X-ray imaging system and X-ray image analysis method

Abstract

To provide an X-ray imaging system and an X-ray image analysis method with which, when inspecting an inspection object that includes an object portion by irradiating the inspection object with an X-ray, it is possible to easily conduct inspection, irrespective of the three-dimensional position in a depth direction of the object portion.SOLUTION: This X-ray imaging system 100 comprises an image generation unit 15 that generates an X-ray image 30a and an X-ray image 30b that are mutually different in the angle of irradiation. The X-ray imaging system 100 further includes a control unit 21 that calculates, on the basis of the positions of a solder ball 104 and a solder ball 106 included in an inspection object 101 in each of the X-ray image 30a and the X-ray image 30b, and the difference in angle between the X-ray irradiation angle for generating the X-ray image 30a and the X-ray irradiation angle for generating the X-ray image 30b, the three-dimensional positions of the solder ball 104 and solder ball 106 by a trigonometrical survey.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to an X-ray imaging system and an X-ray image analysis method, and more particularly to an X-ray imaging system and an X-ray image analysis method for irradiating an object to be examined with X-rays. [Background technology]

[0002] Conventionally, devices that irradiate objects to be inspected with X-rays are known (see, for example, Patent Document 1).

[0003] Patent Document 1 discloses a device (fluoroscopic inspection device) that emits an X-ray beam toward an electronic substrate on which a semiconductor element, which is the object to be inspected, is mounted. This fluoroscopic inspection device acquires a transmitted image of the semiconductor element by detecting the emitted X-ray beam. Based on the transmitted image of the semiconductor element, the fluoroscopic inspection device described in Patent Document 1 determines soldering defects by measuring the presence or absence of air bubbles in the solder of the semiconductor element. Furthermore, the fluoroscopic inspection device described in Patent Document 1 removes noise images caused by foreign matter overlapping the semiconductor element by combining two images acquired by detecting X-ray beams incident at different angles. [Prior art documents] [Patent Documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2011-196983 [Overview of the project] [Problems that the invention aims to solve]

[0005] However, when inspecting an object by irradiating it with X-rays, as in the fluoroscopic inspection device described in Patent Document 1, the image generated for inspection is a two-dimensional projection image of the irradiated X-rays. Therefore, it is difficult to detect the three-dimensional depth position of the target part included in the object being inspected. For example, if electronic components with the same shape are mounted on both the front and back surfaces of a substrate, it becomes difficult to distinguish whether the part determined to be abnormal is located on the front or back surface of the substrate. As a result, it is conceivable that inspection becomes difficult due to the difficulty in detecting the three-dimensional position of the target part included in the object being inspected. Furthermore, when noise images caused by foreign objects overlapping the object being inspected are removed by combining two images, as in the fluoroscopic inspection device described in Patent Document 1, it is conceivable that the target part included in the object being inspected may be removed. For example, if electronic components that are the target part to be inspected are mounted so that they overlap on both the front and back surfaces of a substrate, the target part on either the front or back surface of the substrate will be removed as noise. As a result, it is conceivable that inspection of the removed target part will become impossible. Therefore, when inspecting an object by irradiating it with X-rays, including the target area, it is desirable to be able to easily perform the inspection regardless of the three-dimensional depth position of the target area.

[0006] This invention was made to solve the above-mentioned problems, and one of its objectives is to provide an X-ray imaging system and an X-ray image analysis method that enable easy inspection of an object to be inspected, including the target part, by irradiating the object with X-rays, regardless of the three-dimensional depth position of the target part. [Means for solving the problem]

[0007] To achieve the above objective, the X-ray imaging system in the first aspect of this invention is: Placed on the circuit boardThe system comprises: an X-ray irradiation unit that irradiates an object to be inspected, including the target portion, with X-rays; an X-ray detection unit that detects the X-rays irradiated from the X-ray irradiation unit; an angle change unit that changes the irradiation angle of the X-rays irradiated onto the object to be inspected; an image generation unit that generates a first image and a second image, which are X-ray images with different irradiation angles, based on the X-rays irradiated by the X-ray irradiation unit and detected by the X-ray detection unit while changing the irradiation angle by the angle change unit; and an image analysis unit that calculates the three-dimensional position of the target portion by triangulation based on the position of the target portion included in each of the first and second images generated by the image generation unit, and the angle difference between the X-ray irradiation angle for generating the first image and the X-ray irradiation angle for generating the second image. The image analysis unit determines whether the target portion is located on the front or back surface of the substrate. . The X-ray imaging system in the second aspect of this invention includes an X-ray irradiation unit that irradiates an object to be inspected, including the target portion, with X-rays; an X-ray detection unit that detects the X-rays irradiated from the X-ray irradiation unit; an angle changing unit that changes the irradiation angle of the X-rays irradiated onto the object to be inspected; an image generation unit that generates a first image and a second image, which are X-ray images with different irradiation angles, based on the X-rays irradiated by the X-ray irradiation unit and detected by the X-ray detection unit while changing the irradiation angle by the angle changing unit; and the position of the target portion included in the object to be inspected in each of the first and second images generated by the image generation unit, and the generation of the first image The system includes an image analysis unit that calculates the three-dimensional position of a target portion by triangulation based on the angle difference between the X-ray irradiation angle for generating the first image and the X-ray irradiation angle for generating the second image, wherein the target portion includes a first target portion and a second target portion that are different from each other, and the image analysis unit calculates the three-dimensional position of the first target portion and the three-dimensional position of the second target portion by triangulation based on the respective positions of the first target portion and the second target portion in each of the first and second images generated by the image generation unit, and the angle difference between the X-ray irradiation angle for generating the first image and the X-ray irradiation angle for generating the second image. The X-ray imaging system in the third aspect of this invention comprises: an X-ray irradiation unit that irradiates an object to be inspected, including the target portion, with X-rays; an X-ray detection unit that detects the X-rays irradiated from the X-ray irradiation unit; an angle changing unit that changes the irradiation angle of the X-rays irradiated onto the object to be inspected; an image generation unit that generates a first image and a second image, which are X-ray images with different irradiation angles, based on the X-rays irradiated by the X-ray irradiation unit and detected by the X-ray detection unit while changing the irradiation angle by the angle changing unit; and an image analysis unit that calculates the three-dimensional position of the target portion by triangulation based on the position of the target portion included in each of the first and second images generated by the image generation unit and the angle difference between the X-ray irradiation angle for generating the first image and the X-ray irradiation angle for generating the second image. The image analysis unit detects the region of the target portion in each of the first and second images by using a trained model generated by machine learning, and obtains the position of the target portion in each of the first and second images based on the detected region of the target portion.

[0008] This invention 4 The X-ray imaging system in this context comprises: an X-ray irradiation unit that irradiates an object to be examined, which includes a first target portion and a second target portion, which are different target portions from each other, with X-rays; an X-ray detection unit that detects the X-rays irradiated from the X-ray irradiation unit; an image generation unit that generates an X-ray image based on the X-rays irradiated by the X-ray irradiation unit and detected by the X-ray detection unit; and an image analysis unit that detects the regions of the first target portion and the second target portion in the X-ray image in a separated state, by using a trained model that has been trained to separately identify overlapping portions and non-overlapping portions in the X-ray image generated by the image generation unit, where the first target portion and the second target portion overlap each other when viewed from the irradiation direction, which is the direction from the X-ray irradiation unit to the X-ray detection unit.

