Method and device for surface extraction of CT volume in X-ray CT

Abstract

The present invention provides a surface extraction method and device for a CT volume in an X-ray CT. The surface extraction method and device for a CT volume in an X-ray CT extract voxels with fixed CT values ​​from volume data acquired by the X-ray CT to generate an isosurface mesh (M), calculate a gradient vector (g) of the CT value of the vertex (p) at each vertex (p) of the mesh (M), generate a plurality of sampling points (S) in the positive and negative directions of the calculated gradient vector (g), calculate the gradient norm (N) of the CT value at each generated sampling point (S), move the vertex to a sampling point (Sm) where the calculated gradient norm is a maximum value (Nm), and thereby correct the vertex (p) of the isosurface.

Description

Technical Field

[0001] The present invention relates to a surface extraction method and device for a CT volume in an X-ray CT, and in particular to a surface extraction method and device for a CT volume in an X-ray CT that can resist CT value fluctuations and achieve high-precision surface extraction. Background Art

[0002] In the 1970s, medical X-ray CT devices entered the practical stage, and based on this technology, X-ray CT devices for industrial products began to appear in the early 1980s. Since then, industrial X-ray CT devices have been used to observe and inspect pores in cast parts, poor welding of welded parts, and defects in circuit patterns of circuit parts that are difficult to confirm from the appearance. On the other hand, in recent years, with the popularization of 3D printers, there has been an increasing demand not only for observing and inspecting the inside of processed products obtained by 3D printers, but also for 3D dimensional measurement of internal structures and its high precision.

[0003] In response to the above technical trend, X-ray CT devices for measurement (see Japanese Patent Publication No. 2002-71345 and Japanese Patent Publication No. 2004-12407) have begun to spread mainly in Germany. In this X-ray CT device for measurement, the measurement object is placed at the center of a rotating table, and X-ray irradiation is performed while the measurement object is rotated.

[0004] exist Figure 1 1 shows the structure of a general X-ray CT apparatus 1 used in measurement. The cover 10 for shielding X-rays includes: an X-ray source 12 that radiates a cone-beam X-ray 13; an X-ray detector 14 that detects the X-ray 13; a rotating table 16 for placing a measurement object (e.g., a workpiece) W and rotating the workpiece W for CT imaging; and an XYZ moving mechanism 18 for adjusting the position and magnification of the workpiece W projected in the X-ray detector 14. The general X-ray CT apparatus 1 is also composed of a controller 20 that controls these devices, a control PC 22 that provides instructions to the controller 20 through user operations, and the like.

[0005] The control PC 22 has a function of displaying the projection image of the workpiece W projected on the X-ray detector 14 and a function of reconstructing a tomographic image from a plurality of projection images of the workpiece W in addition to controlling each device.

[0006] like Figure 2As shown, the X-rays 13 emitted from the X-ray source 12 pass through the workpiece W on the rotating table 16 and reach the X-ray detector 14. While the workpiece W is rotating, the X-ray detector 14 obtains perspective images (projection images) of the workpiece W in all directions, and reconstructs the workpiece W using a CT reconstruction algorithm such as a back-projection method or a successive approximation method, thereby generating a tomographic image of the workpiece W.

[0007] By controlling the XYZ axes of the XYZ moving mechanism 18 and the θ axis of the rotating table 16 , the position of the workpiece W can be moved, and the imaging range (position, magnification) and imaging angle of the workpiece W can be adjusted.

[0008] In order to obtain a tomographic image or volume data (a collection of stereoscopic images or tomographic images of the workpiece W in the Z-axis direction) of the workpiece W which is the final target of the X-ray CT apparatus 1 , a CT scan of the workpiece W is performed.

[0009] CT scanning consists of two processes: obtaining a fluoroscopic image (projection image) of the workpiece W and CT reconstruction. In the fluoroscopic image obtaining process, the rotating table 16 on which the workpiece W is mounted is rotated continuously at a fixed speed or intermittently at a fixed step length during X-ray irradiation to obtain a fluoroscopic image of the workpiece W in the entire circumference direction (fixed interval). The obtained fluoroscopic image in the entire circumference direction (fixed interval) is CT reconstructed using a CT reconstruction algorithm such as a back-projection method or a successive approximation method, thereby generating a CT image such as Figure 3 The workpiece shown in the example (in Figure 3 : The tomographic image or volume data of the master ball (B).

