane according to the perspective projection principle;

[0010] (4) According to the coordinates of the corresponding feature points on the actual plane and the ideal plane obtained in step (3), solve using a linear equation to obtain the transformation relationship between the actual plane and the ideal plane, and calibrate the projection image accordingly to obtain the calibrated projection image.

[0011] The specific content of step (2) is as follows:

[0012] (21) According to the normal vector n0 = (0, 0, 1) of the ideal plane when the detector is in the ideal position, and the ideal plane passes through the origin, the equation of the ideal plane is obtained as follows: ;

[0013] (22) Calculate the rotation transformation matrix of the normal vector of the actual plane and the normal vector of the ideal plane according to the current rotation angle , and obtain the normal vector of the actual plane as n a = ([[]] , , ), then , and calculate the equation of the actual plane accordingly: .

[0014] The rotation transformation matrix in step (22) is specifically:

[0015] The rotation angle of the current detector relative to its ideal position around the axis parallel to its vertical side and the rotation angle θ of the current detector relative to its ideal position around the axis parallel to its horizontal side obtained through calibration are used to calculate the first rotation transformation matrix between the actual plane and the ideal plane where the current detector is located:

[0016] .

[0017] The specific content of step (3) is: Select several feature points on the actual plane, calculate the straight-line equation between each feature point and the X-ray source, and calculate the coordinates of the feature points corresponding to each feature point on the ideal plane in combination with the plane equations of the ideal plane and the actual plane.

[0018] The specific content of step (4) is:

[0019] (41) Obtain the transformation relationship a between the points on the ideal plane p0 and the corresponding feature points on the actual plane p :

[0020]

[0021] Among them, ;

[0022] Based on this, it is obtained that:

[0023]

[0024] (42) Using the selected feature points, solve the transformation matrix a between the ideal plane p0 and the actual plane p ;

[0025] Among them,

[0026]

[0027]

[0028] where m is the number of selected feature points;

[0029] (43) Solve the corresponding points of the points on the actual plane on the ideal plane respectively to obtain the calibrated projection image.

[0030] The step (***43***) further includes: obtaining the pixel values of each pixel of the image after transformation on the actual plane by using the cubic interpolation algorithm.

[0031] The method further includes the following steps:

[0032] Obtain the rotation angle η of the current detector relative to its ideal position around the axis perpendicular to the image plane through calibration, and perform a rotation transformation on the calibrated projection image accordingly to obtain the final calibrated image.

[0033] During the rotation transformation of the calibrated projection image, the rotation transformation matrix .

[0034] Beneficial effects: The present invention solves the problem of deformation of the projection image caused by the rotation and tilt between the actual detector and the ideal detector, completes image calibration before the reconstruction of the projection image, ensures the imaging effect of the projection image, and effectively eliminates artifacts. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 is a flowchart of the present invention;

[0036] Figure 2 in (a) is a schematic diagram of the world coordinate system established based on the C-arm machine, Figure 2 in (b) is the axis of the detector relative to the ideal coordinate system Figure 2 Note: In the translation, the number in step (***43***) in the original text should be a specific number, but it is not clear in the provided text, so it is retained as it is in the translation. Also, the specific content of the tags a , , etc. remains unchanged as they are likely specific identifiers in the patent context.In (c), the detector has a rotation angle relative to axis, and the schematic diagram is as follows. Figure 2 In (d), the detector has a rotation angle θ relative to axis, and the schematic diagram is as follows;

[0037] Figure 3 In (a), the detector has a rotation angle relative to axis, and the projection schematic diagram is as follows when the rotation angle exists. Figure 3 In (b), the projection image collected by the detector is as follows, Figure 3 and in (c), the schematic diagram of the effect after calibrating the projection image error is as follows;

[0038] Figure 4 In (a), the projection schematic diagram is as follows when the detector has a rotation angle θ relative to axis, Figure 4 in (b), the projection image collected by the detector is as follows, Figure 4 and in (c), the schematic diagram of the effect after calibrating the projection image error is as follows;

[0039] Figure 5 is the relationship diagram between the ideal plane and the actual plane when there are rotation angles and θ at the same time;

[0040] Figure 6 is the characteristic dot matrix diagram selected in different regions of the image;

[0041] Figure 7 In (a), the projection schematic diagram is as follows when the detector has a rotation angle η relative to axis, Figure 7 in (b), the projection image collected by the detector is as follows, Figure 7 and in (c), the schematic diagram of the effect after calibrating the projection image error is as follows. Specific Embodiment

[0042] The present invention will be further illustrated below in conjunction with the accompanying drawings and specific embodiments.