[0009] This invention 5 The X-ray image analysis method in this phase is: Placed on the circuit boardThe steps include: acquiring a first image and a second image, which are X-ray images with different irradiation angles, generated based on X-rays detected by irradiating an object containing the target part while changing the irradiation angle; calculating the three-dimensional position of the target part by triangulation based on the position of the target part included in the object in each of the first and second images, and the angle difference between the X-ray irradiation angle for generating the first image and the X-ray irradiation angle for generating the second image; A step of determining whether the target part is located on the front or back surface of the substrate, It is equipped with. [Effects of the Invention]

[0010] The above 1 ~3 X-ray imaging system in the context of the above, and the above-mentioned 5 In this X-ray image analysis method, the three-dimensional position of the target part is calculated by triangulation based on the position of the target part included in the object being inspected in each of the first and second images, and the angular difference between the X-ray irradiation angle for generating the first image and the X-ray irradiation angle for generating the second image. This makes it possible to obtain the three-dimensional position of the target part calculated by triangulation based on the position of the target part in the first image, the position of the target part in the second image, and the angular difference in the X-ray irradiation angle between the first and second images. Therefore, the object being inspected can be inspected while identifying the three-dimensional position of the target part. As a result, when inspecting an object by irradiating it with X-rays, including the target part, the inspection can be easily performed regardless of the three-dimensional depth position of the target part.

[0011] Furthermore, the above 4In this aspect, the X-ray imaging system uses a trained model that has been learned to separately identify overlapping portions where a first target portion and a second target portion in an X-ray image generated by an image generation unit overlap each other when viewed from an irradiation direction that is a direction from an X-ray irradiation unit toward an X-ray detection unit, and non-overlapping portions that do not overlap each other, thereby detecting, in a separated state, the region of the first target portion and the region of the second target portion in the X-ray image based on the separately identified overlapping portions and non-overlapping portions. Accordingly, even when the region of the first target portion and the region of the second target portion have mutually overlapping portions, use of the trained model enables accurate detection of the region of the first target portion and the region of the second target portion in a separated state. Therefore, inspection of an inspection object can be performed with the region of the first target portion and the region of the second target portion separated from each other. As a result, when inspecting an inspection object by irradiating the inspection object including target portions with X-rays, the inspection can be easily performed regardless of the three-dimensional depth-direction positions of the target portions. [BRIEF DESCRIPTION OF THE DRAWINGS]

[0012] [Figure 1] It is a block diagram showing the overall configuration of the X-ray imaging system according to the first embodiment of the present invention. [Figure 2] It is a schematic diagram showing the configuration of an inspection object including solder balls. [Figure 3] It is a diagram showing an example of an X-ray image 30a. [Figure 4] It is a diagram showing an example of an X-ray image 30b having an irradiation angle different from that of the X-ray image 30a. [Figure 5] It is a block diagram for explaining the functional configuration of a control unit. [Figure 6] It is a diagram for explaining acquisition of a discrimination result image using a trained model. [Figure 7] It is a diagram for explaining separation of a front solder ball region and a back solder ball region. [Figure 8] It is a diagram for explaining calculation of a three-dimensional position of a solder ball by triangulation. [Figure 9] It is a diagram showing calculation results of three-dimensional positions calculated by triangulation. [Figure 10] It is a flowchart for explaining the X-ray image analysis method according to the first embodiment. [Figure 11] It is a block diagram showing the overall configuration of an X-ray imaging system according to a second embodiment of the present invention. [Figure 12] It is a diagram for explaining acquisition of a discrimination result image using a trained model according to the second embodiment. MODE FOR CARRYING OUT THE INVENTION

[0013] Hereinafter, embodiments embodying the present invention will be described based on the drawings.

[0014] [First Embodiment] (Overall Configuration of X-ray Imaging System) An X-ray imaging system 100 according to a first embodiment of the present invention will be described with reference to FIGS. 1 to 9.

[0015] As shown in FIG. 1, the X-ray imaging system 100 according to the first embodiment is a system that images the inside of an inspection object 101 by detecting X-rays that have passed through the inspection object 101. The X-ray imaging system 100 is used, for example, for imaging the inside of the inspection object 101 as an object in non-destructive inspection applications.

[0016] As shown in Figure 2, the object to be inspected 101 is an electronic device including a substrate 102. The substrate 102 has a surface 102a on the Z1 direction side (X-ray detection unit 12 side) and a back surface 102b on the Z2 direction side (X-ray irradiation unit 11 side). An electronic component 103 is mounted on the surface 102a of the substrate 102. An electronic component 105 is mounted on the back surface 102b of the substrate 102. The electronic component 103 is electrically connected to the surface 102a, which is the Z1 direction side of the substrate 102, by a plurality of solder balls 104 (bumps). The plurality of solder balls 104 are arranged in a grid pattern on the surface 102a of the substrate 102 in a regular manner. The electronic component 105 is electrically connected to the back surface 102b, which is the Z2 direction side of the substrate 102, by a plurality of solder balls 106 (bumps). The multiple solder balls 106, like the multiple solder balls 104, are arranged in a grid pattern on the back surface 102b of the substrate 102 in a regular manner. That is, the electronic components 103 and 105 are connected to the substrate 102 by a BGA (Ball Grid Array). Note that the solder ball 104 is an example of the "target portion," "first target portion," and "first solder ball" in the claims. Also, the solder ball 106 is an example of the "target portion," "second target portion," and "second solder ball" in the claims.

[0017] For example, in the object under inspection 101, 39 solder balls 104 are arranged in a grid pattern in 3 vertical and 13 horizontal rows on the surface 102a of the substrate 102, and similarly, 39 solder balls 106 are arranged in a grid pattern in 3 vertical and 13 horizontal rows on the back surface 102b. Each of the multiple solder balls 104 on the surface 102a of the substrate 102 and each of the multiple solder balls 106 on the back surface 102b are arranged to overlap each other when viewed from a direction perpendicular to the surface 102a of the substrate 102 (Z direction). Electronic components 103 and 105 include, for example, electronic circuits such as ICs (integrated circuits). The X-ray imaging system 100 performs non-destructive testing for abnormalities such as voids and bridges in the multiple solder balls 104 and multiple solder balls 106. In addition to the electronic components 103, the substrate 102 also has surface-mounted resistors or capacitors (not shown) mounted on it.

[0018] As shown in Figure 1, the X-ray imaging system 100 comprises a fluoroscopy device 10 and an analysis device 20. The fluoroscopy device 10 generates X-ray images 30a (see Figure 3) and 30b (see Figure 4) by performing X-ray imaging on the object to be inspected 101. The analysis device 20 performs analysis processing on the generated X-ray images 30a and 30b. Each of the fluoroscopy device 10 and the analysis device 20 has a communication module and transmits and receives information from each other via a network or the like. Note that X-ray image 30a is an example of the "first image" and "X-ray image" in the claims. Also, X-ray image 30b is an example of the "second image" and "X-ray image" in the claims.

[0019] The fluoroscopy apparatus 10 comprises an X-ray irradiation unit 11, an X-ray detection unit 12, a stage 13, a drive unit 14, and an image generation unit 15. The X-ray irradiation unit 11 irradiates an object to be inspected 101, which includes a plurality of solder balls 104 and a plurality of solder balls 106, with X-rays. The X-ray irradiation unit 11 includes an X-ray tube, which is an X-ray source that irradiates X-rays when power is supplied from a power supply unit (not shown). The X-ray detection unit 12 detects the X-rays irradiated from the X-ray irradiation unit 11. The X-ray detection unit 12 outputs an electrical signal corresponding to the detected X-rays. The X-ray detection unit 12 includes, for example, an FPD (Flat Panel Detector), which is an X-ray detector. The X-ray irradiation unit 11 and the X-ray detection unit 12 are located inside a housing (not shown) of the fluoroscopy apparatus 10. The drive unit 14 is an example of an "angle changing unit" in the claims.