[0010] A mesh of a desired surface shape can be generated from the generated volume data for evaluation and analysis (see Japanese Patent Application Laid-Open No. 2018-40790). Summary of the invention

[0011] Problem that the invention aims to solve

[0012] Regarding the basic method of generating a mesh of a surface shape based on volume data, there is a method of generating a mesh by extracting voxels of a fixed CT value (isosurface mesh generation). However, in general, the surface shape of volume data has fluctuations in the CT value caused by artifacts (noise). In this method, the surface shape is directly affected by the fluctuations in the CT value caused by artifacts, etc., and there is a problem that the surface shape cannot be extracted with high precision by only determining the isosurface of the surface based on a fixed CT value.

[0013] Solutions for solving problems

[0014] The present invention has been made to solve the above-mentioned conventional problems, and an object of the present invention is to resist CT value fluctuation and realize high-precision surface extraction.

[0015] In the present invention, when performing surface extraction of a CT volume in X-ray CT, voxels with fixed CT values ​​are extracted from volume data acquired by X-ray CT to generate an isosurface mesh, and at each vertex of the mesh, the gradient information of the CT value is used to correct the vertices of the isosurface, thereby solving the above-mentioned problem.

[0016] Here, the vertex of the isosurface can be corrected in the following manner: calculate the gradient vector of the CT value of the vertex, generate multiple sampling points in the positive and negative directions of the calculated gradient vector, calculate the gradient norm of the CT value at each generated sampling point, and move the vertex to the sampling point with the largest calculated gradient norm.

[0017] In addition, the CT value of the apex can be calculated by the following formula using the Feldkamp method.

[0018]

[0019] Here, the coefficient α(θ, p)=d sod / (d sod +d z )

[0020] θ: rotation angle

[0021] d sod : Distance between line source and rotation center

[0022] S filtered (θ, p); filter corrected projection value (using Shepp-Logan filter)

[0023] In addition, the present invention provides a device for performing surface extraction of a CT volume in X-ray CT, the device being characterized in that it comprises: a unit for generating an isosurface mesh by extracting voxels with fixed CT values ​​from volume data acquired by X-ray CT; and a unit for correcting the vertices of the isosurface using gradient information of the CT value at each vertex of the mesh.

[0024] Here, the unit for modifying the vertices of the isosurface may include:

[0025] A unit for calculating the gradient vector of the CT value of a vertex;

[0026] A unit for generating multiple sampling points in the positive and negative directions of the calculated gradient vector;

[0027] A unit for calculating the gradient norm of the CT value at each generated sampling point; and

[0028] The cell that moves the vertex to the sampling point where the calculated gradient norm is the largest.

[0029] Effects of the Invention

[0030] According to the present invention, by using the gradient information of the CT value to correct the isosurface, high-precision surface extraction can be achieved.

[0031] These features, advantages and other features and advantages of the present invention will become apparent from the following detailed description of the preferred embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The preferred embodiments are described with reference to the accompanying drawings, wherein like reference numerals are used to identify like elements throughout the drawings, and wherein:

[0033] Figure 1 This is a cross-sectional view showing the overall structure of a conventional X-ray CT apparatus used for measurement.

[0034] Figure 2 It is a perspective view showing the arrangement of main parts of the X-ray CT apparatus of the prior art.

[0035] Figure 3 FIG. 1 is a diagram showing an overview of CT reconstruction by the X-ray CT apparatus of the related art.

[0036] Figure 4 It is a flowchart showing the processing procedure of the embodiment of the present invention.

[0037] Figure 5 This is a perspective view for explaining an example of a method for analyzing and calculating a CT value gradient according to an embodiment of the present invention.

[0038] Figure 6 FIG. 1 is a diagram showing an example of extracting a plurality of sampling points in the gradient ±g directions according to the embodiment of the present invention.