[0043] Figure 1 is the flow chart of the present invention. As Figure 1 shown, the method for calibrating the inclination of the projection image of the present invention includes the following steps:

[0044] (1) Obtain the spatial coordinates of the geometric center O of the calibration phantom, and use the geometric center O of the calibration phantom as the origin, and use the rotation axis of the C-arm machine as the axis, and use the direction from the origin to the X-ray source when the C-arm of the C-arm machine is in the initial position as the axis, and determine the axis according to the right-hand rule to establish the world coordinate system , such as Figure 2 shown in (a) of

[0045] When the detector is in the ideal position, on the image projected on the detector, taking the projection of the X-ray source S on the detector as the origin, with the direction along the image plane downward as the axis, and the direction perpendicular to the image plane and outward as the axis, determine the axis according to the right-hand rule, and establish an ideal plane coordinate system , such as Figure 2 shown in (b) of Figure 2 shown in (c) of Figure 2 shown in (d) of

[0046] (2) The C-arm machine performs rotational scanning on the calibration phantom to obtain the projection image of the calibration phantom on the detector; wherein, the calibration phantom is cylindrical, and two layers of steel balls are arranged in parallel with a spacing of D inside the calibration phantom, and each layer of steel balls is placed on a standard circular orbit with a radius of R. The standard circular orbits are symmetrically arranged on the upper and lower sides of the center of the calibration phantom, so that the two layers of steel balls are symmetrically arranged on the upper and lower sides of the center of the calibration phantom;

[0047] (3) Through calibration, it is obtained that the detector has a rotation angle with respect to the axis and a rotation angle θ with respect to the axis;

[0048] Figure 2 (c) in is a schematic diagram of the rotation angle of the detector with respect to the ideal plane coordinate system Figure 2 axis, and (d) in Figure 5 is a schematic diagram of the rotation angle of the detector with respect to the axis. In the present invention, according to step (2), the elliptic equation formed by the projections of the two layers of steel balls on the calibration phantom on the detector is calculated, and the corresponding elliptic parameters are solved accordingly, and based on this, the rotation angles and θ of the current detector relative to its ideal position are calculated according to the distance from the ray source to the detector;

[0049] In the present invention, many calibration methods have been provided in the prior art to obtain the rotation angle i of the detector with respect to the y axis, the rotation angle θ with respect to the x i axis, and the rotation angle η with respect to the z i axis, which will not be elaborated here one by one;

[0050] (4) Judge the current rotation angle Whether and θ are both 0. If both are 0, jump to step (11); otherwise, go to step (5);

[0051] (5) Let the normal vector of the ideal plane p0 when the detector is at the ideal position be n0 = (0, 0, 1), and the ideal plane p0 passes through the origin. Therefore, the equation of the ideal plane p0 can be obtained as: ;

[0052] (6) From the current rotation angle and θ, the rotation transformation matrix between the normal vectors of the actual plane p a of the detector at the actual position and the ideal plane p0 can be obtained:

[0053]

[0054] The normal vector of the actual plane p a is n a = ( , , ). Then, can be obtained, and the equation of the actual plane p a can be obtained: ;

[0055] Among them, the actual plane p a after rotation corresponds to the obtained x a , y a , z a , that is, the actual coordinate system ;

[0056] (7) Given that the distance between the X-ray source S and the detector at the ideal position is s, then the coordinates of the X-ray source S in the ideal plane coordinate system are (0, 0, s). Let the coordinates of any point k on the actual plane p a be (x k , y k , z k ). Then the equation of the straight line Sa is: ;

[0057] (8) From the equation of the straight line Sk and the equation of the actual plane p a , the coordinates of the intersection point of the straight line Sk and the ideal plane p0 can be obtained as , where ;

[0058] (9) Repeat (7) and (8) to calculate the coordinates of the corresponding points on the ideal plane p0 of the 25 feature points selected on the actual plane p Figure 6 as shown; a ;

[0059] (10) Obtain the transformation relationship between the points on the ideal plane p0 and the corresponding feature points on the actual plane p a :

[0060]

[0061] Among them, ;

[0062] It can be obtained that:

[0063]

[0064] Therefore, according to the 25 pairs of feature points obtained in step (9), the linear equation At = 0 can be used to solve the transformation matrix between the ideal plane p0 and the actual plane p a ;

[0065] Among them,

[0066]

[0067] , m = 25;

[0068] Solve accordingly to obtain The values of the elements in, and then obtain the transformation matrix And calibrate the image accordingly, and use the cubic interpolation algorithm to obtain the pixel values of each pixel on the image obtained after the transformation of the actual plane p a The pixel values of each pixel on the image obtained after the transformation are obtained to obtain the calibrated projection image;

[0069] (11) Obtain the current value η of the rotation angle of the detector around the z i axis relative to its ideal position, and determine whether the rotation angle η of the actual detector relative to its ideal position is 0; if the value of η is 0, execute step (13); otherwise, execute step (12);

[0070] (12) Perform a rotation transformation with an angle of η on the calibrated projection image to correct the rotation angle η of the detector;

[0071] According to the calibrated image obtained in step (10), perform a rotation transformation on it, where the rotation transformation matrix ; And use the cubic interpolation algorithm to obtain the pixel values of each pixel for the image after this rotation transformation;

[0072] (13) Output the finally calibrated image.