[0020] The object to be inspected 101 is placed on the stage 13. The stage 13 is configured to rotate around the Y-axis by the drive unit 14. The drive unit 14 changes the positioning angle of the object to be inspected 101 by rotating the stage 13, on which the object to be inspected 101 is placed, around the Y-axis. In other words, the drive unit 14 changes the irradiation angle of the X-rays irradiated onto the object to be inspected 101 by changing the positioning angle of the object to be inspected 101 relative to the X-ray irradiation unit 11. The irradiation angle is expressed, for example, as the angle of the X-ray irradiation direction with respect to a direction perpendicular to the mounting surface of the stage 13 on which the object to be inspected 101 is placed. The drive unit 14 operates based on signals from the image generation unit 15. The drive unit 14 includes, for example, a servo motor. Note that "X-ray irradiation direction" means the direction from the X-ray irradiation unit 11 toward the X-ray detection unit 12. More specifically, the X-ray irradiation direction is described as the three-dimensional depth direction from the X-ray source of the X-ray irradiation unit 11, which is a point light source, toward the center of the detection surface of the X-ray detection unit 12.

[0021] As shown in Figures 3 and 4, the image generation unit 15 generates X-ray images 30a and 30b based on the X-rays detected by the X-ray detection unit 12. X-ray images 30a and 30b are X-ray images with different irradiation angles, generated based on X-rays detected by the X-ray detection unit 12 after being irradiated by the X-ray irradiation unit 11 while the irradiation angle is changed by the drive unit 14. For example, X-ray image 30a is an image acquired when the irradiation direction (Z direction), which is the direction from the X-ray irradiation unit 11 toward the X-ray detection unit 12, is approximately equal to the direction perpendicular to the main surface (surface 102a) of the substrate 102. X-ray image 30b is an image acquired when the arrangement angle of the object to be inspected 101 is rotated 15 degrees around the Y axis from the Z direction.

[0022] X-ray images 30a and 30b include multiple solder balls 104 and multiple solder balls 106 arranged in a grid pattern with regularity. In X-ray images 30a and 30b, the multiple solder balls 104 and multiple solder balls 106 are arranged on the substrate 102 of the object under inspection 101 so as to overlap each other, and therefore the images include areas where the solder balls 104 and solder balls 106 overlap each other. In addition, in X-ray images 30a and 30b, the irradiation angles are different, so the positional relationship between the solder balls 104 and solder balls 106 is different.

[0023] The image generation unit 15 is a computer that includes, for example, a processor such as a CPU (Central Processing Unit) and a memory device for storing information. The image generation unit 15 controls the operation of each part of the fluoroscopy device 10. The image generation unit 15 controls the irradiation of X-rays by the X-ray irradiation unit 11 by, for example, controlling a power supply device (not shown). The image generation unit 15 also controls the operation of the drive unit 14. The image generation unit 15 then outputs the generated X-ray images 30a and 30b to the analysis device 20.

[0024] As shown in Figure 1, the analysis device 20 has a control unit 21 and a storage unit 22. The analysis device 20 is, for example, a personal computer that is communicatively connected to the fluoroscopy device 10. The control unit 21 includes a CPU, ROM (Read Only Memory), and RAM (Random Access Memory), etc. The control unit 21 may also include a processor such as a GPU (Graphics Processing Unit) or an FPGA (Field-Programmable Gate Array) configured for image processing. Note that the control unit 21 is an example of the "image analysis unit" in the claims.

[0025] The storage unit 22 is configured to store various programs and parameters executed by the control unit 21. The storage unit 22 includes non-volatile memory such as an HDD (Hard Disk Drive) or an SSD (Solid State Drive). The storage unit 22 also stores a trained model 50 (see Figure 6) used for analyzing X-ray images 30a and 30b.

[0026] Furthermore, a display unit 23 and an operation unit 24 are connected to the analysis device 20. The display unit 23 includes, for example, a liquid crystal monitor. The display unit 23 displays images and text information under the control of the control unit 21. The operation unit 24 receives input operations from the operator. The operation unit 24 includes, for example, a keyboard and a pointing device such as a mouse. The operation unit 24 outputs operation signals to the control unit 21 based on the received input operations.

[0027] (Analysis of X-ray images using an analysis device) The analysis device 20 determines whether there are any abnormalities in the solder balls 104 and 106 based on the X-ray images 30a and 30b generated by the image generation unit 15. Specifically, the analysis device 20 detects the regions of the solder balls 104 and 106 from two X-ray images 30a and 30b, which have different irradiation angles. The analysis device 20 also calculates the three-dimensional positions of the solder balls 104 and 106, including their positions in the irradiation direction (Z direction) (positions perpendicular to the substrate 102), by triangulation. In other words, the analysis device 20 calculates the three-dimensional positions (coordinates) of the solder balls 104 and 106, including their positions in the three-dimensional depth direction, in the X-ray images 30a and 30b, by triangulation. The analysis device 20 then determines whether there are any abnormalities in the detected solder balls 104 and 106 based on the regions of the solder balls 104 and 106, and also determines whether the abnormal part is located on the surface 102a or the back surface 102b of the substrate 102 (i.e., whether it is a solder ball 104 or a solder ball 106) based on the three-dimensional position of the abnormal part.

[0028] Specifically, as shown in Figure 5, the control unit 21 includes a region detection unit 41, a coordinate acquisition unit 42, an angle difference acquisition unit 43, a position calculation unit 44, and a determination unit 45. The control unit 21 functions as the region detection unit 41, the coordinate acquisition unit 42, the angle difference acquisition unit 43, the position calculation unit 44, and the determination unit 45 by executing various programs (software) stored in the storage unit 22. That is, in Figure 5, the region detection unit 41, the coordinate acquisition unit 42, the angle difference acquisition unit 43, the position calculation unit 44, and the determination unit 45 are illustrated as software-based functional blocks. However, it is not limited to this, and some or all of the region detection unit 41, the coordinate acquisition unit 42, the angle difference acquisition unit 43, the position calculation unit 44, and the determination unit 45 may be configured by dedicated hardware circuits.

[0029] As shown in Figure 6, the region detection unit 41 uses the trained model 50 to acquire identification result images 31a and 31b from the X-ray images 30a and 30b.

[0030] As shown in Figure 7, the identification result image 31a is a label image that identifies three types of regions in the X-ray image 30a: an overlapping region 61 where solder balls 104 and solder balls 106 on both sides of the substrate 102 overlap when viewed from the irradiation direction (Z direction); a non-overlapping region 62 which contains only one solder ball 104 or solder ball 106 without overlapping; and a background region 63 which does not contain either solder ball 104 or solder ball 106. In the identification result image 31a, the overlapping region 61 is colored gray, the non-overlapping region 62 is colored white, and the background region 63 is colored black. Note that in Figures 6 and 7, gray is represented by hatching. The color coding in the identification result image 31b is the same as in the identification result image 31a.