[0039] Figure 7 This is a diagram showing a state where vertices are moved to sampling points according to an embodiment of the present invention.

[0040] Figure 8 It is a perspective view showing a stepping cylinder as an example of a measurement object in the related art. DETAILED DESCRIPTION

[0041] The embodiments of the present invention are described in detail below with reference to the accompanying drawings. In addition, the present invention is not limited to the contents described in the following embodiments and examples. In addition, the constituent elements in the embodiments and examples described below include constituent elements that can be easily thought of by a person skilled in the art, substantially the same constituent elements, and constituent elements of the so-called equivalent scope. Furthermore, the constituent elements disclosed in the embodiments and examples described below can be appropriately combined for use, or can be appropriately selected for use.

[0042] In the algorithm of the present invention, volume data generated by X-ray CT is used as input, and mesh data obtained by surface extraction is used as output.

[0043] exist Figure 4 An overview of the processing procedure in the embodiment of the present invention is shown in FIG.

[0044] First, in step 100, an isosurface mesh M is generated according to volume data.

[0045] Each voxel of the volume data has a CT value, and voxels with a fixed CT value can be extracted to generate an isosurface mesh. As a method for selecting a CT value, there is a method of selecting by confirming a histogram of the CT value of the volume data. The histogram usually has peaks for each workpiece material and air. For example, when it is desired to extract the workpiece outer surface (the boundary surface between the air and the workpiece), the middle value (CT value) of the peak of the air and the workpiece outer surface material is selected on the histogram. In addition, an appropriate CT value corresponding to the desired surface shape in the selected volume data is used to generate an isosurface mesh.

[0046] Next, the process proceeds to step 110 to extract a vertex p constituting the mesh M.

[0047] Next, the process proceeds to step 120 to calculate the gradient vector g of the extracted mesh vertex p. The gradient vector g can be calculated as follows.

[0048] First, for example, the CT value f(p) at the vertex p(x, y, z) is calculated by using the following formula using the Feldkamp method (FDK method) which is a commonly used back-projection method (see Figure 5 ).exist Figure 5 In the figure, 30 is a computer for performing calculations.

[0049]

Number 1

[0050]

[0051] Here, the coefficient α(θ, p)=d sod / (d sod +d z )

[0052] θ: rotation angle

[0053] d sod : Distance between line source and rotation center

[0054] S filteered (θ, p): filter-corrected projection value (using Shepp-Logan filter)

[0055] Furthermore, when obtaining the filter-corrected projection value, a Ram-Lak filter or a Kak-Slaney filter can be used in addition to the Shepp-Logan filter.

[0056] Next, the gradient vector g(p) of the CT value at the point p(x, y, z) is calculated using the following equation: Here, in order to simplify the calculation, it can be assumed that α(θ, p)=α(θ) (α does not depend on p).

[0057]

Number 2

[0058]

[0059] Here, R -θ : The three-dimensional rotation matrix of angle -θ

[0060] Differentiation in the detector coordinate system (calculation of the center difference of the filter-corrected projection values)

[0061] Furthermore, the point p can be regarded as a continuous value, and interpolation processing can be appropriately performed in calculations such as calculation of projection values.

[0062] Then, enter step 130, such as Figure 6 As illustrated, multiple sampling points S are extracted on the gradient vector ±g passing through the mesh vertex p. For example, the extraction range and extraction interval can be arbitrarily set based on the voxel size. For example, the sampling range can be set to about ±4 voxels, and the sampling interval can be set to about 0.1 to 0.2 times the voxel size.

[0063] Next, enter step 140, such as Figure 7 As illustrated, the gradient norm N of all the extracted sampling points S is calculated from the absolute value of the gradient vector g by the following formula.

[0064]

Number 3

[0065] N=|g(S)|···(3)

[0066] Next, the process proceeds to step 150 to derive a sampling point Sm that has a maximum value Nm among the gradient norm N of the sampling point S.

