X-ray CT apparatus, method, and program

Abstract

To highly accurately analyze an internal structure of an object even when the object is curved in its longitudinal direction.SOLUTION: An X-ray CT apparatus according to the present disclosure includes a first generation unit 110, a determination unit 115, and a second generation unit 120. The first generation unit 110 generates first to N-th images (N is a natural number equal to or greater than 2) according to the irradiation result of X-rays to an object. The determination unit 115 determines first to N-th center points respectively indicating the centers of the cross-sectional regions of the first to N-th images. The second generation unit 120 generates first to M-th (M is a natural number equal to or greater than 2) cross-sectional images according to the first to N-th images. The first to M-th cross-sectional images respectively include first to M-th points. The k-th cross-sectional image (1≤k≤M) is the image that represents the cross-sectional region of the object on a k-th plane which is orthogonal to a k-th vector heading from the i-th center point to the j-th center point (1≤i<j≤N) within the near-field region of the k-th point.SELECTED DRAWING: Figure 12

Description

Technical Field

[0001] The present disclosure relates to an X-ray CT apparatus, method, and program.

Background Art

[0002] Japanese Patent Application Laid-Open No. 2023-070250 discloses an X-ray CT (Computed Tomography) apparatus. This apparatus includes an X-ray light source and a stage. The X-ray light source irradiates X-rays toward a sample contained in a cooling container. The stage has a sample placement surface (support surface) and supports the cooling container on that surface.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] An X-ray CT apparatus is used to analyze the internal structure of an object by generating a plurality of cross-sectional images of the object. Each of the plurality of cross-sectional images is basically generated so as to represent a cross-sectional area of the object in a plane orthogonal to the normal direction of the support surface of a support member such as a stage.

[0005] An object such as an electric cable can frequently bend in its longitudinal direction depending on the actual use environment. Therefore, observing the object in a bent state by CT scanning is extremely important from the viewpoint of quality evaluation. However, when the object is bent in its longitudinal direction, the way (shape) in which the cross-sectional area of the object appears in a plurality of cross-sectional images may change significantly. As a result, it may not be possible to accurately analyze the internal structure of the object.

[0006] The present disclosure has been made to solve the above problems, and an object thereof is to provide an X-ray CT apparatus, a method, and a program that enable high-precision analysis of the internal structure of an object even when the object is curved in its longitudinal direction.

Means for Solving the Problems

[0007] The X-ray CT apparatus according to one aspect of the present disclosure includes a light source, a support member, a first generation unit, a determination unit, and a second generation unit. The light source irradiates an object with X-rays. The support member has a support surface for supporting the object. The first generation unit generates first to Nth images (N is a natural number of 2 or more), each of which represents a cross-sectional area of the object in a plane orthogonal to the normal vector of the support surface, according to the irradiation result of X-rays on the object. The determination unit determines first to Nth center points each indicating the center of the cross-sectional area of the first to Nth images. The second generation unit generates first to Mth cross-sectional images (M is a natural number of 2 or more) according to the first to Nth images. The first to Mth cross-sectional images each include first to Mth points. The kth cross-sectional image (1 ≤ k ≤ M) is an image representing the cross-sectional area of the object in the kth plane orthogonal to the kth vector directed from the ith center point to the jth center point (1 ≤ i < j ≤ N) within the vicinity of the kth point. The first point is included in the first image, and the hth point (2 ≤ h ≤ M) is separated from the (h - 1)th point by a predetermined distance in the direction of the (h - 1)th vector.

[0008] A method according to an aspect of the present disclosure is a method for generating an image of an object using an X-ray CT apparatus. The X-ray CT apparatus includes a light source and a support member. The light source irradiates the object with X-rays. The support member has a support surface for supporting the object. The method includes generating first to Nth images each representing a cross-sectional area of the object in a plane orthogonal to the normal vector of the support surface according to the irradiation result of the X-rays on the object, determining first to Nth center points each indicating the center of the cross-sectional area of the first to Nth images, and generating first to Mth cross-sectional images according to the first to Nth images. The first to Mth cross-sectional images each include first to Mth points. The kth cross-sectional image is an image representing a cross-sectional area of the object in a kth plane orthogonal to a kth vector directed from an ith center point to a jth center point within a neighborhood area of the kth point. The first point is included in the first image. The hth point is separated from the (h - 1)th point by a predetermined distance in the direction of the (h - 1)th vector.

[0009] A program according to an aspect of the present disclosure is a program for causing a computer to execute the above method.

Advantages of the Invention

[0010] According to the present disclosure, even when the object is curved in its longitudinal direction, the internal structure of the object can be analyzed with high accuracy.

Brief Description of the Drawings

[0011]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Figure 17

Figure 18

Figure 19

Figure 20

Figure 21

Figure 22

Figure 23

Figure 24

Figure 25

Figure 26

Figure 27

Figure 28

Figure 29

Figure 30

Figure 31

Figure 32

Figure 33

Figure 34

Figure 35

Figure 36

Figure 37

Figure 38

Figure 39

Figure 40

[0012] Description of Embodiments of the Present Disclosure First, embodiments of the present invention will be listed and described.

[0013] (1) The X-ray apparatus 1 according to one aspect of the present disclosure includes a light source 3, a stage 4, a first generation unit 110, a determination unit 115, and a second generation unit 120. The light source 3 irradiates the cable 9 with X-rays 3a. The stage 4 has a support surface 4a for supporting the cable 9. The first generation unit 110 generates normal slice images 200_1 to 200_N (N is a natural number of 2 or more), each representing a cross-sectional area of the cable 9 in a plane orthogonal to the normal vector of the support surface 4a, according to the irradiation result of the X-rays on the cable 9. The determination unit 115 determines center points CP1 to CPN respectively indicating the centers of the cross-sectional areas of the cable 9 in the normal slice images 200_1 to 200_N. The second generation unit 120 generates oblique slice images 350_1 to 350_M (M is a natural number of 2 or more) according to the normal slice images 200_1 to 200_N. The oblique slice images 350_1 to 350_M each include points P1 to PM. The oblique slice image 350_k (1 ≦ k ≦ M) is an image representing the cross-sectional area of the cable 9 in a plane 355(k) orthogonal to a vector Vij(k) from the center point CPi to the center point CPj (1 ≦ i < j ≦ N) of the cable 9 within the vicinity area of the point Pk. The point P1 is included in the normal slice image 200_1. The point Ph (2 ≦ h ≦ M) is separated from the point P(h - 1) by a predetermined distance in the direction of the vector Vij(h - 1).

[0014] Even when the cable 9 is curved in its longitudinal direction, the direction of the vector Vij(k) is approximately equal to the local longitudinal direction of the cable 9 as seen from the point Pk. Therefore, with the configuration of (1) above, the plane 355(k) is approximately equal to a plane orthogonal to the local longitudinal direction at the point Pk. Accordingly, any of the first to M-th cross-sectional images accurately represents a plane (cross-section) orthogonal to the longitudinal direction of the object. As a result, even when the object is curved in its longitudinal direction, the internal structure of the object can be analyzed with high accuracy based on these cross-sectional images.

[0015] (2) In the above (1), each of the normal slice images 200_1 to 200_N includes n×m pixels (n and m are natural numbers) to which feature amounts are respectively assigned. Each of the diagonal slice images 350_k includes n×m pixels to which feature amounts are respectively assigned. The number of pixels of the diagonal slice image 350_k may be different from n×m, but since it is often n×m unless there is a special purpose, the following description will proceed on the assumption that the number of pixels is n×m. The first generation unit 110 generates each of the normal slice images 200_1 to 200_N such that the coordinates of each of the n×m pixels of the image are determined in a reference coordinate system in which the direction of the normal vector is the Z-axis direction. The second generation unit 120 includes a coordinate system setting unit 125, a coordinate conversion unit 130, and an assignment unit 135. The coordinate system setting unit 125 has the direction of the vector Vij(k) as the z(k)-axis direction and sets a k-th local coordinate system for determining the coordinates of each of the n×m pixels of the diagonal slice image 350 according to the center points CP1 to CPN of the cable 9 in the normal slice image group. The coordinate conversion unit 130 converts the coordinates of each of the n×m pixels of the diagonal slice image 350 in the k-th local coordinate system into the k-th corresponding coordinates, which are the coordinates of the pixel in the reference coordinate system according to the angle θ(k), which is the angle formed by the normal vector and the vector Vij(k). The assignment unit 135 assigns the feature amount of each of the n×m pixels of the diagonal slice image 350 according to the feature amount of at least one pixel included in the n×m pixels of the normal slice images 200_1 to 200_N, which has coordinates within the vicinity region R(X, Y, Z) of the k-th corresponding coordinates of the pixel.

[0016] If the k-th corresponding coordinates of the pixels of the diagonal slice image 350_k coincide with any of the coordinates of the n×m pixels of the normal slice images 200_1 to 200_N, it is preferable that the feature amount of the pixel of the diagonal slice image 350_k is the feature amount of the pixel having the coordinates that coincide with the k-th corresponding coordinates. However, in most cases, the k-th corresponding coordinates do not coincide with any of the coordinates of the n×m pixels of the normal slice images 200_1 to 200_N. Therefore, it is important to appropriately determine the feature amount of each pixel of the diagonal slice image 350_k. By adopting the configuration of (2) above, the feature amount of each pixel of the diagonal slice image 350_k is generated according to the feature amount of at least one pixel included in the n×m pixels of the normal slice images 200_1 to 200_N in the vicinity of the k-th corresponding coordinates of the pixel. Therefore, even when the k-th corresponding coordinates do not coincide with any of the coordinates of the n×m pixels of the normal slice images 200_1 to 200_N, the feature amount of each pixel of the diagonal slice image 350_k can be appropriately determined.

[0017] (3) In (2) above, the first generation unit 110 generates the normal slice images 200_1 to 200_N such that the Z coordinate of each of the n×m pixels of the normal slice image 200_f (2 ≤ f ≤ N) is separated by a length L in the Z-axis direction from the Z coordinate of each of the n×m pixels of the image of the normal slice image 200_(f - 1). The second generation unit 120 generates the diagonal slice images 350_1 to 350_M such that a predetermined distance in the k-th local coordinate system is equal to the length L. It is also possible to generate these diagonal slice images such that the predetermined distance is different from the length L.

[0018] If the cable 9 is not curved in its longitudinal direction, the normal slice images 200_1 to 200_N can accurately analyze the cross-section of the cable 9 when generated at intervals of length L. Therefore, it is preferable that the cross-sectional images of the oblique slice images 350_1 to 350_M are also generated at intervals of length L along the longitudinal direction of the cable 9, similar to the normal slice images 200_1 to 200_N. By adopting the configuration of (3) above, since the distance between the point P(k - 1) and the point Pk is the length L, the oblique slice images 350_1 to 350_M are also generated at intervals of length L along the longitudinal direction of the cable 9. Therefore, the internal structure of the cable 9 can be analyzed with higher accuracy.