[0031] The trained model 50 is generated by machine learning to distinguish the overlapping portion 61, the non-overlapping portion 62, and the background portion 63 separately. The trained model 50 is generated by machine learning using deep learning, with the input training data 30t, which mimics X-ray images 30a and 30b, and the output training data 31t, in which the regions corresponding to the overlapping portion 61, non-overlapping portion 62, and background portion 63 in the input training data 30t are color-coded. For example, in the generation of the trained model 50, machine learning based on U-Net, a type of Fully Convolutional Network (FCN), is performed. The trained model 50 is generated by training it to perform an image transformation (image reconstruction) on each pixel of the input image, coloring the overlapping portion 61 gray, the non-overlapping portion 62 white, and the background portion 63 black. The region detection unit 41 obtains identification result images 31a and 31b as identification results (inference results) from the trained model 50, in which the overlapping portion 61, the non-overlapping portion 62, and the background portion 63 are identified separately.

[0032] Then, as shown in Figure 7, the region detection unit 41 detects the regions of solder balls 104 and solder balls 106 in X-ray image 30a and X-ray image 30b, respectively. The region detection unit 41 detects the regions of multiple solder balls 104 and multiple solder balls 106 in X-ray image 30a and X-ray image 30b, respectively, by using a trained model 50 generated by machine learning. Specifically, the region detection unit 41 detects the regions of multiple solder balls 104 and multiple solder balls 106 in X-ray image 30a and X-ray image 30b, respectively, in a separated state, based on the overlapping portion 61 and non-overlapping portion 62 that are separately identified in the identification result image 31a and identification result image 31b. Although Figure 7 shows an example of an identification result image 31a including one solder ball 104 and one solder ball 106, the same process is performed individually for each of the multiple solder balls 104 and each of the multiple solder balls 106.

[0033] In detail, the region detection unit 41 acquires each of the non-overlapping portions 62 as a detection region that is the region of either the solder ball 104 or the solder ball 106. Then, the region detection unit 41 detects the solder ball 104 region and the solder ball 106 region separately by adding the overlapping portion 61 adjacent to the detection region to the detection region. That is, the region detection unit 41 acquires regions 64 and 65 in a state where the solder ball 104 region and the solder ball 106 region are separated. However, although regions 64 and 65 are in a state where the solder ball 104 region and the solder ball 106 region are separated, it is not known at this point which of regions 64 and 65 is the solder ball 104 region and which is the solder ball 106 region.

[0034] Therefore, in order to determine which of the detected regions 64 and 65 is the region of solder ball 104 and which is the region of solder ball 106, a triangulation calculation is performed.

[0035] <Calculation using triangulation> The coordinate acquisition unit 42 acquires the positions of each of the multiple solder balls 104 and each of the multiple solder balls 106 in each of the X-ray images 30a and 30b, based on regions 64 and 65, which are either the region of the multiple solder balls 104 or the region of the multiple solder balls 106 detected by the unit. Specifically, the coordinate acquisition unit 42 acquires the centroid positions of regions 64 and 65, which are the region of the multiple solder balls 104 and the region of the multiple solder balls 106 detected by the unit, as the positions of each of the solder balls 104 and solder balls 106 in each of the X-ray images 30a and 30b. For example, the coordinate acquisition unit 42 performs a process to extract the contour of each of the regions 64 and 65 detected from each of the X-ray images 30a and 30b. Then, based on the extracted contour lines, the coordinate acquisition unit 42 acquires the centroid position of the region enclosed by the contour lines. The coordinate acquisition unit 42 then acquires the coordinates of the acquired centroid position in the X-ray images 30a and 30b, respectively. Here, "centroid position" refers to the position of the mathematical geometric center of the region enclosed by the extracted contour lines. In other words, the centroid position is the average position of the position coordinates of all points included in the region enclosed by the extracted contour lines.

[0036] The angle difference acquisition unit 43 acquires the angle difference between the X-ray irradiation angle for generating X-ray image 30a and the X-ray irradiation angle for generating X-ray image 30b. The fluoroscopy device 10 acquires the irradiation angles when generating X-ray image 30a and X-ray image 30b, respectively. The fluoroscopy device 10 then outputs the acquired irradiation angles, each associated with X-ray image 30a and X-ray image 30b, to the analysis device 20. The angle difference acquisition unit 43 acquires the angle difference between the irradiation angle of X-ray image 30a and the irradiation angle of X-ray image 30b, based on the irradiation angles associated with each of the X-ray image 30a and X-ray image 30b.

[0037] Then, as shown in FIG. 8, in the first embodiment, the position calculation unit 44 processes, in each of the X-ray image 30a and the X-ray image 30b generated by the image generation unit 15, each position of the plurality of solder balls 104 and the plurality of solder balls 106 included in the inspection object 101, and calculates the three-dimensional positions of each of the plurality of solder balls 104 and the three-dimensional positions of each of the plurality of solder balls 106 by triangulation, based on the angle difference between the X-ray irradiation angle for generating the X-ray image 30a and the X-ray irradiation angle for generating the X-ray image 30b. That is, the position calculation unit 44 does not perform reconstruction processing to generate a three-dimensional model from the X-ray image 30a and the X-ray image 30b like CT (Computed Tomography), and calculates three-dimensional positions including the positions of the solder balls 104 and 106 in the irradiation direction (Z direction) by triangulation.

[0038] Specifically, for example, the position calculation unit 44 sets the centroid position of the region 64 in the X-ray image 30a as point P0´(x0´, y0´, z0´), and sets the centroid position of the region 64 in the X-ray image 30b as point P θ ´(x θ ´,y θ ´,z θ ´). It is assumed that the irradiation angle of the X-ray image 30b is rotated by θ from the irradiation angle of the X-ray image 30a. At this time, let the three-dimensional position of the solder ball 104 or the solder ball 106 corresponding to the region 64 at the time when X-rays are irradiated to generate the X-ray image 30a be point P0(x0, y0, z0), and let the three-dimensional position of the solder ball 104 or the solder ball 106 corresponding to the region 64 at the time when X-rays are irradiated to generate the X-ray image 30b after the irradiation angle is changed by θ be point P θ (x θ ,y θ ,z θ ), and let the coordinates of the rotation center be C(t x ,0,t z ), then the coordinates of point P0 are expressed by the following equations (1) and (2) based on calculation by triangulation.

Mathematical Expression

[0039] As shown in Figure 9, when the results of calculating the position based on each of the regions 64 and each of the regions 65 are plotted, it is possible to separate the points into regions where the numerical value indicating the position in the Z direction is large (closer to the X-ray detection unit 12) and regions where the numerical value indicating the position in the Z direction is small (further from the X-ray detection unit 12). In this way, it is possible to detect whether region 64 or region 65 corresponds to the multiple solder balls 104 located on the Z1 side or the multiple solder balls 106 located on the Z2 side.

[0040] Note that both X-ray image 30a and X-ray image 30b contain multiple solder balls 104 and multiple solder balls 106, respectively. For example, by performing multiple X-ray imaging while slightly changing the irradiation angle from the X-ray imaging to acquire X-ray image 30a to the X-ray imaging to acquire X-ray image 30b, the correspondence between each of the multiple solder balls 104 and multiple solder balls 106 in X-ray image 30a and each of the multiple solder balls 104 and multiple solder balls 106 in X-ray image 30b can be obtained.