[0067] Next, the process proceeds to step 160 to move the vertex p to the sampling point Sm. This is because the deviation of the sampling point S is large at the boundary between the workpiece and the air, or at the boundary between different workpiece materials.

[0068] After step 160 is completed, the process proceeds to step 170 to check whether the correction of all vertices p has been completed.

[0069] If the correction is not completed, the process returns to step 110 to correct the next vertex p.

[0070] On the other hand, when it is determined in step 170 that corrections have been completed for all vertices p, the process ends.

[0071] According to the simulations conducted by the inventors, Figure 8 The measurement values ​​of the stepping cylinders shown in the example have an outer diameter of 20 mm for the first layer and 60 mm for the fifth layer. The results show that in the previous isosurface mesh method, the errors in the first and second layers were particularly large, but according to the present invention, the errors in the first and second layers were halved.

[0072] In addition, Figure 3 The ruby ​​sphere of the standard sphere B shown as an example was evaluated. In the conventional method, the difference between the actual measured diameter of the sphere and the diameter during fitting was 8 to 9 μm, but this difference was reduced to 0 to 2 μm according to the method of the present invention. The same tendency was observed for all the spheres.

[0073] In this embodiment, the FDK method is used, so the result of differentiating the formula for obtaining the CT value can be easily obtained, and the gradient information of the CT value can be easily obtained. In addition, the method of obtaining the gradient information of the CT value is not limited to the FDK method.

[0074] In the above-described embodiment, the gradient information is set to the gradient norm, but the gradient information is not limited thereto. The measurement object is also not limited to the workpiece.

[0075] It is obvious to those skilled in the art that the above described embodiments are only illustrative and represent the application of the principles of the present invention. Those skilled in the art can easily design various other embodiments without departing from the spirit and scope of the present invention.

[0076] The disclosure of Japanese Patent Application No. 2018-226040 filed on November 30, 2018 including the specification, drawings and claims is incorporated herein by reference in its entirety.

Claims

1. A method for surface extraction of a CT volume in X-ray CT, characterized in that: When performing surface extraction of CT volumes in X-ray CT, Extract voxels with fixed CT values ​​from the volume data acquired by X-ray CT to generate an isosurface mesh. At each vertex of the isosurface mesh, gradient information of the CT value is used to correct the vertex of the isosurface mesh to remove artifacts from the obtained CT image, The vertices of the isosurface mesh are corrected in the following manner: Calculate the gradient vector of the CT value of the vertex, Generate multiple sampling points in the positive direction and multiple sampling points in the negative direction of the calculated gradient vector, Calculate the gradient norm of the CT value at each generated sampling point, and Move the vertex to the sampling point where the calculated gradient norm is the largest, The plurality of sampling points in the positive direction and the plurality of sampling points in the negative direction are generated at predetermined intervals.

2. The surface extraction method of CT volume in X-ray CT according to claim 1, characterized in that: Using the Feldkamp method, the CT value of the vertex is calculated by the following formula, Here, the coefficient α(θ,p)=d sod / (d sod +dz) θ: rotation angle d sod : Distance between line source and rotation center S filtered (θ, p) : filter-corrected projection value obtained using the Shepp-Logan filter.

3. A surface extraction device for a CT volume in an X-ray CT, for performing surface extraction of a CT volume in an X-ray CT, wherein the surface extraction device for a CT volume in an X-ray CT is characterized by comprising: Extracting voxels with fixed CT values ​​from volume data acquired by X-ray CT to generate cells of an isosurface mesh; and A unit for modifying the vertices of the isosurface mesh using the gradient information of the CT value at each vertex of the isosurface mesh to remove artifacts from the obtained CT image, in, The vertices of the isosurface mesh are corrected in the following way: Calculate the gradient vector of the CT value of the vertex, Generate multiple sampling points in the positive direction and multiple sampling points in the negative direction of the calculated gradient vector, Calculate the gradient norm of the CT value at each generated sampling point, and Move the vertex to the sampling point where the calculated gradient norm is the largest, The plurality of sampling points in the positive direction and the plurality of sampling points in the negative direction are generated at predetermined intervals.

Images