[0019] (4) The method according to one aspect of the present disclosure is a method for generating an image of an object using the X-ray apparatus 1. The X-ray apparatus 1 includes a light source 3 and a stage 4. The light source 3 irradiates the cable 9 with X-rays 3a. The stage 4 has a support surface 4a for supporting the cable 9. The method includes a step (S105) of generating normal slice images 200_1 to 200_N (N is a natural number of 2 or more), each representing a cross-sectional area of the cable 9 in a plane orthogonal to the normal vector of the support surface 4a, according to the irradiation result of the X-rays on the cable 9; a step (S110) of determining center points CP1 to CPN respectively indicating the centers of the cross-sectional areas of the cable 9 in the normal slice images 200_1 to 200_N; and a step (S115) of generating oblique slice images 350_1 to 350_M (M is a natural number of 2 or more) according to the normal slice images 200_1 to 200_N. The oblique slice images 350_1 to 350_M each include points P1 to PM. The oblique slice image 350_k (1 ≤ k ≤ M) is an image representing the cross-sectional area of the cable 9 in a plane 355(k) orthogonal to the vector Vij(k) from the center point CPi of the cable 9 to the center point CPj (1 ≤ i < j ≤ N) of the cable 9 within the vicinity region RG(k) of the point Pk. The point P1 is included in the normal slice image 200_1. The point Ph (2 ≤ h ≤ M) is separated from the point P(h - 1) by a predetermined distance in the direction of the vector Vij(h - 1).

[0020] By adopting the configuration of (4) above, similar to the case of (1) above, the internal structure of the cable 9 can be analyzed with high precision.

[0021] (5) In the above (4), each of the normal slice images 200_1 to 200_N includes n×m pixels (n and m are natural numbers) to which feature amounts are respectively assigned. Each of the diagonal slice images 350_k includes n×m pixels to which feature amounts are respectively assigned. The number of pixels of the diagonal slice image 350_k may be different from n×m, but hereinafter, the description will proceed assuming n×m. Each of the normal slice images 200_1 to 200_N is generated such that the coordinates of each of the n×m pixels of the image are determined in a reference coordinate system with the direction of the normal vector as the Z-axis direction. The step (S115) of generating the diagonal slice images 350_1 to 350_M includes a step (S155) of setting a k-th local coordinate system for determining the coordinates of each of the n×m pixels of the diagonal slice image 350, having the direction of the vector Vij(k) as the z(k)-axis direction, according to the center points CP1 to CPN of the cable 9 in the normal slice image group; a step (S160) of converting the coordinates of each of the n×m pixels of the diagonal slice image 350 in the k-th local coordinate system into the k-th corresponding coordinates, which are the coordinates of the pixel in the reference coordinate system according to the angle θ(k), which is the angle formed by the normal vector and the vector Vij(k); and steps (S165, S170) of assigning the feature amount of each of the n×m pixels of the diagonal slice image 350 according to the feature amount of at least one pixel included in the n×m pixels of the normal slice images 200_1 to 200_N, which have coordinates within the vicinity region R(X, Y, Z) of the k-th corresponding coordinates of the pixel.

[0022] By adopting the configuration of (5) above, similar to the case of (2) above, the internal structure of the cable 9 can be analyzed with high precision.

[0023] (6) In the above (5), for each of the n×m pixels of the normal slice images 200_1 to 200_N, the Z coordinate of each pixel of the normal slice image 200_f (2 ≤ f ≤ N) is generated to be separated by a length L in the Z-axis direction from the Z coordinate of each of the n×m pixels of the image of the normal slice image 200_(f - 1). The oblique slice images 350_1 to 350_M are generated such that a predetermined distance in the k-th local coordinate system is equal to the length L. The predetermined distance may be different from the length L, but in many cases, it is easier for subsequent analysis if it is equal to the length L.

[0024] By adopting the configuration of the above (6), similar to the case of the above (3), the internal structure of the cable 9 can be analyzed with high precision.

[0025] (7) In the above (5), the cable 9 includes a plurality of strands 25. The method further includes, for each of the plurality of strands 25, a step (S305) of determining a k-th center coordinate, which is the center coordinate of the strand in the oblique slice image 350_k, and, for each of the plurality of strands 25, a step (S310) of converting the k-th center coordinate of the strand into a k-th corresponding center coordinate, which is the center coordinate of the strand in the reference coordinate system according to the angle θ(k).

[0026] For each element wire 25, when deriving the locus of the center coordinates (XY coordinates) of the element wire, it is assumed that the way the element wire is imaged (shape) does not vary significantly from image to image. However, in the normal slice images 200_1 to 200_N, the way each element wire 25 is imaged may change significantly depending on its Z coordinate. Therefore, it is difficult to accurately derive the locus of the center coordinates of each element wire 25 in the reference coordinate system using the normal slice images 200_1 to 200_N. By adopting the configuration of (7) above, the locus of the center coordinates of each element wire 25 is determined using the oblique slice images 350_1 to 250_M. In the oblique slice images 350_1 to 250_M, for each element wire 25, only its position changes depending on its z coordinate, and its shape hardly changes. Therefore, the locus of the k-th center coordinates is accurately derived. Thus, according to S305 and S310, the locus of the center coordinates of each element wire in the reference coordinate system can be accurately derived.

[0027] (8) In (5) above, the cable 9 includes a plurality of element wires 25. The method further includes a step of generating a high-resolution image 380_k by performing image processing for increasing the resolution of the oblique slice image 350_k (S302), a step of determining, for each of the plurality of element wires 25, the k-th center coordinates which are the center coordinates of the element wire in the high-resolution image 380_k (S305A), and a step of converting, for each of the plurality of element wires 25, the k-th center coordinates of the element wire into the k-th corresponding center coordinates which are the center coordinates of the element wire in the reference coordinate system according to the angle θ(k) (S310A).

[0028] By adopting the configuration of (8) above, the resolution of the oblique slice image 350_k can be increased by image processing such as super-resolution processing. As a result, the portion representing the cross-section of each element wire 25 in the high-resolution image 380_k becomes clearer. Consequently, the center (k-th center coordinates) of each element wire is determined with higher accuracy. Therefore, the locus of the center coordinates of each element wire in the reference coordinate system can be accurately derived. Note that the processing for increasing the resolution of the image is not limited to super-resolution processing, and various methods such as classical filter processing and segmentation using deep learning are applicable.

[0029] (9) In the above (7) or (8), the method further includes a classification process step (S315) of classifying each of the plurality of strands 25 into one of a plurality of groups according to the locus of the k-th center coordinates of the strand.

[0030] When the cable 9 is not curved (straight along the Z direction), the cross-sectional areas (cross-sections) of the cable 9 are parallel to each other. In this case, for example, the locus of the center coordinates of each strand 25 in the cable 9 changes in a predetermined periodic pattern in the normal slice images 200_1 to 200_N (reference coordinate system). Therefore, each strand 25 can be appropriately classified by a predetermined clustering algorithm such as the k-means method. On the other hand, when the cable 9 is curved, the locus of the center coordinates of each strand 25 does not change in such a pattern in the normal slice images 200_1 to 200_N. As a result, there is a possibility that each strand 25 may not be appropriately classified by the algorithm. The inventor noticed that even when the cable 9 is curved, the locus of the center coordinates (k-th center coordinates) of each strand 25 changes in a periodic pattern in the oblique slice images 350_1 to 350_M (high-resolution images 380_1 to 380_M). By adopting the configuration of the above (9), since each strand 25 is classified according to the locus of the k-th center coordinates of the strand, the above situation is avoided. As a result, each strand 25 can be appropriately classified using a predetermined algorithm such as the k-means method.

[0031] (10) In the above (7) or (8), the method further includes a step (S320) of generating a locus image 820 representing the locus of the k-th corresponding center coordinates of each of the plurality of strands 25 in the reference coordinate system.

[0032] By adopting the configuration of (10) above, for each elementary wire 25, the locus of the k-th corresponding center coordinates of the elementary wire in the reference coordinate system can be easily recognized by the user. As a result, even when the cable 9 is curved, the state of each elementary wire inside the cable 9 in the reference coordinate system (real space) can be appropriately and easily recognized by the user.

[0033] (11) A program according to an aspect of the present disclosure is a program for causing a computer to execute the method described in (5) above.

[0034] By adopting the configuration of (11) above, similar to the case of (1), the internal structure of the cable 9 can be analyzed with high accuracy. [Details of Embodiments of the Present Disclosure] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings. The same or corresponding parts in the drawings are denoted by the same reference numerals and their description will not be repeated. Each of the embodiments and their modifications may be combined with each other as appropriate.

[0035] <Embodiment 1> FIG. 1 is a diagram showing the hardware configuration of an X-ray CT apparatus according to Embodiment 1. Referring to FIG. 1, an X-ray CT apparatus (hereinafter also simply referred to as an “X-ray apparatus”) 1 includes a light source 3, a stage 4, a driving device 6, an X-ray detector 7, and a computer 10.

[0036] The light source 3 irradiates an object with X-rays 3a. In this example, the object is a cable 9. The cable 9 is an electric wire cable and is fixed by a jig 27. The shape of the cable 9 will be described in detail later. As will be shown later, the target of the method of the present invention is not limited to electric wire cables and can be widely applied to samples that can be curved in the longitudinal direction.

[0037] Stage 4 is a support member having a support surface 4a for supporting cable 9 (fixture 27). A plane parallel to the support surface 4a is defined as the XY plane. Stage 4 is rotatable about its rotation axis 4r. The rotation axis 4r extends along the direction of the normal vector of the support surface 4a (Z direction). The Z direction is orthogonal to the XY plane. The coordinate system defined by the X-axis, Y-axis, and Z-axis is also referred to as the "reference coordinate system".

[0038] The drive device 6 is provided below the cable 9 and includes a motor (not shown). The drive device 6 rotates the stage 4 about the rotation axis 4r or moves the stage 4 along the Z direction.

[0039] The X-ray detector 7 includes a plurality of detection elements 8. Each detection element 8 includes a scintillator and a photodiode (both not shown). The scintillator converts the X-ray 3a that has passed through the cable 9 into light. The photodiode detects the light converted by the scintillator.

[0040] While the light source 3 irradiates the cable 9 with the X-ray 3a, the drive device 6 rotates the stage 4. As a result, the X-ray 3a is irradiated from all circumferential directions of the cable 9. Consequently, an X-ray fluoroscopic image of the cable 9 is captured by the X-ray detector 7.

[0041] The computer 10 includes a control device 11, an input interface 12, a storage device 13, a display device 14, and a communication device 15.