[0041] The determination unit 45 determines whether there are any abnormalities in the solder balls 104 and 106 contained in the X-ray images 30a and 30b, based on the regions identified in the identification result image 31a and the identification result image 31b. For example, from among regions 64 and 65 in the identification result image 31a, the determination unit 45 acquires the region detected as the region of the solder ball 104 based on the position in the irradiation direction obtained from the three-dimensional position calculated by the position calculation unit 44. The determination unit 45 then detects the area and shape of each of the multiple solder balls 104 by detecting each region of the multiple solder balls 104 in the X-ray image 30a. The determination unit 45 then determines whether there are any abnormalities such as voids, poor wetting, bridging, and sputter in each of the multiple solder balls 104, based on the area and shape of each of the detected multiple solder balls 104. The determination unit 45 may display on the display unit 23 the portion of the X-ray image 30a that has been determined to have an abnormality. The determination unit 45 similarly determines whether or not there is an abnormality in the solder ball 106.

[0042] The determination unit 45 may determine whether or not there is an abnormality based on whether or not the area of ​​the detected solder ball 104 or 106 falls within a predetermined range. In that case, the size of the predetermined range considered normal may be corrected based on the calculated position (distance) of the irradiation direction.

[0043] (X-ray image analysis method according to the first embodiment) Next, with reference to Figure 10, the X-ray image analysis method according to the first embodiment will be described. Note that the control processing in steps 301 to 307 is performed by the execution of a program by the control unit 21.

[0044] First, in step 301, X-ray images 30a and 30b are obtained, which are X-ray images with different irradiation angles. X-ray images 30a and 30b are generated based on X-rays irradiated while changing the irradiation angle of the X-rays to the object under inspection 101, which includes multiple solder balls 104 and multiple solder balls 106, by changing the arrangement angle of the object under inspection 101.

[0045] Next, in step 302, the identified image 31a and identified image 31b are obtained from the X-ray image 30a and X-ray image 30b using the trained model 50, thereby detecting (estimating) the separately identified superimposed portion 61 and non-superimposed portion 62.

[0046] Next, in step 303, based on the detected superimposed portion 61 and non-superimposed portion 62, regions 64 and 65 are obtained in which the regions of multiple solder balls 104 and the regions of multiple solder balls 106 in each of the X-ray image 30a and X-ray image 30b are separated.

[0047] Next, in step 304, the positions of each of the multiple solder balls 104 and multiple solder balls 106 contained in the object to be inspected 101 are obtained in each of the X-ray images 30a and 30b generated by the image generation unit 15. That is, based on the regions 64 and 65, which are the regions of the multiple solder balls 104 and the regions of the multiple solder balls 106 detected in step 303, the positions of each of the multiple solder balls 104 and multiple solder balls 106 in each of the X-ray images 30a and 30b are obtained. Specifically, the centroid positions of each of the detected regions 64 and 65 are obtained.

[0048] Next, in step 305, the angle difference between the X-ray irradiation angle for generating X-ray image 30a and the X-ray irradiation angle for generating X-ray image 30b is obtained.

[0049] Next, in step 306, based on the positions of the multiple solder balls 104 and each of the multiple solder balls 106 obtained in step 304 and the angle difference obtained in step 305, the three-dimensional position of each of the multiple solder balls 104, including its position in the three-dimensional depth direction which is the direction of illumination (direction perpendicular to the substrate 102: Z direction), is calculated by triangulation.

[0050] Next, in step 307, based on the calculated three-dimensional position, the regions of the multiple solder balls 104 and the regions of the multiple solder balls 106 are detected, and a determination is made as to whether or not there is an abnormality in each of the multiple solder balls 104 and each of the multiple solder balls 106.

[0051] (Effects of the first embodiment) In the first embodiment, the following effects can be obtained.

[0052] In the first embodiment, as described above, the X-ray imaging system 100 calculates the three-dimensional positions of the solder balls 104 and 106 by triangulation based on the positions of the solder balls 104 and 106 (target portion) contained in the object to be inspected 101 in each of the X-ray images 30a (first image) and 30b (second image), and the angular difference between the X-ray irradiation angle for generating the X-ray image 30a and the X-ray irradiation angle for generating the X-ray image 30b. As a result, the three-dimensional positions of the solder balls 104 and 106 calculated by triangulation can be obtained based on the positions of the solder balls 104 and 106 in the X-ray image 30a, the positions of the solder balls 104 and 106 in the X-ray image 30b, and the angular difference in the X-ray irradiation angles between the X-ray image 30a and the X-ray image 30b. Therefore, the inspection object 101 can be inspected while identifying the three-dimensional positions of the solder balls 104 and 106. As a result, when inspecting the inspection object 101, which includes the solder balls 104 and 106, by irradiating it with X-rays, the inspection can be easily performed regardless of the three-dimensional depth position of the solder balls 104 and 106.

[0053] Furthermore, in the first embodiment described above, additional effects can be obtained by configuring it as follows.

[0054] In other words, in the first embodiment, as described above, the drive unit 14 (angle changing unit) changes the irradiation angle by changing the positioning angle of the object to be inspected 101 relative to the X-ray irradiation unit 11, the position of the object to be inspected 101, the positioning angle of the X-ray irradiation unit 11 relative to the object to be inspected 101, and the position of the X-ray irradiation unit 11, which is at least one of these. The control unit 21 (image analysis unit) calculates the three-dimensional positions of the solder balls 104 and solder balls 106 (target parts) by triangulation without performing reconstruction processing on the X-ray image 30a (first image) and X-ray image 30b (second image). Here, if a three-dimensional model is generated and the three-dimensional positional relationship is calculated by performing image reconstruction processing, the time required for calculation processing increases compared to when the positional relationship is calculated by triangulation. In contrast, in the first embodiment, the control unit 21 is configured to calculate the three-dimensional positions of the solder balls 104 and 106 (target portion) by triangulation without performing reconstruction processing on the X-ray images 30a and 30b. This suppresses an increase in the time required to obtain the three-dimensional positions of the solder balls 104 and 106, and reduces the processing load on the control unit 21 (image analysis unit).

[0055] Furthermore, in the first embodiment, as described above, the object to be inspected 101 includes two different solder balls, 104 (first target portion) and 106 (second target portion). The control unit 21 (image analysis unit) calculates the three-dimensional position of solder ball 104 and solder ball 106 by triangulation based on the positions of solder ball 104 and solder ball 106 in each of the X-ray images 30a and 30b generated by the image generation unit 15, and the angle difference between the X-ray irradiation angle for generating X-ray image 30a and the X-ray irradiation angle for generating X-ray image 30b. With this configuration, the three-dimensional position of solder ball 104 and solder ball 106 can be obtained by calculation processing using triangulation in the control unit 21. Therefore, solder ball 104 and solder ball 106 can be identified from each other in X-ray images 30a and 30b based on their three-dimensional positions. Furthermore, even when solder balls 104 and 106 overlap each other, the three-dimensional positions of solder ball 104 and solder ball 106 can be obtained, allowing inspection of solder balls 104 and 106 while distinguishing them from each other.

[0056] Furthermore, in the first embodiment, as described above, the substrate 102 includes a plurality of solder balls 104 (first solder balls) arranged in a grid pattern on the surface 102a of the substrate 102, and a plurality of solder balls 106 (second solder balls) arranged in a grid pattern on the back surface 102b of the substrate 102. The control unit 21 (image analysis unit) calculates the three-dimensional position of each of the plurality of solder balls 104 and the plurality of solder balls 106 by triangulation based on the position of each of the plurality of solder balls 104 and the plurality of solder balls 106 in each of the X-ray images 30a (first image) and X-ray image 30b (second image), and the angle difference between the irradiation angle of the X-ray image 30a and the irradiation angle of the X-ray image 30b. With this configuration, the three-dimensional positions of the solder balls 104 and 106 located on the front surface 102a and back surface 102b of the substrate 102 can be obtained. Based on the obtained three-dimensional positions, it is possible to detect whether each of the solder balls 104 (solder ball 106) in the X-ray images 30a and 30b is located on the front surface 102a or the back surface 102b of the substrate 102. Therefore, when determining whether or not there are abnormalities in the solder balls 104 and 106 by analyzing the X-ray images 30a and 30b, it is possible to easily identify which surface of the substrate 102, front surface 102a or back surface 102b, the solder ball 104 or solder ball 106 determined to be abnormal is located on.