[0073] Figure 3 (a) in is the angular error of the actual detector relative to its ideal position around the y i axis Schematic diagram of Figure 3 In (b), there is an angular error around the y i axis. When the calibration phantom's projected image on the actual detector is calibrated by the present invention, the projected image shown in (c) is obtained. When there is an angular error around the y Figure 3 axis, the projected image of the calibration phantom on the actual detector, after being calibrated by the present invention, obtains the projected image shown in (c).

[0074] Figure 4 In (a), it is a schematic diagram of the relative positions of the X-ray source, calibration phantom, and detector, where the detector has an angular error θ around the x i axis relative to its ideal position. Figure 4 In (b), when there is an angular error θ around the x i axis, the projected image of the calibration phantom on the actual detector, after being calibrated by the present invention, obtains the projected image shown in (c). Figure 5 When there is an angular error θ around the x

[0075] Figure 7 In (a), it is a schematic diagram of the relative positions of the X-ray source, calibration phantom, and detector, where the detector has an angular error η around the z i axis relative to its ideal position. Figure 7 In (b), when there is an angular error η around the z i axis, the projected image of the calibration phantom on the actual detector, after being calibrated by the present invention, obtains the projected image shown in (c). Figure 7 When there is an angular error η around the z

[0076] The present invention solves the problem of the projected image deformation caused by the rotation and tilt of the actual detector relative to the ideal detector. It completes image calibration before reconstructing the projected image, ensures the imaging effect of the projected image, and effectively eliminates artifacts.

[0077] The preferred embodiments of the present invention have been described in detail above. However, the present invention is not limited to the specific details in the above embodiments. Within the scope of the technical concept of the present invention, various equivalent transformations (such as quantity, shape, position, etc.) can be made to the technical solutions of the present invention, and these equivalent transformations all belong to the protection scope of the present invention.

Claims

1. A method for tilt calibration of a projected image, characterized in that: Including the steps: (1) Obtain the projection image, and obtain the rotation angle of the actual plane where the current detector is located relative to the ideal plane where its ideal position is located through calibration; (2) Calculate the rotation transformation matrix between the normal vector of the actual plane of the detector and the normal vector of the ideal plane according to the current rotation angle, and calculate and obtain the plane equations of the ideal plane and the actual plane. Specifically: (21) According to the normal vector \(n_0=(0, 0, 1)\) of the ideal plane when the detector is in the ideal position, and the ideal plane passes through the origin, the equation of the ideal plane is obtained as follows: ; (22) The rotation angle of the current detector relative to its ideal position about the axis parallel to its vertical side obtained through calibration and the rotation angle θ of the current detector relative to its ideal position about the axis parallel to its horizontal side are used to calculate the first rotation transformation matrix between the actual plane where the current detector is located and the ideal plane : ； The normal vector of the actual plane is obtained as n a = ([[]] , , ), then is obtained, and based on this, the equation of the actual plane is calculated as: ; (3) Obtain the coordinates of the corresponding feature points on the ideal plane for several feature points selected on the actual plane according to the perspective projection principle; (4) According to the coordinates of the corresponding feature points on the actual plane and the ideal plane obtained in step (3), use a linear equation to solve for the transformation relationship between the actual plane and the ideal plane, and calibrate the projection image accordingly to obtain the calibrated projection image. Specifically: (41) Obtain the transformation relationship between the points on the ideal plane p0 and the corresponding feature points on the actual plane p a :​ ； Among them, ; Accordingly, it is obtained that: ； (42) Solve the transformation matrix between the ideal plane p0 and the actual plane p by using the linear equation At = 0 according to the selected feature points a between ; Wherein, ； ； where m is the number of selected feature points; (43) Solve the corresponding points on the ideal plane for the points on the actual plane respectively, and obtain the calibrated projection image.

2. The tilt calibration method according to claim 1, wherein: The specific content of step (3) is: Select several feature points on the actual plane, calculate the straight-line equation between each feature point and the X-ray source, and calculate the coordinates of the corresponding feature points on the ideal plane in combination with the plane equations of the ideal plane and the actual plane.

3. The tilt calibration method according to claim 1, characterized in that: Step (43) further includes: Using a cubic interpolation algorithm to obtain the pixel values of each pixel of the image after transformation on the actual plane.

4. The tilt calibration method according to any one of claims 1 to 3, characterized in that: It further includes the following steps: Obtain the rotation angle η of the current detector relative to its ideal position around the axis perpendicular to the image plane through calibration, and perform a rotation transformation on the calibrated projection image accordingly to obtain the final calibrated image.

5. The tilt calibration method according to claim 4, characterized in that: In the rotation transformation of the calibrated projection image, its rotation transformation matrix .

Images