[0042] The control device 11 controls the light source 3, the drive device 6, the display device 14, and the communication device 15. The control device 11 is also configured to function as a processing device that executes various processes, and includes a CPU (Central Processing Unit) 11a and a memory 11b. The CPU 11a executes various arithmetic processes by executing the programs stored in the memory 11b. The control device 11 generates volume data of the cable 9 according to the irradiation result of the X-ray 3a from the light source 3 to the cable 9 (specifically, the detection result of the X-ray detector 7). Specifically, the control device 11 reconstructs an X-ray fluoroscopic image of the cable 9 using a CT reconstruction algorithm such as the back-projection method or the successive approximation method, thereby generating volume data. This data will be described in detail later.

[0043] The input interface 12 receives user operations by the user of the X-ray device 1. The storage device 13 is a non-volatile memory device, for example, a hard disk or an SSD (Solid State Drive). The storage device 13 stores the volume data of the cable 9. The display device 14 is configured to display various images in response to user operations. The communication device 15 is configured to communicate with the information processing device 16.

[0044] The information processing device 16 is, for example, a PC (Personal Computer), and is provided outside the X-ray device 1. The information processing device 16 includes a control device 17, a communication device 18, a storage device 19A, and a display device 19B. The control device 17 controls the communication device 18 and the display device 19B. The control device 17 also functions as a processing device that executes various processes, and includes a CPU 17a and a memory 17b. The CPU 17a executes various arithmetic processes. The memory 17b stores the programs executed by the CPU 17a. The communication device 18 communicates with the X-ray device 1 through a wired or wireless connection. The storage device 19A stores various data used by the control device 17. The display device 19B displays various screens.

[0045] FIG. 2 is a view showing the appearance of the cable 9 and the jig 27. Referring to FIG. 2, in this example, the cable 9 is fixed by the jig 27 in a curved state in its longitudinal direction. A tape (not shown) is wound around the cable 9, thereby maintaining the posture of the cable 9. The X-axis, Y-axis, and Z-axis are the same as those shown in FIG. 1. Each of Za to Zc is an example of a predetermined Z coordinate. Note that the method for maintaining the posture of the cable 9 is not limited to the method using a tape, and for example, a method using some dedicated jig may also be used.

[0046] FIG. 3 is a view showing the internal structure of the cable 9. Referring to FIG. 3, this example shows a plane (cross-section) perpendicular to the longitudinal direction of the cable 9. The cable 9 includes a covering portion 21 and a plurality of strand bundles 22. The plurality of strand bundles 22 includes strand bundles 22_r (r = 1, 2,..., J). In this example, J = 19.

[0047] FIG. 4 is a view showing in detail the configuration of each strand bundle 22_r. Referring to FIG. 4, each strand bundle 22_r includes a plurality of strands 25. Let the number of strands 25 be K. In this example, K = 11. Each strand 25 is twisted in the Z direction. The cable 9 includes a total of H (= J × K) strands 25.

[0048] Each of the H strands 25 is classified into one of the first to J groups depending on which strand bundle 22 the strand belongs to. For example, if the strand 25 is included in the strand bundle 22_1, this strand is classified into the first group. If the strand 25 is included in the strand bundle 22_J, this strand is classified into the Jth group.

[0049] FIG. 5 is a diagram showing an example of volume data of cable 9 generated using a CT reconstruction algorithm. Referring to FIG. 5, the volume data 50 includes n×m×N voxels. Each of n and m is a natural number of 2 or more, for example, 256. N is a natural number of 2 or more, and is determined by observation conditions of X-ray CT or the like. A feature amount (pixel value) is assigned to each voxel. The feature amount is, for example, luminance. The length L represents the length of one side of each voxel, and in many cases, the value of the voxel size in X-ray CT observation corresponds to the length L as it is. An X coordinate, a Y coordinate, and a Z coordinate are assigned to each voxel.

[0050] FIG. 6 is a diagram for explaining a planar image defined by voxels having the same Z coordinate. Referring to FIG. 6, the center point of each voxel is also represented as point cn(X, Y, Z) in association with the X coordinate, the Y coordinate, and the Z coordinate of the voxel. A planar image defined by points cn(X, Y, Z) having the same Z coordinate is also referred to as a "normal slice image". In this example, normal slice images 200_1,... 200_(w - 1), 200_w,... 200_N are generated. Each normal slice image is also represented as normal slice image 200. A group of images composed of these normal slice images is also represented as a "normal slice image group". The normal slice image group corresponds to the volume data 50.

[0051] Each normal slice image 200 represents a slice image of cable 9 in the XY plane and includes a plurality of (in this example, n×m) pixels. To each pixel, the feature amount (luminance) of the voxel corresponding to the pixel and the coordinates (X, Y, Z) of point cn(X, Y, Z) are assigned. Each point cn(X, Y, Z) of normal slice image 200_w is separated from each point cn(X, Y, Z) of normal slice image 200_(w - 1) by a certain distance (length L) in the Z direction. In other words, each normal slice image 200 is separated by length L in the Z direction. The minimum Z coordinate and the maximum Z coordinate among the Z coordinates of the normal slice image group are also represented as Z1 and ZN, respectively.

[0052] FIG. 7 is a diagram showing the appearance of the cable in a comparative example where the cable is not curved. Referring to FIG. 7, in the comparative example, unlike the cable 9 (FIG. 2), the cable 9A is fixed by the jig 27 in a straight state along the Z direction. The structure of the cable 9A is the same as that of the cable 9 (see FIGS. 3 and 4).

[0053] FIG. 8 is a diagram showing an example of a plurality of normal slice images generated in the comparative example. Referring to FIG. 8, each normal slice image 210 is perpendicular to the Z direction (parallel to the XY plane) and represents the cross-sectional area of the cable 9A in the XY plane. This cross-sectional area represents the cross-section of each strand bundle (each strand) of the cable 9A.

[0054] The cable 9A is not curved in its longitudinal direction, and this longitudinal direction is always parallel to the Z direction. As a result, each normal slice image 210 is always generated so as to represent the cross-section of the cable 9A in a plane (XY plane) perpendicular to the longitudinal direction of the cable 9A. Further, each normal slice image 210 is generated at intervals of a certain distance (length L in this example). When each normal slice image 210 is generated in this way, the cross-section of the cable 9A can be analyzed with high accuracy using each normal slice image 210.

[0055] Referring to FIG. 2 again, an object such as the cable 9 can be frequently curved in its longitudinal direction depending on the actual use environment. Therefore, observing the object in a curved state by CT scan is extremely important from the viewpoint of evaluating the quality of the object. However, when the object is curved in its longitudinal direction (for example, when the object is the cable 9), the way the cross-sectional area of the object appears in a plurality of normal slice images may change significantly. This will be explained below.

[0056] Each of FIGS. 9 to 11 is a diagram showing an example of the normal slice image 200 in Embodiment 1. Referring to FIGS. 9 to 11, the normal slice images 200a to 200c are respectively generated corresponding to Za to Zc (FIG. 2). Each of the normal slice images 200a to 200c represents the cross-sectional area of the cable 9 (each strand bundle 22) in the XY plane. Each of the numerous thin lines in these images corresponds to the strand 25 included in the strand bundle 22.

[0057] In Embodiment 1, since the cable 9 is curved in its longitudinal direction, the normal slice image 200 often does not represent the cross-sectional area of the cable 9 in a plane perpendicular to the longitudinal direction. For example, while the cross-sectional area of the cable 9 in the normal slice image 200b is perpendicular to the longitudinal direction, each of the cross-sectional areas of the cable 9 in the normal slice images 200a and 200c is not perpendicular to the longitudinal direction. As a result, the shape of the cross-sectional area of the cable 9 (the shape of each strand 25 in the cross-sectional area and the shape of the entire cable 9 as an aggregate thereof) in the normal slice images 200a to 200c can vary significantly depending on the Z coordinate. In this case, it may not be possible to accurately analyze the internal structure of the cable 9 using the normal slice image 200.

[0058] The X-ray apparatus 1 according to Embodiment 1 has a configuration for dealing with such problems. This will be described below.

[0059] FIG. 12 is a block diagram showing the functional configuration of the control device 11 of the X-ray apparatus 1. Referring to FIG. 12, the control device 11 includes a first generation unit 110, a determination unit 115, a second generation unit 120, and a display control unit 122.

[0060] The first generation unit 110 generates normal slice images 200_1 to 200_N of the cable 9 according to the irradiation result of the X-ray 3a from the light source 3 to the cable 9. These normal slice images are generated corresponding to the volume data 50. The first generation unit 110 generates each of the normal slice images 200_1 to 200_N such that the coordinates (X, Y, Z) of each of the n×m pixels of the image are defined in the reference coordinate system. Specifically, the first generation unit 110 generates each normal slice image such that the Z coordinate of each pixel of the image is constant. The first generation unit 110 generates the normal slice images 200_1 to 200_N such that the Z coordinate of each of the n×m pixels of the normal slice image 200_f (2≤f≤N) is separated by a certain distance (length L) in the Z-axis direction from the Z coordinate of each of the n×m pixels of the (f - 1)th normal slice image (see FIG. 6).

[0061] The determination unit 115 determines first to Nth center points respectively indicating the centers of the cross-sectional areas of the cable 9 in the normal slice images 200_1 to 200_N (determines the center points of the cable 9 reflected in each normal slice image 200). For example, for each normal slice image 200, the determination unit 115 utilizes the fact that the cable 9 is brighter than the surrounding air in the image, binarizes the image, thereby dividing the image into a white area or a black area, and determines the center point of the cross-sectional area of the cable 9 in the image (specifically, the coordinates of the center point in the reference coordinate system) according to the result of the division. In one example, the determination unit 115 calculates the average value of the X coordinates (average X coordinate) of a plurality of pixels corresponding to the white area and the average value of the Y coordinates (average Y coordinate) of these pixels, and determines the point having the average X coordinate and the average Y coordinate as the X coordinate and the Y coordinate respectively as the center point of the cable 9. The determination unit 115 may determine the first to Nth center points of the cable 9 using a method such as deep learning. The method for determining these center points is not limited. In this embodiment, the center point of the cable 9 is determined by image processing for each normal slice image 200. However, for example, the cable 9 itself may be recognized as a three-dimensional object, and the center points may be determined collectively for all the normal slice images 200.

[0062] The second generation unit 120 generates the first to M-th (M is a natural number of 2 or more) cross-sectional images according to the volume data 50. Each of these cross-sectional images represents the cross-sectional area of the cable 9 in a plane substantially perpendicular to the local longitudinal direction of the cable 9. The image group composed of these cross-sectional images is also referred to as an "oblique slice image group". Each of these cross-sectional images is also referred to as an "oblique slice image". The second generation unit 120 includes a coordinate system setting unit 125, a coordinate conversion unit 130, and an assignment unit 135. These functions will be described in detail later. The display control unit 122 controls the display device 14 to display each oblique slice image in response to a user operation.