[0057] Furthermore, in the first embodiment, as described above, the control unit 21 (image analysis unit) detects the regions of solder balls 104 and solder balls 106 (target portion) in each of the X-ray image 30a (first image) and X-ray image 30b (second image), and acquires the centroid position of the detected regions of solder balls 104 and solder balls 106 as the positions of solder balls 104 and solder balls 106 in each of the X-ray image 30a and X-ray image 30b. With this configuration, even if the size or shape of solder balls 104 and solder balls 106 differs in each of the X-ray image 30a and X-ray image 30b, the three-dimensional position (coordinates) of solder balls 104 and solder balls 106 can be obtained by performing calculations on the acquired centroid position without performing calculations on all the pixels included in the regions of solder balls 104 and solder balls 106. Therefore, by obtaining the centroid position of the regions of solder ball 104 and solder ball 106, the processing burden of the calculations required to calculate the three-dimensional positions of solder ball 104 and solder ball 106 by triangulation can be reduced.

[0058] Furthermore, in the first embodiment, as described above, the control unit 21 (image analysis unit) uses a trained model 50 generated by machine learning to detect the regions of solder balls 104 and solder balls 106 (target areas) in each of the X-ray images 30a (first image) and 30b (second image), and acquires the positions of solder balls 104 and solder balls 106 in each of the X-ray images 30a and 30b based on the detected regions of solder balls 104 and solder balls 106. Here, when detecting regions using a trained model 50 generated by machine learning, regions can be detected with higher accuracy than when detecting regions by binarization using a threshold. Therefore, by using the trained model 50, the positions of solder balls 104 and solder balls 106 can be acquired with high accuracy based on the detected regions. As a result, the three-dimensional positions of solder balls 104 and solder balls 106 can be acquired with high accuracy by triangulation.

[0059] Furthermore, in the first embodiment, as described above, the solder balls 104 (first target portion) and 106 (second target portion) are different from each other, and the solder balls 104 and 106 have portions that overlap each other when viewed from the irradiation direction (Z direction), which is the direction from the X-ray irradiation unit 11 toward the X-ray detection unit 12, in at least one of the X-ray image 30a (first image) and the X-ray image 30b (second image). The control unit 21 (image analysis unit) uses a trained model 50 generated by machine learning to detect the region of solder ball 104 and the region of solder ball 106 in a separated state, and calculates the three-dimensional position of solder ball 104 and the three-dimensional position of solder ball 106 by obtaining the respective positions of solder ball 104 and solder ball 106 in each of the X-ray image 30a and X-ray image 30b based on the detected regions of solder ball 104 and solder ball 106. With this configuration, even if there is an overlap between the area of ​​solder ball 104 and the area of ​​solder ball 106, the trained model 50 can be used to accurately detect the solder ball 104 and the area of ​​solder ball 106 separately. Therefore, the inspection of the object to be inspected 101 can be performed with the solder ball 104 and the area of ​​solder ball 106 separated. As a result, when inspecting the object to be inspected 101, which includes solder balls 104 and solder ball 106, by irradiating the object to be inspected 101 with X-rays, the inspection can be easily performed regardless of the three-dimensional depth position of solder balls 104 and solder ball 106.

[0060] Furthermore, in the first embodiment, as described above, the control unit 21 (image analysis unit) uses a trained model 50 that has been trained to separately identify the overlapping portion 61 where the solder ball 104 (first target portion) and the solder ball 106 (second target portion) overlap each other when viewed from the irradiation direction, and the non-overlapping portion 62 where they do not overlap. Based on the separately identified overlapping portion 61 and non-overlapping portion 62, the control unit detects the regions of the solder ball 104 and the regions of the solder ball 106 in each of the X-ray images 30a (first image) and X-ray image 30b (second image). With this configuration, the trained model 50 can separately identify the overlapping portion 61 and the non-overlapping portion 62. Therefore, by combining the overlapping portion 61 and the non-overlapping portion 62 based on the identification result by the trained model 50, the regions of the solder ball 104 and the regions of the solder ball 106 can be detected accurately and easily. As a result, even when there is an overlap between the area of ​​solder ball 104 and the area of ​​solder ball 106, the three-dimensional positions of solder ball 104 and solder ball 106 can be calculated accurately and easily using triangulation.

[0061] (Effects of the X-ray image analysis method according to the first embodiment) The X-ray image analysis method of the first embodiment can provide the following effects.

[0062] In the X-ray image analysis method of the first embodiment, as described above, the three-dimensional positions of the solder balls 104 and 106 (target portion) contained in the object to be inspected 101 are calculated by triangulation based on the positions of the solder balls 104 and 106 (target portion) in each of the X-ray images 30a (first image) and 30b (second image), and the angle difference between the X-ray irradiation angle for generating the X-ray image 30a and the X-ray irradiation angle for generating the X-ray image 30b. As a result, the three-dimensional positions of the solder balls 104 and 106 calculated by triangulation can be obtained based on the positions of the solder balls 104 and 106 in the X-ray image 30a, the positions of the solder balls 104 and 106 in the X-ray image 30b, and the angle difference in the X-ray irradiation angle between the X-ray image 30a and the X-ray image 30b. Therefore, the inspection object 101 can be inspected while identifying the three-dimensional positions of the solder balls 104 and 106. As a result, when inspecting the inspection object 101, including the solder balls 104 and 106, by irradiating it with X-rays, an X-ray image analysis method can be provided that allows for easy inspection regardless of the three-dimensional depth position of the solder balls 104 and 106.

[0063] [Second Embodiment] A second embodiment will be described with reference to Figures 11 and 12. Unlike the first embodiment, which uses a trained model 50 that is trained to identify the overlapping portion 61 and the non-overlapping portion 62 separately, this second embodiment uses a trained model 250 that is trained to identify a target region 261 that includes both the overlapping portion and the non-overlapping portion without distinguishing between them. In the figures, parts with the same configuration as in the first embodiment are denoted by the same reference numerals.

[0064] (Configuration of the X-ray imaging system according to the second embodiment) As shown in Figure 11, the X-ray imaging system 200 according to the second embodiment includes a fluoroscopy device 10 and an analysis device 220. In the X-ray imaging system 200, the presence or absence of abnormalities in the solder balls 104 and solder balls 106 on the object to be inspected 101 is determined, similar to the X-ray imaging system 100 of the first embodiment. The configuration of the fluoroscopy device 10 is the same as in the first embodiment. The analysis device 220 includes a control unit 221 and a storage unit 222. The hardware configurations of the control unit 221 and the storage unit 222 are the same as those of the control unit 21 and the storage unit 22 of the first embodiment, respectively.

[0065] The control unit 221 includes a region detection unit 241. The control unit 221 functions as the region detection unit 241 by executing various programs (software) stored in the storage unit 222. That is, in Figure 11, the region detection unit 241 is shown as a software-based functional block. However, the region detection unit 241 may also be configured with a dedicated hardware circuit. The other configurations of the control unit 221 are the same as in the first embodiment. The storage unit 222 stores the learned model 250.