[0063] Each of the oblique slice images accurately represents a plane (cross-section) orthogonal to the longitudinal direction of the cable 9. As a result, the internal structure of the cable 9 can be analyzed with high accuracy based on the oblique slice image group. Hereinafter, when the number of oblique slice images is equal to the number of normal slice images (in the case of N = M: specifically, when one oblique slice image is corresponded to each normal slice image), how the second generation unit 120 generates the oblique slice images will be specifically described.

[0064] Each of FIGS. 13 to 19 is a diagram for explaining a specific method for generating an oblique slice image.

[0065] Referring to FIG. 13, the center point CPw (1 ≤ w ≤ N) of the cable 9 indicates the center point of the cross-sectional area of the cable 9 in the normal slice image 200_w. As described above, the coordinates (X coordinate, Y coordinate, and Z coordinate) of the center point CPw in the reference coordinate system are determined by the determination unit 115. Each center point CPw of the cable 9 is also represented as the center point CP. The vectors e X , e Y , e Z are unit vectors (basis vectors) on the X-axis, Y-axis, and Z-axis, respectively.

[0066] Referring to FIG. 14, the z(w) axis is locally defined from the center point CPw of the cable 9. In this example, it is defined by the vector from the center point CPw of the cable 9 to the center point CP(w+1) of the cable 9. The unit vector in the direction of the z(w) axis is also represented as the vector e z(w) . The direction of the vector e z(w) is approximately equal to the local longitudinal direction of the cable 9 as seen from the center point CPw of the cable 9. The center point CP(w+1) of the cable 9 is separated from the center point CPw of the cable 9 by a distance d(w) in the direction of the vector e z(w) . The angle formed by the z(w) axis and the Z axis is also represented as the angle θw. The angle θw is calculated according to the inner product of the vector e Z and the vector e z(w) . The x-component, y-component, and z-component of the vector e z(w) are also represented as e z(w)x , e z(w)y , and e z(w)z , respectively. The vector e z(w) is represented as follows. As shown in Equation 1 below, the magnitude of the vector e z(w) is normalized to 1.

[0067]

Equation

[0068] Referring to FIG. 15, the cross product ep(w) is the cross product of the vector e Z and the vector e z(w) . The x-component, y-component, and z-component of the cross product ep(w) are also represented as px, py, and pz, respectively. px, py, and pz are represented as follows.

[0069]

Equation

[0070] The cross product ep(w) (= e Z × e z(w) ) is represented as follows. As shown in Equation 3 below, the magnitude of the cross product ep(w) is normalized to 1.

[0071]

Number

[0072] The rotation matrix for rotating the vector in the reference coordinate system by an angle θw around the cross product ep(w) is also denoted as matrix R(θw). Matrix R(θw) corresponds to the operation of converting the Z-axis to the z(w)-axis and is expressed as follows.

[0073]

Number

[0074] Referring to FIG. 16, the x(w)-axis and the y(w)-axis are orthogonal to each other, and each of these axes is perpendicular to the z(w)-axis. The plane defined by the x(w)-axis and the y(w)-axis is also denoted as plane 305(w). Since plane 305(w) is defined as a plane perpendicular to the z(w)-axis, it is approximately equal to the plane perpendicular to the local longitudinal direction at the center point CPw of cable 9. Since matrix R(θw) corresponds to the operation of converting the Z-axis to the z(w)-axis, it also corresponds to the operation of converting the X-axis and the Y-axis to the x(w)-axis and the y(w)-axis, respectively.

[0075] The multiplication result of matrix R(θw) and vector e X is denoted as vector e x(w) The multiplication result of matrix R(θw) and vector e Y is denoted as vector e y(w) The multiplication result of matrix R(θw) and vector e Z is denoted as vector e z(w) Vectors e x(w) , e y(w) , e z(w) are expressed as follows in the reference coordinate system.

[0076]

Number

[0077] [Number]

[0078] [Number]

[0079] Substituting Equation (2) into Equation (7) leads to Equation (1). This is obvious from the definition of the matrix R(θw).

[0080] Vector e x(w) , e y(w) , e z(w) are unit vectors on the x(w)-axis, y(w)-axis, and z(w)-axis, respectively. The coordinate system with these unit vectors as basis vectors is also referred to as the "w-th local coordinate system". In this example, the w-th local coordinate system is locally set for each center point CPw of the cable 9, and has the direction of the vector e z(w) as the z-axis direction. In this embodiment, it is assumed that the origin of the w-th local coordinate system is the center point CPw. The coordinates in the reference coordinate system can be converted to the coordinates in the w-th local coordinate system according to the angle θw (specifically, the matrix R(θw)).

[0081] The image defined by the plane 305(w) (the plane image orthogonal to the vector e z(w) ) corresponds to the w-th diagonal slice image (300_w) (1 ≤ w ≤ N = M). Each diagonal slice image 300 includes n × m pixels (n, m are natural numbers). A feature amount is assigned to each pixel of the diagonal slice image 300. In this example, the feature amount is luminance. The control device 11 (the second generation unit 120) generates the diagonal slice image 300_w such that the coordinates of each of the n × m pixels of the image are defined in the w-th local coordinate system. Specifically, the control device 11 generates the diagonal slice image 300_w such that the coordinates of each pixel of the image are defined in the plane 305(w) (the z(w)-coordinate of each pixel is constant).

[0082] According to the above procedure, diagonal slice images 300_1, … 300_w, … 300_N are generated (see Fig. 17).

[0083] Hereinafter, the coordinate system setting unit 125, the coordinate conversion unit 130, and the assignment unit 135 (all shown in Fig. 12) will be described.

[0084] The coordinate system setting unit 125 sets the z(w)-axis for each value of w according to the center points CP1, … CPw, … CPN of the cable 9. The coordinate system setting unit 125 defines vectors e z(w) , e y(w) , e z(w) for each value of w, and sets the w-th local coordinate system by defining them, thereby defining the plane 305(w) (see Fig. 16).

[0085] The coordinate conversion unit 130 converts the coordinates (x, y, z) of each of the n×m pixels of the diagonal slice image 300_w in the w-th local coordinate system into the w-th corresponding coordinates (XYZ representation), which are the coordinates of the pixel in the reference coordinate system, according to the angle θ(w). Specifically, the coordinate conversion unit 130 adds the components (XYZ representation) of the vectors obtained by multiplying the x-coordinate and the y-coordinate (both scalars) of the pixel in the w-th local coordinate system by the vectors e x(w) , e y(w) respectively to the coordinate values of the center point CPw in the reference coordinate system to convert the local coordinates of the pixel into the w-th corresponding coordinates of the pixel. By the above conversion process, the coordinates of each pixel of the diagonal slice image in the reference coordinate system are derived.

[0086] Referring to FIG. 18, the assignment unit 135 assigns the feature amount of each of the n×m pixels pxe of the diagonal slice image 300_w. Specifically, the assignment unit 135 assigns to each of these pixels according to the feature amount of at least one voxel of the volume data 50 having coordinates within the neighborhood region R(X, Y, Z) of the w-th corresponding coordinates of the pixel. The feature amount is the same as the feature amount of at least one pixel having coordinates within the neighborhood region R(X, Y, Z) of the w-th corresponding coordinates of the pixel. The at least one pixel is included in a plurality of pixels (n×m×N pixels) of the normal slice image group. The length of one side of the pixel pxe is, for example, length L. The length of one side is not limited, but in order to facilitate subsequent data analysis, it is often aligned with the length of one side of the pixels of the original normal slice image 200.

[0087] Referring to FIG. 19, each point of the n×m pixels pxe of the diagonal slice image 300_w is also represented as a point p(x, y, z) in association with the x coordinate, y coordinate, and z coordinate of this point in the w-th local coordinate system. The points pna(X, Y, Z) to pnh(X, Y, Z) of the pixels of the normal slice image group are within the neighborhood region R(X, Y, Z) of the point p(x, y, z). The neighborhood region R(X, Y, Z) is defined as a region within a range of a predetermined threshold distance from the point p(x, y, z) (specifically, the point determined by its XYZ representation) in the reference coordinate system.

[0088] For example, the assignment unit 135 assigns the feature amount of the pixel of the point p(x, y, z) to be the same as the feature amount of the pixel having the point closest to the point p(x, y, z) among pna(X, Y, Z) to pnh(X, Y, Z). In this example, the closest point is the point pne(X, Y, Z) where the distance ds from the point p(x, y, z) is the smallest. The assignment unit 135 may weight the feature amounts of pna(X, Y, Z) to pnh(X, Y, Z) according to the distances between the point p(x, y, z) and each of pna(X, Y, Z) to pnh(X, Y, Z), and assign the feature amount of the pixel of the point p(x, y, z) according to the result of the weighting.

[0089] If the w-th corresponding coordinates of the pixel pxe of the diagonal slice image 300_w match any of the coordinates of the n×m×N pixels of the normal slice image group at the reference coordinates (when the point p(x, y, z) matches any of the points pna(X, Y, Z) to pnh(X, Y, Z)), it is preferable that the feature amount of the pixel pxe is the feature amount of the pixel having the coordinates that match the w-th corresponding coordinates. However, in most cases, the w-th corresponding coordinates do not match any of the coordinates of the n×m×N pixels of the normal slice image group. Therefore, it is important to appropriately determine the feature amount of each pixel pxe.

[0090] In Embodiment 1, the feature amount of each pixel pxe of the diagonal slice image 300_w is generated according to the feature amount of at least one of the n×m×N pixels of the normal slice image group within the neighborhood region R(X, Y, Z) of the w-th corresponding coordinates of the pixel (for example, the pixel having the point pne(X, Y, Z)). Therefore, even when the w-th corresponding coordinates of the pixel pxe do not match any of the coordinates of the n×m×N pixels of the normal slice image group, the feature amount of each pixel pxe of the diagonal slice image 300_w can be appropriately determined.

[0091] FIG. 20 is a diagram showing how the cross-sectional area of the cable 9 is represented in each diagonal slice image 300. Referring to FIGS. 20(A) to (C), the inventor confirmed that the way (shape) the cross-sectional area of the cable 9 appears is substantially constant and does not change in the diagonal slice images 300_1 to 300_N (for example, diagonal slice images 300_a to 300_c) (1 ≦ a < b < c ≦ N). The inventor also confirmed that in these slice images, the way each strand 25 appears in the cross-sectional area of the cable 9 is also substantially constant.

[0092] FIG. 21 is a flowchart showing an example of the processing executed by the control device 11 in Embodiment 1. This flowchart shows the procedure of a method for generating an image of the cable 9 using the X-ray CT apparatus 1, and is started when the detection of the X-ray 3a by the X-ray detector 7 is completed. Each step of this flowchart is performed by the CPU 11a executing a program stored in the memory 11b. Hereinafter, the steps are abbreviated as "S".

[0093] Referring to FIG. 21, the control device 11 generates a normal slice image group (volume data 50) according to the irradiation result of the X-ray 3a on the cable 9 (S105).

[0094] The control device 11 determines the center point CP of the cable 9 in each normal slice image 200 of the image group. In other words, the control device 11 determines the coordinates of the center points CP1,... CPw,... CPN of the cable 9 in the normal slice image 200 in the reference coordinates.

[0095] The control device 11 generates an oblique slice image group according to the normal slice image group (the result of the determination of these coordinates) (S115).

[0096] The control device 11 controls the display device 14 to display an oblique slice image 300 (see FIG. 20 for example) in response to a user operation (S120).

[0097] FIG. 22 is a flowchart showing the detailed procedure of the process of S115 in FIG. 21. At the start of this flowchart, it is assumed that the value of w is 1.

[0098] Referring to FIG. 22, the control device 11 sets the w-th local coordinate system (plane 305(w)) by setting the z(w) axis according to the result of the determination of the coordinates of the center points CP1,... CPw,... CPN of the cable 9 in the normal slice image group (S110) (S155).

[0099] For each of the n×m pixels pxe of the oblique slice image 300_w corresponding to the plane 305(w), the control device 11 converts the coordinates of the pixel in the w-th local coordinate system (the coordinates of the point p(x, y, z)) into the w-th corresponding coordinates (XYZ representation) of the pixel (S160).

[0100] For each pixel pxe, the control device 11 determines the feature amount of at least one pixel of the normal slice image group having coordinates within the neighborhood region R(X, Y, Z) of the w-th corresponding coordinates of the pixel (S165).

[0101] The control device 11 assigns the feature amount of each pixel pxe according to the feature amount of at least one pixel determined in relation to the pixel. Thereby, the oblique slice image 300_w is generated (S170).

[0102] The control device 11 reflects whether the value of w has reached N (S175). If the value of w has not yet reached N (NO in S175), the control device 11 increments the value of w (S180). Thereafter, the process returns to S155. As a result, S155 to S170 are repeated until the value of w reaches N, and an oblique slice image group is generated. When the value of w reaches N (YES in S175), the process proceeds to S120 in FIG. 21.

[0103] As described above, according to the first embodiment, an oblique slice image group is generated using the normal slice image group (volume data 50). Each oblique slice image accurately represents the cross-sectional area of the cable 9 in a plane (cross-section) orthogonal to the local longitudinal direction of the cable 9. As a result, the internal structure of the cable 9 can be accurately analyzed.

[0104] <Modification Example of the First Embodiment> In Embodiment 1, the interval between the diagonal slice image 300_w and the diagonal slice image 300_(w + 1) (the distance d(w) between diagonal slice images in the local longitudinal direction) is not necessarily constant. However, from the viewpoint of highly accurate analysis of the cross-section of the cable 9, it is preferable that this interval is constant (for example, length L) in the same manner as the normal slice images 200_1 to 200_N.

[0105] In this modification, the second generation unit 120 generates a group of diagonal slice images such that the above interval is constant (for example, length L). In this case, the number (M) of diagonal slice images is larger than the number (N) of normal slice images (M > N).

[0106] Each of FIGS. 23 and 24 is a diagram for explaining a method for generating a group of diagonal slice images in this modification.

[0107] Referring to FIG. 23, the diagonal slice images in this modification are also represented as diagonal slice images 350_k (k is a natural number: 1 ≦ k ≦ M). As described below, also in this modification, the control device 11 (second generation unit 120) generates diagonal slice images 350_1 to 350_M according to the normal slice images 200_1 to 200_N (volume data 50).

[0108] The method for generating the diagonal slice image 350_1 is the same as the method for generating the diagonal slice image 300_1. In other words, the diagonal slice image 350_1 is the same as the diagonal slice image 300_1.

[0109] A point included in the diagonal slice image 350_k is also denoted as point Pk. For example, the diagonal slice image 350_1 includes point P1. Similarly, the diagonal slice images 350_2, …, 350_M each include points P2, …, PM, respectively. In the following description, it is assumed that point P1 is the same as the center point CP1 of the cable 9 in the normal slice image 200_1, but it may be different from this center point CP1. Similarly, points P2, …, PM are each assumed to be the center points of the diagonal slice images 350_2, …, 350_M, respectively, but are not limited to these center points.

[0110] The vicinity region of point Pk is also denoted as vicinity region RG(k). For example, the vicinity region RG(1) is the vicinity region of point P1. The vicinity region RG(k) is a region within a predetermined reference distance from point Pk. Each of the center points CP1, CP2, CP3, … of the cable 9 is the same as that in Embodiment 1 (see, for example, FIGS. 13 to 15).

[0111] The method for generating each of the diagonal slice images 350_2, …, 350_M is different from the method for generating each of the diagonal slice images 300_2, …, 350_N. This will be described below.

[0112] Referring to FIG. 24, each diagonal slice image is also denoted as diagonal slice image 350. A vector from the center point CPi of the cable 9 to the center point CPj (1 ≦ i < j ≦ N) of the cable 9 within the vicinity region RG(k) is also denoted as vector Vij(k). In this modification, j = i + 2, but the value of j may be other than 2 as long as both of the center points CPi and CPj of the cable 9 are within the vicinity region Rk.

[0113] The z(k)-axis is locally defined from point Pk and is defined by vector Vij(k) in this example (having the same direction as the direction of vector Vij(k)). The unit vector in the direction of z(k) is denoted as vector e z(k) as well. The angle formed by the Z-axis and vector e z(k) is also denoted as angle θ(k). Vector e z(k)The direction is approximately equal to the local longitudinal direction of the cable 9 as seen from the point Pk. The vector e x(k) and the vector e y(k) are each defined to be orthogonal to the vector e z(k) . The vector e x(k) and the vector e y(k) are orthogonal to each other. These vectors define a plane orthogonal to the vector e z(k) . This plane is also denoted as the plane 355(k). The plane 355(k) is approximately equal to the plane orthogonal to the local longitudinal direction at the point Pk.

[0114] The vector e x(k) , e y(k) , e z(k) as the base vectors is also referred to as the "k-th local coordinate system". In this example, the k-th local coordinate system is locally set for each point Pk and has the direction of the vector e z(k) as the z-axis direction. In this variant, the origin of the k-th local coordinate system is assumed to be the point Pk. The coordinates in the k-th local coordinate system can be converted to the coordinates in the reference coordinate system according to the angle θ(k).

[0115] The oblique slice image 350_k corresponds to an image representing the cross-sectional area of the cable 9 in the plane 355(k). The coordinates of each of the n×m pixels px of the oblique slice image 350_k are defined in the k-th local coordinate system. The length of one side of the pixel px is length L in this example.

[0116] The point P(k + 1) is separated from the point Pk by a certain predetermined distance in the direction of the vector e z(k) . In other words, the points of the point Ph (2 ≤ h ≤ M) are separated from the point P(h - 1) by a predetermined distance in the direction of the vector e z(h-1) . This distance is assumed to be equal to the length L in this example, but is not limited to the length L.

[0117] In this way, the point P(k + 1) is determined according to the point Pk. Therefore, the control device 11 can derive the coordinates of the point P(k + 1) according to the coordinates of the point Pk in the reference coordinate system. As a result, the control device 11 can derive the coordinates (X coordinate, Y coordinate, and Z coordinate) of the points P2, P3, … PM according to the coordinates of the point P1 in the reference coordinate system. The Z coordinate of the point PM is determined to be equal to or less than ZN (FIG. 6).

[0118] The control device 11 determines the coordinates of the center points CPi and CPj of the cable 9 (the coordinates of these center points in the reference coordinate system) within the vicinity region RG(k) according to the coordinates of the point Pk, and thereby determines the vector Vij(k). The control device 11 sets the z(k) axis according to the vector Vij(k), and thereby the vectors e x(k) , e y(k)) , e z(k) are defined. The control device 11 sets the k-th local coordinate system according to the vectors e x(k) , e y(k) , e z(k) . For each value of k, the control device 11 defines the plane 355(k) by setting the k-th local coordinate system, and thereby generates the oblique slice images 350_2, … 350_M.

[0119] The method for assigning the feature amount of each pixel to each of the n × m pixels px of each oblique slice image 350 is the same as the method described in FIGS. 18 and 19. Specifically, the control device 11 (coordinate conversion unit 130) converts the coordinates of each of the n × m pixels px of the oblique slice image 350_k into the k-th corresponding coordinates (XYZ representation), which are the coordinates of the pixel in the reference coordinate system, according to the angle θ(k). More specifically, the control device 11 converts the x coordinate and the y coordinate (both scalars) of the pixel in the k-th local coordinate system into the vectors e x(k) , e y(k)By adding the components (XYZ representation) of the vectors obtained by multiplying each by the coordinates of the point Pk in the reference coordinate system, the local coordinates of the pixel can be converted into the k-th corresponding coordinates of the pixel. After the above coordinate transformation, the control device 11 (assignment unit 135) assigns the feature amount of each of the n×m pixels px of the diagonal slice image 350_k according to the feature amounts of at least one pixel having coordinates within the neighborhood region R(X, Y, Z) of the k-th corresponding coordinates of the pixel. The at least one pixel is included in the n×m×N pixels of the normal slice images 200_1 to 200_N.

[0120] FIG. 25 is a diagram showing experimental data indicating the change in the X coordinate of the center point group of the cable 9 in the normal slice image group, the point group obtained by performing a smoothing process on the center group, and the point group composed of each point Pk in the diagonal slice image group.

[0121] Referring to FIG. 25, the point set 505 shows the transition of the X coordinate of the center point CPw of the cable 9 reflected in the normal slice image group. The point set 510 shows the transition of the X coordinate of the points obtained as a result of performing a smoothing process in the Z-axis direction on these center points CPw. This smoothing process is executed to reduce the discontinuity (FIG. 26) of the point set 505. The point set 515 shows the transition of the X coordinate of the point group composed of each point Pk in the diagonal slice image group. In this example, for the sake of convenience, these point sets are shown as if they were almost the same, but in reality, as will be described below, these point sets are different from each other.

[0122] FIG. 26 is a diagram showing in detail the point sets 505, 510, and 515 in the region RNG of FIG. 25. For ease of illustration, each of these point sets is linked by a line.

[0123] As shown by the point sets 505, 510, and 515, for each Z coordinate, there is a slight difference between the X coordinate of point Pk in the oblique slice image group and the X coordinate of the center point CPw of cable 9 (or the X coordinate of the smoothed point) in the normal slice image group. However, this difference is at most the length L (the length of one side of a pixel px), and is so small that it can be ignored from a practical perspective. The inventor has confirmed that this difference does not cause any inconvenience.

[0124] Note that also in this modification example, similar to the example of FIG. 20, it was confirmed that in the oblique slice images 350_1 to 350_M, the way (shape) in which the cross-sectional area of cable 9 is reflected is almost constant and does not change.