[0066] As shown in Figure 12, the region detection unit 241, similar to the region detection unit 41 in the first embodiment, uses a trained model 250 to acquire identification result images 231a and 231b from the X-ray images 30a and 30b generated by the fluoroscopy device 10. Unlike the first embodiment, the identification result images 231a and 231b are label images in which both the overlapping portion of the X-ray image 30a where solder balls 104 and 106 overlap when viewed from the irradiation direction (Z direction) and the non-overlapping portion containing only one solder ball 104 or solder ball 106 without overlapping are colored white, and the background portion 63 which does not contain either solder ball 104 or solder ball 106 is colored black.

[0067] In other words, unlike the trained model 50 of the first embodiment, the trained model 250 is trained and generated by machine learning to identify the target region 261 which includes both the overlapping and non-overlapping portions, without distinguishing between the overlapping and non-overlapping portions. The trained model 250 is generated by training it using deep learning to perform an image transformation (image reconstruction) on each pixel of the input image, which colors the target region 261 white and the background portion 63 black. The region detection unit 241 identifies the target region 261 in the X-ray image 30a and the X-ray image 30b, respectively, by obtaining identification result images 231a and 231b, in which the target region 261 and the background portion 63 are identified separately as a result of identification by the trained model 250.

[0068] The region detection unit 241 then detects areas within the identified target region 261 where the brightness value is smaller than a predetermined identification threshold as overlapping areas. Specifically, the region detection unit 241 extracts only the portion included in the target region 261 from each of the X-ray image 30a and X-ray image 30b. The region detection unit 241 then performs a binarization process on the extracted portion based on a predetermined identification threshold to distinguish between areas with relatively small brightness values ​​(dark) and areas with relatively large brightness values ​​(bright) from the portion included in the target region 261 in each of the X-ray image 30a and X-ray image 30b.

[0069] The region detection unit 241 then detects regions with a brightness value smaller than a predetermined identification threshold (relatively dark areas) as overlapping areas. The region detection unit 241 then detects regions with a brightness value larger than a predetermined identification threshold (relatively bright areas) as non-overlapping areas. In this way, the region detection unit 241 identifies overlapping and non-overlapping areas from the target region 261 and detects the regions of multiple solder balls 104 and the regions of multiple solder balls 106 in each of the X-ray images 30a and 30b in a separated state using control processing similar to that of the region detection unit 41 in the first embodiment. That is, in the state where the regions of multiple solder balls 104 and the regions of multiple solder balls 106 in each of the X-ray images 30a and 30b are separated, it is unclear which is the region of solder ball 104 and which is the region of solder ball 106, similar to regions 64 and 65 in the first embodiment.

[0070] The control unit 221 performs calculation processing using triangulation, similar to the first embodiment, to calculate the three-dimensional position (position in the direction of irradiation) of the solder balls 104 and the three-dimensional position (position in the direction of irradiation) of the solder balls 106 based on the regions of the detected solder balls 104 and the regions of the detected solder balls 106. Then, similar to the first embodiment, the control unit 221 determines whether or not there are abnormalities in the solder balls 104 and solder balls 106.

[0071] Furthermore, the other configurations of the X-ray imaging system 200 according to the second embodiment are the same as those of the first embodiment described above.

[0072] (Effects of the second embodiment) In the second embodiment, the following effects can be obtained.

[0073] In the second embodiment, as described above, the control unit 221 (image analysis unit) identifies the target region 261 in each of the X-ray images 30a (first image) and 30b (second image) by using a trained model 250 that has been trained to identify a target region 261 that includes both the overlapping portion and the non-overlapping portion of the solder ball 104 (first target portion) and the solder ball 106 (second target portion), without distinguishing between the overlapping portion where the solder ball 104 and the solder ball 106 overlap each other when viewed from the irradiation direction, and by detecting the region with a brightness value smaller than a predetermined identification threshold as the overlapping portion in the identified target region 261, thereby detecting the region of the solder ball 104 and the region of the solder ball 106 in each of the X-ray images 30a and 30b in a separated state. With this configuration, a trained model 250 is used that is trained to identify a target region 261 that includes both overlapping and non-overlapping parts, without distinguishing between them. Therefore, when generating training data to create the trained model 250, it is easier to generate the training data compared to when the training data is generated in a way that distinguishes between overlapping and non-overlapping parts. As a result, the burden of generating the trained model 250 can be reduced.

[0074] Furthermore, the other effects of the second embodiment are the same as those of the first embodiment described above.

[0075] [Differentiation] It should be noted that the embodiments disclosed herein are illustrative and not restrictive in all respects. The scope of the present invention is defined by the claims rather than by the description of the embodiments above, and further includes all modifications (exceptions) within the meaning and scope equivalent to the claims.

[0076] For example, the first and second embodiments described above show examples of calculating the three-dimensional positions of solder balls 104 and 106 that have overlapping portions, but the present invention is not limited thereto. In the present invention, the three-dimensional position of a target portion that does not have overlapping portions when viewed from the direction of X-ray irradiation may be calculated. Alternatively, instead of solder balls, voids (holes) contained in the object to be inspected may be used as the target portion, and the three-dimensional position may be calculated. In that case, the object to be inspected is not limited to electronic equipment including a substrate, but may also be a molded product made of metal or resin.

[0077] Furthermore, while the first and second embodiments described above show examples in which the centroid position of the detected solder ball 104 and solder ball 106 (target portion) is obtained as the position of solder ball 104 and solder ball 106, the present invention is not limited thereto. In the present invention, a position other than the centroid position may be obtained as the position of solder ball 104 and solder ball 106.

[0078] Furthermore, while the first and second embodiments described above show examples of detecting the regions of solder balls 104 and solder balls 106 (target areas) in X-ray image 30a (first image) and X-ray image 30b (second image), respectively, by using a trained model 50 (250) generated by machine learning, the present invention is not limited thereto. In the present invention, the regions of the target areas may be detected by performing image processing such as binarization on the first and second images without using a trained model.

[0079] Furthermore, in the first and second embodiments described above, examples were shown in which the drive unit 14 (angle changing unit) is configured to change the irradiation angle of X-rays irradiated onto the object to be inspected 101 by changing the positioning angle of the object to be inspected 101 relative to the X-ray irradiation unit 11, but the present invention is not limited thereto. In the present invention, the angle changing unit may be configured to change the irradiation angle of X-rays irradiated onto the object to be inspected by changing the positioning angle of the X-ray irradiation unit relative to the object to be inspected (the irradiation direction of X-rays irradiated from the X-ray irradiation unit to the X-ray detection unit). For example, the irradiation angle may be changed by rotating the X-ray irradiation unit and the X-ray detection unit without changing the positioning angle of the object to be inspected. Alternatively, the irradiation angle may be changed by changing the positioning position of the object to be inspected or the positioning position of the X-ray irradiation unit.

[0080] Furthermore, in the first and second embodiments described above, the image generation unit 15, which generates the X-ray image 30a (first image) and the X-ray image 30b (second image), is provided separately from the control unit 21 (image analysis unit), which performs triangulation to analyze the three-dimensional positions of the solder balls 104 and solder balls 106 (target portion) included in the X-ray image 30a and the X-ray image 30b. However, the present invention is not limited to this. In the present invention, the generation of the first and second images and the analysis of the three-dimensional positions of the target portion by triangulation may be performed by a common control device.