[0125] As described above, in this modification example, the control device 11 (the second generation unit 120) generates the oblique slice images 350_1 to 350_M such that the distance between the oblique slice images 350 in the k-th local coordinate system is equal to the length L.

[0126] By adopting such a configuration, the oblique slice images 350_1 to 350_M are also generated at regular intervals along the longitudinal direction of cable 9. Therefore, the internal structure of cable 9 can be analyzed with higher accuracy.

[0127] <Embodiment 2> Since each of the oblique slice images 350_1 to 350_M represents the cross-sectional area of cable 9 in the plane 355(k) (FIG. 24), it also represents the cross-section of each strand 25 of cable 9 in the plane 355(k). These oblique slice images are useful for tracing the path of each strand 25 in cable 9 (hereinafter, also referred to as "strand tracing").

[0128] In Embodiment 2, the control device 11 controls the communication device 15 to transmit data indicating the diagonal slice images 350_1 to 350_M to the information processing device 16. The communication device 18 of the information processing device 16 receives the data. The control device 17 executes generatrix tracing according to the data. In the description of Embodiment 2 (and its modification), it is assumed that the diagonal slice images 350_1 to 350_M are mainly used as the diagonal slice image group, but the diagonal slice images 300_1 to 300_M may be used instead of these images.

[0129] FIG. 27 is a diagram for explaining how the control device 17 traces the path of each generatrix 25. Referring to FIG. 27, the control device 17 determines the coordinates (central coordinates) of the center point of the cross-section of the generatrix 25 included in the diagonal slice image 350_k. This determination may use classical binarization processing or advanced processing such as deep learning. This center point is, for example, the center point cp_1,..., the center point cp_s,..., or the center point cp_H′ (1 ≤ s ≤ H′, where H = J × K). H′ represents the number of generatrices 25 included in the diagonal slice image 350_k. In practice, H′ is a value close to H and often varies depending on the quality of image analysis. In the following description, assuming an ideal situation, H′ is equal to H. The generatrix 25 having the center point cp_s is also represented as the generatrix 25(s).

[0130] According to the determined center point having the center coordinates and the center point having the coordinates closest to the determined coordinates among the coordinates of the center points of the cross-sections of the plurality of generatrices 25 included in the diagonal slice image 350_(k + 1), the line segment SG(s,k) is defined. When the line segments SG(s,k) of the generatrix 25(s) are aggregated for each value of k, the path pth of the generatrix 25(s) (the locus of the center point cp_s in the diagonal slice images 350_1 to 350_M) is traced.

[0131] By executing the process of tracing the path pth for each generatrix 25 in this way, generatrix tracing is executed. The path pth of each generatrix 25 derived by generatrix tracing accurately represents the actual path of each generatrix 25.

[0132] The strand tracking is premised on the fact that the way the cross-section of each strand 25 is imaged (shape) does not vary significantly from one slice image to another. In Embodiment 2, since the way the cross-section of each strand 25 is imaged in the oblique slice images 350_1 to 350_M is almost constant and does not change regardless of the value of k, strand tracking can be accurately performed using the group of oblique slice images.

[0133] FIG. 28 shows data representing the transition of the center coordinates of each strand 25. Specifically, data 700 shows the k-th center coordinate of each strand 25 for each oblique slice image 350_k (k-th local coordinate system). The k-th center coordinate is the center coordinate of the strand 25 in the oblique slice image 350_k. These center coordinates are defined in the corresponding k-th local coordinate system. The data 700 is stored in the storage device 19A.

[0134] FIG. 29 is a diagram for explaining the reason why the path of each strand 25 can be accurately tracked in Embodiment 2. FIG. 29(A) shows the reference coordinate system, each local coordinate system, and the plane of each oblique slice image 350 (plane 355(k)). The inclination of these planes varies for each local coordinate system.

[0135] FIG. 29(B) shows the virtual coordinate system and the plane of each oblique slice image 350 in the virtual coordinate system. The virtual coordinate system is used to represent the center coordinates of the cross-sectional area of each strand 25 in the oblique slice image 350 under the assumption that each z(k) axis is aligned. The center coordinate is defined in the corresponding k-th local coordinate system. The virtual coordinate system is defined by the x′ axis, y′ axis, and z′ axis. Each of these axes is defined by the corresponding k-th local coordinate system. In the virtual coordinate system, the cable 9 is represented as if it is straight along the z′ axis. As a result, the amount of movement of the center coordinates of the cross-sectional area of each strand 25 between the oblique slice images 350_k and 350_(k + 1) becomes small, facilitating strand tracking.

[0136] FIG. 30 is a block diagram showing a functional configuration of a control device 17 of the information processing apparatus 16 in the second embodiment. Referring to FIG. 30, the control device 17 includes a center coordinate determination unit 170, a coordinate conversion unit 175, a classification unit 180, a trajectory image generation unit 185, and a display control unit 190.

[0137] For each generatrix 25, the center coordinate determination unit 170 determines a k-th center coordinate, which is the center coordinate of the generatrix in the oblique slice image 350_k. The k-th center coordinate is, for example, the coordinate of a center point cp_s (see FIG. 27) in the k-th local coordinate system.

[0138] For each generatrix 25, the coordinate conversion unit 175 converts the k-th center coordinate of the generatrix into a k-th corresponding center coordinate (XYZ representation), which is the center coordinate of the generatrix in the reference coordinate system according to the angle θ(k). The functions of the classification unit 180, the trajectory image generation unit 185, and the display control unit 190 will be described in detail later.

[0139] When deriving the trajectory (path pth) of the center coordinate of each generatrix 25, it is assumed that the way the cross-section of the generatrix appears (shape) does not change significantly from image to image. However, in the normal slice image group, the way the cross-section of each generatrix 25 appears may change significantly depending on its Z coordinate (see FIGS. 9 to 11). Therefore, it is difficult to accurately derive the trajectory of the center coordinate of each generatrix 25 in the reference coordinate system using the normal slice image group.

[0140] In the second embodiment, due to the functions of the center coordinate determination unit 170 and the coordinate conversion unit 175, the trajectory of the center coordinate of each generatrix 25 is determined using the oblique slice image group. As described above, in the oblique slice image group, the way the cross-section of each generatrix 25 appears is almost constant and does not change regardless of its z coordinate (see FIG. 20). Therefore, the trajectory of the k-th center coordinate can be accurately derived using the oblique slice image group. As a result, as described below, the trajectory of the center coordinate of each generatrix 25 in the reference coordinate system can be accurately derived by the coordinate conversion process by the coordinate conversion unit 175 (accurate generatrix tracking can be performed).

[0141] FIG. 31 is a diagram for explaining the coordinate conversion process by the coordinate conversion unit 175. Referring to FIG. 31, the data 700 is the same as that in FIG. 28. The coordinate group cg′(s) represents a set of central coordinates of the generatrix 25_(s) in the virtual coordinate system (1 ≤ s ≤ H), and represents the locus of the center point cp_s of the generatrix 25_(s). The data 700 represents the coordinate group cg′(s) for each generatrix 25_(s). Based on the coordinate group cg′(s), the path pth of the corresponding generatrix 25(s) is derived.

[0142] The data 750 indicates the k-th corresponding central coordinates (XYZ representation of the center point cp_s) of each generatrix 25 for each diagonal slice image 300_k. By the coordinate conversion process by the coordinate conversion unit 175, the data 700 is converted into the data 750. The data 750 is stored in the storage device 19A. The coordinate group cg(s) corresponds to a set of central coordinates of the generatrix 25_(s) in the reference coordinate system (1 ≤ s ≤ H), and represents the locus of the center point cp_s of the generatrix 25_(s). The data 750 represents the coordinate group cg(s) for each generatrix 25_(s).

[0143] The inventor confirmed using the data 700 that the locus of the k-th central coordinates (specifically, its x coordinate and y coordinate) of each generatrix 25 in the diagonal slice images 350_1 to 350_M changes in a periodic pattern. For example, it was confirmed that the x coordinate (xs_k) and y coordinate (ys_k) of the generatrix 25(s) in the virtual coordinate system change in a periodic pattern.

[0144] FIG. 32 is another diagram for explaining the coordinate conversion process by the coordinate conversion unit 175. Referring to FIG. 32, (A) represents the k-th central coordinates (coordinates of the center point cp_s) of the generatrix 25 in the virtual coordinate system. This coordinate is determined based on the vectors e x(k) and the vector e y(k) and is assumed to be (α, β). The starting points of the vectors e x(k) and the vector e y(k) are the center point of the diagonal slice image 350_k (the point Pk as the origin of the k-th local coordinate system).

[0145] Due to the conversion process by the coordinate conversion unit 175, (α, β) is converted to its corresponding center coordinates (α´, β´) (see Fig. 32(B)). These coordinates are defined based on the vector e x(k) and the vector e y(k) in the reference coordinate system. The above conversion process is performed for each generatrix 25(s) by adding the components (XYZ representation) of the vector Vαβ obtained by multiplying the x - coordinate and y - coordinate (α, β) of the center coordinates of the generatrix in the k - th local coordinate system by the vectors e x(k) , e y(k) respectively to the coordinate values of the point Pk in the reference coordinate system. The above conversion process is executed for each generatrix 25(s) (center point cp_s) and for each oblique slice image 300_k. Thereby, the data 750 is generated.

[0146] Referring to Fig. 30 again, the functions of the classification unit 180, the trajectory image generation unit 185, and the display control unit 190 will be described.

[0147] The classification unit 180 executes a classification process for each of the H(=J×K) generatrices 25 to classify the generatrix into one of a plurality of groups according to the trajectory (path pth) of the k - th center coordinates of the generatrix. In this example, the classification process is executed for each value of s according to the coordinate group cg′(s). Thereby, it is determined which of the generatrix bundles 22_1~22_J each of the H generatrices 25 is included in. The classification process is executed using a predetermined clustering algorithm such as the k - means method. The method for the classification process is not particularly limited as long as it uses the data 700.

[0148] When the cable is not curved as in the foregoing comparative example (when it is straight along the Z direction), the cross-sectional areas (cross-sections) of the cables are parallel to each other (see FIGS. 7 and 8). In this case, for example, the locus of the center coordinates (x coordinate and y coordinate) of each strand 25 in the cable 9A changes in a predetermined periodic pattern in the normal slice images 210_1 to 210_N (reference coordinate system). Therefore, each strand of the cable 9A can be appropriately classified by a predetermined algorithm such as the k-means method.

[0149] On the other hand, when the cable is curved (see FIG. 2), the locus of the center coordinates of each strand 25 of the cable 9 does not change in such a pattern in the normal slice images 200_1 to 200_N. As a result, there is a possibility that each strand 25 may not be appropriately classified by the k-means method or the like. The inventor focused on the fact that the locus (path pth) of the k-th center coordinates of each strand 25 changes in a periodic pattern in the oblique slice images 350_1 to 350_M.

[0150] According to the function of the classification unit 180, since each strand 25 is classified according to the locus of the k-th center coordinates, the above situation is avoided. As a result, each strand 25 can be appropriately classified using a predetermined algorithm such as the k-means method.

[0151] The locus image generation unit 185 generates a locus image representing the locus of the k-th corresponding center coordinates of each strand 25 in the reference coordinate system and the result of the classification process for the strand. The locus image may represent only the locus of the k-th corresponding center coordinates of each strand 25. In this case, the locus image may be a character image representing only the numbers corresponding to the locus (for example, the numbers represented by the coordinate groups cg(1),... cg(H) in FIG. 31). The locus image will be described in detail later. The display control unit 190 controls the display device 19B to display the locus image in response to a user operation.

[0152] FIG. 33 is a diagram for explaining the results of strand tracking and classification processing in the second embodiment.

[0153] Referring to FIG. 33, the virtual trajectory image 810 represents the trajectory (thin line) of the k-th center coordinates of each generatrix 25 in the virtual coordinate system. Specifically, the virtual trajectory image 810 shows, for each generatrix 25, the trajectory (path pth) of a plurality of coordinate points each having coordinates included in the coordinate group cg′(r). The virtual trajectory image 810 further shows, depending on the luminance of the thin line corresponding to each generatrix, to which group among the first to J-th groups each generatrix 25 is classified (the result of the classification process).

[0154] The trajectory image 820 is generated by the conversion process of the coordinate conversion unit 175 using the virtual trajectory image 810. The trajectory image 820 represents the trajectory of the k-th corresponding center coordinates of each generatrix 25 in the reference coordinate system. Specifically, the trajectory image 820 shows, for each generatrix 25, the trajectory of a plurality of coordinate points each having coordinates included in the coordinate group cg(r). The trajectory image 820 shows, depending on the luminance of the thin line corresponding to each generatrix, to which group among the first to J-th groups each generatrix 25 is classified (the result of the classification process).

[0155] Each of the virtual trajectory image 810 and the trajectory image 820 is preferably a color image. In this case, each of these images shows, depending on the color of the thin line corresponding to each generatrix, to which group among the first to J-th groups each generatrix 25 is classified. For example, the color of each thin line is determined so that the colors of the thin lines corresponding to the generatrices 25 belonging to the same group are the same.

[0156] According to the trajectory image 820, for each generatrix 25, the trajectory of the k-th corresponding center coordinates of the generatrix in the reference coordinate system and the result of the classification process for each generatrix can be easily recognized by the user. As a result, even when the cable 9 is curved, the state of each generatrix 25 inside the cable 9 in the reference coordinate system (real space) can be appropriately and easily recognized by the user. Note that the trajectory image may be an image representing a two-dimensional graph showing the correspondence between the X coordinate or the Y coordinate of the trajectory and the Z coordinate, rather than a three-dimensional image like FIG. 33.

[0157] FIG. 34 is a flowchart showing an example of the processing executed by the control device 17 in the second embodiment. At the start of this flowchart, the value of k is assumed to be 1. Each step of the flowchart in FIG. 34 is performed by the control device 17 (CPU 17a) executing a program stored in the memory 17b.

[0158] Referring to FIG. 34, the control device 17 determines the k-th center coordinates in the diagonal slice image 350_k for each prime line 25 (S305).

[0159] The control device 17 converts the k-th center coordinates of each prime line 25 into the k-th corresponding center coordinates in the reference coordinate system according to the angle θ(k) (S310).

[0160] The control device 17 determines whether k has reached M (S312). Specifically, the control device 17 derives the Z coordinate of the point Pk in the reference coordinates and determines whether this Z coordinate exceeds ZN (FIG. 6). If k is less than M (NO in S312), the control device 11 increments the value of k (S313). Thereafter, the process returns to S305. As a result, until k reaches M (YES in S312), S310 and S312 are repeated, thereby generating the data 700 and 750. When k reaches M (when the Z coordinate of the point Pk exceeds ZN), the process proceeds to S315.

[0161] The control device 17 executes a classification process for each prime line 25 according to the data 700 (S315). The control device 17 generates a trajectory image 820 according to the data 750 (the trajectory of each prime line 25) and the result of the classification process (S320). Thereafter, the control device 17 controls the display device 19B to display the trajectory image 820 in response to a user operation (S325).

[0162] In order to accurately perform the strand tracking, it is assumed that the way (shape) the cross-section of each strand 25 is imaged does not vary significantly from image to image. In the normal slice images 200_1 to 200_N, the way the cross-section of each strand 25 is imaged may vary significantly depending on the Z coordinate (see FIGS. 9 to 11). Therefore, it is difficult to accurately perform the strand tracking of the cable 9 curved in the longitudinal direction using the normal slice image group.

[0163] In contrast, in Embodiment 2, the locus of the center coordinates of each strand 25 is determined using the oblique slice image group. In the oblique slice image group, for the cross-section of each strand 25, only the position changes depending on the z(k) coordinate, and its shape is almost constant and does not change (see FIG. 20). As a result, strand tracking can be accurately performed using the oblique slice image group.

[0164] <Modification Example of Embodiment 2> In this modification example, for each value of k, the control device 17 generates the k-th high-resolution image by performing image processing for increasing the resolution of the oblique slice image 350_k. The above image processing is super-resolution processing in this example, but is not limited thereto. The inventor has confirmed that various image processes such as super-resolution processing using a convolutional neural network or segmentation are effective. As will be described below, in strand tracking, each high-resolution image is used in place of the oblique slice image that is the generation source of the image.

[0165] FIG. 35 is a block diagram showing the functional configuration of the control device 17 of the information processing device 16 in this modification example. Referring to FIG. 35, the control device 17 includes a super-resolution processing unit 165, a center coordinate determination unit 170A, a coordinate conversion unit 175A, a classification unit 180A, a locus image generation unit 185A, and a display control unit 190A.

[0166] For each value of k, the super-resolution processing unit 165 generates a high-resolution image 380_k by performing super-resolution processing on the diagonal slice image 350_k. The super-resolution processing unit 165 thereby generates high-resolution images 380_1 to 380_M. These high-resolution images are also referred to as the "high-resolution image group".

[0167] For each generatrix 25, the center coordinate determination unit 170A determines the k-th center coordinate (high-resolution k-th center coordinate), which is the center coordinate of the generatrix in the high-resolution image 380_k.

[0168] For each generatrix 25, the coordinate conversion unit 175A converts the k-th center coordinate of the generatrix into the k-th corresponding center coordinate (high-resolution k-th corresponding center coordinate (XYZ representation)), which is the center coordinate of the generatrix in the reference coordinate system according to the angle θ(k). The method of this conversion process is the same as the method of the conversion process by the coordinate conversion unit 175 in the second embodiment, except that the high-resolution image 380_k is used instead of the diagonal slice image 350_k.

[0169] According to the super-resolution processing unit 165, the resolution of the diagonal slice image 350_k is increased by super-resolution processing. As a result, the portion representing the cross-section of each generatrix 25 in the high-resolution image 380_k becomes clearer. Consequently, the center (k-th center coordinate) of each generatrix 25 is determined with higher accuracy by the center coordinate determination unit 170A. Therefore, according to the conversion process by the coordinate conversion unit 175A, the locus (coordinate group cg(s)) of the center coordinates of each generatrix 25 in the reference coordinate system can be derived with higher accuracy.

[0170] For each of the H (= J × K) generatrices 25, the classification unit 180A executes a classification process of classifying the generatrix into one of a plurality of groups according to the locus (cg′(r)) of the k-th center coordinate of the generatrix.

[0171] For each of the H generatrices 25, the locus image generation unit 185A generates a locus image (an image similar to the locus image 820) representing the locus of the k-th corresponding center coordinates of the generatrix in the reference coordinate system and the result of the classification process for the generatrix. The display control unit 190A controls the display device 19B to display the locus image in response to a user operation.

[0172] Image processing such as super-resolution processing or segmentation assumes that the way the cross-section of each generatrix 25 appears does not vary significantly from image to image. Therefore, even if image processing such as super-resolution processing is performed on each of the normal slice images 200_1 to 200_N, the image processing may not succeed, and there is a possibility that the center of each generatrix 25 cannot be determined with high accuracy as described above. The reason is that in these normal slice images, the way the cross-section of each generatrix 25 appears can vary significantly depending on the Z coordinate (see FIGS. 9 to 11).

[0173] In contrast, in this modification, the super-resolution processing is executed for each oblique slice image 350_k. In these images, the cross-section of each generatrix 25 only changes in position depending on the z(k) coordinate, and its shape is almost constant and does not change. Therefore, the super-resolution processing succeeds, and thereby the center of each generatrix 25 can be determined with high accuracy. As a result, the generatrix tracking and classification processes can be executed with higher accuracy.

[0174] FIG. 36 is a diagram for specifically explaining the outline of the processing procedure in this modification. Referring to FIG. 36, first, each normal slice image 200 is generated by CT scanning. Then, for each normal slice image 200, the center point CPw of the cross-sectional area of the cable 9 in the image is determined by binarizing each pixel of the image. Next, each oblique slice image 350_k is generated. Thereafter, each high-resolution image 380_k is generated, and generatrix tracking is executed according to these high-resolution images.

[0175] FIG. 37 is a flowchart showing an example of the processing executed by the control device 17 in this modification. Referring to FIG. 37, S315 to 325 are the same as those shown in FIG. 34. This flowchart is different from the flowchart (FIG. 34) of the second embodiment in that S302 is added and S305A to S313A are executed in place of S305 to S313. In other respects, the flowchart of FIG. 37 is the same as that of FIG. 34. Each step of the flowchart of FIG. 37 is performed by the control device 17 (CPU 17a) executing a program stored in the memory 17b. At the start of this flowchart, the value of k is 1.

[0176] The control device 17 generates a high-resolution image 380_k by performing super-resolution processing on the diagonal slice image 300_k (S302).

[0177] The control device 17 determines, for each generatrix 25, a k-th center coordinate that is the center coordinate of the generatrix in the high-resolution image 380_k (S305A).

[0178] The control device 17 converts, for each generatrix 25, the k-th center coordinate of the generatrix into a k-th corresponding center coordinate that is the center coordinate of the generatrix in the reference coordinate system according to the angle θ(k) (S310A).

[0179] The control device 17 determines whether k has reached M (S312A). If k is less than M (NO in S312A), the control device 11 increments the value of k (S313A). Thereafter, the process returns to S302. As a result, until k reaches M (YES in S312A), S302 to S310A are repeated. When k reaches M, the process proceeds to S315.

[0180] As described above, according to this modification, the center (k-th center coordinate) of each generatrix 25 is determined with higher accuracy by image processing such as super-resolution processing. As a result, the generatrix tracking and classification processing can be performed with higher accuracy.

[0181] <Embodiment 3> The analysis method using the oblique slice image group is effective for objects different from the cable 9 as long as the object is curved in the longitudinal direction. In Embodiment 3, an example where the object is a bus bar will be described.

[0182] FIG. 38 is a diagram for explaining the shape of the bus bar. Referring to FIG. 38, the bus bar 90 is fixed by the jig 27. The bus bar 90 is different from the cable 9 that can be curved in its longitudinal direction depending on the posture at the point where it is originally curved in its longitudinal direction, but is the same as the cable 9 at the point where it is curved in the longitudinal direction. Each of ZA to ZC is an example of the Z coordinate in the reference coordinate system. In the following description, a normal slice image group or an oblique slice image group is used to analyze the internal structure of the bus bar 90.

[0183] FIG. 39 is a diagram for explaining an example of a normal slice image generated for the bus bar 90. This example is described as a comparative example.

[0184] Referring to FIG. 39, (A) to (C) of FIG. 39 show normal slice images 500A to 500C, respectively. The normal slice images 500A to 500C are generated corresponding to the XY planes whose Z coordinates are ZA to ZC, respectively. Since the bus bar 90 is curved in its longitudinal direction, the normal slice images 500 (for example, 500A to 500C) hardly represent the cross-sectional area of the bus bar 90 in a plane perpendicular to the longitudinal direction. As a result, the shape of the cross-sectional area of the bus bar 90 in the normal slice image 500 changes significantly depending on the Z coordinate. In this case, it may not be possible to accurately analyze the internal structure of the bus bar 90 using the normal slice image 500.

[0185] FIG. 40 is a diagram for explaining an example of an oblique slice image generated for the bus bar 90 in Embodiment 3.

[0186] Referring to FIGS. 40(A) to (C), the inventor confirmed that in a plurality of oblique slice images 600 (for example, oblique slice images 600_a to 600_c (1 ≦ a < b < c ≦ M)), the cross-section of the jig 27 moves left and right, and the way (shape) the cross-section of the bus bar 90 is reflected is substantially constant and does not change regardless of the z coordinate (k) (1 ≦ k ≦ M: in this example, k = a, b, or c). Therefore, using the oblique slice image group, the internal structure of objects other than the cable 9 such as the bus bar 90 can be accurately analyzed.

[0187] <Other Modification Examples> In the above, the functions of the super-resolution processing unit 165, the center coordinate determination unit 170 (170A), the coordinate conversion unit 175 (175A), the classification unit 180 (180A), the trajectory image generation unit 185 (185A), and the display control unit 190 (190A) are assumed to be performed by the control device 17 of the information processing device 16. However, these functions may be performed by the control device 11 of the X-ray device 1.

[0188] It should be considered that the embodiments disclosed this time are illustrative in all respects and not restrictive. The scope of the present invention is shown not by the above description but by the claims, and it is intended that all modifications within the meaning and scope equivalent to the claims are included.

Explanation of Reference Numerals

[0189] 1 X-ray device 3 Light source 3a X-rays 4 Stage 4a Support surface 4r Rotation axis 6 Driving device 7 X-ray detector 8 Detection element 9, 9A Cable 10 Computer 11, 17 Control device 11b, 17b Memory 12 Input interface 13, 19A Storage device 14, 19B Display device 15, 18 Communication device 16 Information processing device 21 Coating part 22 Strand bundle 25 Strand 27 Fixture 50 Volume data 90 Bus bar 110 First generation unit 115 Judgment unit 120 Second generation unit 122, 190, 190A Display control unit 125 Coordinate system setting unit 130, 175, 175A Coordinate conversion unit 135 Allocation unit 165 Super-resolution processing unit 170, 170A Center coordinate judgment unit 180, 180A Classification unit 185, 185A Trajectory image generation unit 200, 200a, 200b, 200c, 210, 500, 500A, 500B, 500C Normal slice image 300, 350, 600 Oblique slice image 305, 355 Plane 380 High-resolution image 505, 510, 515 Point set 700, 750 Data 810 Virtual trajectory image 820 Trajectory image CP, CP1, CP2, CP3, CPi, CPj, CPw, cp Center point L Length P, P1, P2, P3, Ph, Pk, cn, p, pna, pne Point R(X, Y, Z) Neighborhood area R(θk), R(θk) Matrix RNG Region SG Line segment Vij(k), eX, eY, eZ, ex, ey, ez Vector cg(s), cg′(s) Coordinate group d(w), ds Distance ep Outer product pth path px, pxe pixels pxf corresponding pixel

Claims

1. a light source that irradiates an object with X-rays, a support member having a support surface for supporting the object, a first generation unit that generates first to Nth images (N is a natural number of 2 or more), each representing a cross-sectional area of the object in a plane orthogonal to the normal vector of the support surface, according to the irradiation result of the X-rays on the object, a determination unit that determines first to Nth center points respectively indicating the centers of the cross-sectional areas of the first to Nth images, a second generation unit that generates first to Mth cross-sectional images (M is a natural number of 2 or more) according to the first to Nth images, each of the first to Mth cross-sectional images includes first to Mth points, the kth cross-sectional image (1 ≤ k ≤ M) is an image representing a cross-sectional area of the object in the kth plane orthogonal to the kth vector directed from the ith center point to the jth center point (1 ≤ i < j ≤ N) within the vicinity of the kth point, the first point is included in the first image, and the hth point (2 ≤ h ≤ M) is separated from the (h - 1)th point by a predetermined distance in the direction of the (h - 1)th vector. An X-ray CT apparatus.

2. each of the first to Nth images includes a plurality of pixels to which feature amounts are respectively assigned, the kth cross-sectional image includes n × m pixels to which the feature amounts are respectively assigned (n, m are natural numbers), the first generation unit generates each of the first to Nth images such that the coordinates of each of the plurality of pixels of the image are defined in a reference coordinate system in which the direction of the normal vector is the Z-axis direction, the second generation unit, a coordinate system setting unit that has the direction of the kth vector as the z-axis direction and sets a kth coordinate system for defining the coordinates of each of the n × m pixels according to the first to Nth center points, a coordinate conversion unit that converts the coordinates of each of the n × m pixels in the kth coordinate system into kth corresponding coordinates, which are the coordinates of the pixel in the reference coordinate system according to the kth angle, which is the angle formed by the normal vector and the kth vector, and an assignment unit that assigns the feature amount of each of the n × m pixels according to the feature amount of at least one pixel included in the plurality of pixels of the first to Nth images having coordinates within the vicinity of the kth corresponding coordinates of the pixel. The X-ray CT apparatus according to Claim 1.

3. The first generation unit generates the first to Nth images such that the Z coordinate of each of the plurality of pixels of the fth image (2 ≤ f ≤ N) is separated from the Z coordinate of each of the plurality of pixels of the (f - 1)th image by a certain distance in the Z-axis direction. The X-ray CT apparatus according to claim 2, wherein the second generation unit generates the first to Mth cross-sectional images such that the predetermined distance in the kth coordinate system is equal to the certain distance.

4. A method for generating an image of an object using an X-ray CT apparatus including a light source that irradiates the object with X-rays and a support member having a support surface for supporting the object, the method comprising: generating first to Nth images (N is a natural number of 2 or more), each representing a cross-sectional area of the object in a plane orthogonal to the normal vector of the support surface, according to an irradiation result of the X-rays on the object; determining first to Nth center points respectively indicating the centers of the cross-sectional areas of the first to Nth images; generating first to Mth cross-sectional images (M is a natural number of 2 or more) according to the first to Nth images, wherein the first to Mth cross-sectional images each include first to Mth points, the kth cross-sectional image (1 ≤ k ≤ M) is an image representing a cross-sectional area of the object in the kth plane orthogonal to the kth vector directed from the ith center point to the jth center point (1 ≤ i < j ≤ N) within the neighborhood of the kth point; the first point is included in the first image, and the hth point (2 ≤ h ≤ M) is separated from the (h - 1)th point by a predetermined distance in the direction of the (h - 1)th vector.

5. Each of the first to Nth images includes a plurality of pixels to which feature amounts are respectively assigned. The kth cross-sectional image includes n × m pixels to which the feature amounts are respectively assigned (n and m are natural numbers). Each of the first to Nth images is generated such that the coordinates of each of the plurality of pixels of the image are defined in a reference coordinate system having the direction of the normal vector as the Z-axis direction. The step of generating the first to Mth cross-sectional images includes: setting a kth coordinate system having the direction of the kth vector as the z-axis direction and for defining the coordinates of each of the n × m pixels according to the first to Nth center points. For each of the n×m pixels, converting the coordinates of the pixel in the k-th coordinate system into k-th corresponding coordinates which are the coordinates of the pixel in the reference coordinate system according to the k-th angle which is the angle formed by the normal vector and the k-th vector; The method according to claim 4, further comprising: for each of the n×m pixels, assigning the feature amount of the pixel according to the feature amount of at least one pixel included in the plurality of pixels of the first to N-th images, the at least one pixel having coordinates within a neighborhood region of the k-th corresponding coordinates of the pixel.

6. The first to N-th images are generated such that the Z coordinate of each of the plurality of pixels of the f-th image (2≤f≤N) is separated by a certain distance in the Z-axis direction from the Z coordinate of each of the plurality of pixels of the (f−1)-th image; The method according to claim 5, wherein the first to M cross-sectional images are generated such that the predetermined distance in the k-th coordinate system is equal to the certain distance.

7. The object is a cable including a plurality of strands; The method according to claim 5, further comprising: for each of the plurality of strands, determining a k-th center coordinate which is the center coordinate of the strand in the k-th cross-sectional image; The method according to claim 5, further comprising: for each of the plurality of strands, converting the k-th center coordinate of the strand into a k-th corresponding center coordinate which is the center coordinate of the strand in the reference coordinate system according to the k-th angle.

8. The object is a cable including a plurality of strands; The method according to claim 5, further comprising: generating a k-th high-resolution image by performing image processing for increasing the resolution of the k-th cross-sectional image; The method according to claim 5, further comprising: for each of the plurality of strands, determining a k-th center coordinate which is the center coordinate of the strand in the k-th high-resolution image; The method according to claim 5, further comprising: for each of the plurality of strands, converting the k-th center coordinate of the strand into a k-th corresponding center coordinate which is the center coordinate of the strand in the reference coordinate system according to the k-th angle.

9. The method according to claim 7 or claim 8, further comprising: for each of the plurality of strands, performing a classification process of classifying the strand into any one of a plurality of groups according to the locus of the k-th center coordinate of the strand.

10. The method according to claim 7 or claim 8, further comprising, for each of the plurality of strands, generating a locus image representing a locus of the k-th corresponding center coordinates of the strand in the reference coordinate system.

11. A program for causing a computer to execute the method according to claim 5.

Images