[0081] Furthermore, in the first and second embodiments described above, the three-dimensional positions of solder balls 104 (first target area) and solder balls 106 (second target area) contained in X-ray images 30a (first image) and 30b (second image), which have different irradiation angles, are calculated by triangulation, and the presence or absence of abnormalities in solder balls 104 and solder balls 106 is determined using the calculated three-dimensional positions. However, the present invention is not limited to this. In the present invention, instead of calculating the three-dimensional positions of the first target area and the second target area, a trained model may be used to detect the regions of the first target area and the second target area in a single X-ray image in a separated state, and the presence or absence of abnormalities in the first target area and the second target area may be determined based on the regions of the first target area and the second target area detected in a separated state.

[0082] [Aspect] Those skilled in the art will understand that the exemplary embodiments described above are specific examples of the following embodiments.

[0083] (Item 1) An X-ray irradiation unit that irradiates the object to be inspected, including the target area, with X-rays, An X-ray detection unit for detecting X-rays emitted from the X-ray irradiation unit, An angle changing unit for changing the irradiation angle of X-rays irradiated onto the object to be inspected, An image generation unit generates a first image and a second image, which are X-ray images with different irradiation angles, based on the X-rays irradiated by the X-ray irradiation unit and detected by the X-ray detection unit while changing the irradiation angle by the angle changing unit. An X-ray imaging system comprising: an image analysis unit that calculates the three-dimensional position of the target portion by triangulation based on the position of the target portion included in the object to be inspected in each of the first image and the second image generated by the image generation unit, and the angle difference between the irradiation angle of the X-rays for generating the first image and the irradiation angle of the X-rays for generating the second image.

[0084] (Item 2) The angle changing unit changes the irradiation angle by changing at least one of the following: the positioning angle of the object to be inspected relative to the X-ray irradiation unit, the position of the object to be inspected, the positioning angle of the X-ray irradiation unit relative to the object to be inspected, and the position of the X-ray irradiation unit. The X-ray imaging system according to item 1, wherein the image analysis unit calculates the three-dimensional position of the target portion by triangulation without performing reconstruction processing on the first image and the second image.

[0085] (Item 3) The aforementioned target portion includes a first target portion and a second target portion that are different from each other. The X-ray imaging system according to item 1 or 2, wherein the image analysis unit calculates the three-dimensional position of the first target portion and the three-dimensional position of the second target portion by triangulation based on the positions of the first target portion and the second target portion in each of the first image and the second image generated by the image generation unit, and the angular difference between the irradiation angle of the X-rays for generating the first image and the irradiation angle of the X-rays for generating the second image.

[0086] (Item 4) The first target portion includes a plurality of first solder balls arranged in a grid pattern on the surface of the substrate, The second target portion includes a plurality of second solder balls arranged in a grid pattern on the back surface of the substrate, The X-ray imaging system according to item 3, wherein the image analysis unit calculates the three-dimensional position of each of the plurality of first solder balls and the plurality of second solder balls by triangulation based on the positions of each of the plurality of first solder balls and the plurality of second solder balls in each of the first image and the second image, and the angular difference between the irradiation angle of the first image and the irradiation angle of the second image.

[0087] (Item 5) The X-ray imaging system according to any one of items 1 to 4, wherein the image analysis unit detects the region of the target portion in each of the first and second images, and acquires the centroid position of the detected region of the target portion as the position of the target portion in each of the first and second images.

[0088] (Item 6) The X-ray imaging system according to any one of items 1 to 5, wherein the image analysis unit detects the region of the target portion in each of the first and second images by using a trained model generated by machine learning, and obtains the position of the target portion in each of the first and second images based on the detected region of the target portion.

[0089] (Item 7) The aforementioned target portion includes a first target portion and a second target portion that are different from each other. The first target portion and the second target portion have portions that overlap each other when viewed from the irradiation direction, which is the direction from the X-ray irradiation unit toward the X-ray detection unit, in at least one of the first image and the second image. The X-ray imaging system according to item 6, wherein the image analysis unit detects the region of the first target portion and the region of the second target portion in a separated state by using the trained model generated by machine learning, and calculates the three-dimensional position of the first target portion and the three-dimensional position of the second target portion by obtaining the positions of the first target portion and the second target portion in the first image and the second image, respectively, based on the detected regions of the first target portion and the second target portion.

[0090] (Item 8) The X-ray imaging system according to item 7, wherein the image analysis unit uses the trained model, which has been trained to separately identify overlapping portions where the first target portion and the second target portion overlap each other when viewed from the irradiation direction, and non-overlapping portions where they do not overlap, and detects the region of the first target portion and the region of the second target portion in each of the first and second images in a separated state based on the separately identified overlapping portions and non-overlapping portions.

[0091] (Item 9) The X-ray imaging system according to item 7, wherein the image analysis unit identifies the target region in each of the first and second images by using the trained model, which has been trained to identify a target region that includes both the overlapping portion and the non-overlapping portion of the first and second target regions, without distinguishing between the overlapping portion and the non-overlapping portion of the first and second target regions when viewed from the irradiation direction, and by detecting as the overlapping portion a region with a brightness value smaller than a predetermined identification threshold within the identified target region, thereby detecting the region of the first target region and the region of the second target region in each of the first and second images in a separated state.

[0092] (Item 10) An X-ray irradiation unit that irradiates an object to be inspected, which includes a first target part and a second target part, which are different target parts from each other, An X-ray detection unit for detecting X-rays emitted from the X-ray irradiation unit, An image generation unit generates an X-ray image based on the X-rays irradiated by the X-ray irradiation unit and detected by the X-ray detection unit, An X-ray imaging system comprising: an image analysis unit that detects the region of the first target portion and the region of the second target portion in the X-ray image in a separated state based on the separately identified overlapping portion and non-overlapping portion, by using a trained model that has been trained to separately identify the overlapping portion and non-overlapping portion of the first target portion and the second target portion in the X-ray image generated by the image generation unit, when viewed from the irradiation direction which is the direction from the X-ray irradiation unit to the X-ray detection unit.

[0093] (Item 11) The steps include acquiring a first image and a second image, which are X-ray images with different irradiation angles, generated based on X-rays detected by irradiating an object to be inspected, including the target portion, while changing the irradiation angle, An X-ray image analysis method comprising the steps of: calculating the three-dimensional position of the target portion included in the object to be inspected in each of the first image and the second image by triangulation, based on the position of the target portion and the angle difference between the irradiation angle of the X-rays for generating the first image and the irradiation angle of the X-rays for generating the second image. [Explanation of symbols]

[0094] 11 X-ray irradiation section 12 X-ray detection unit 14. Drive unit (angle changing unit) 15 Image generation unit 21, 221 Control Unit (Image Analysis Unit) 30a X-ray image (first image, X-ray image) 30b X-ray image (second image, X-ray image) 50,250 pre-trained models 61 Overlapping parts 62 Non-superimposed portion 100, 200 X-ray imaging systems 101 Items to be inspected 102 circuit boards 102a surface 102b Reverse side 104 Solder ball (target part, first target part, first solder ball) 106 Solder ball (target part, second target part, second solder ball) 261 Target Area

Claims

1. An X-ray irradiation unit that irradiates an object to be inspected, including a target portion placed on a substrate, An X-ray detection unit that emits X-rays irradiated from the aforementioned X-ray irradiation unit, An angle changing unit for changing the irradiation angle of X-rays irradiated onto the object to be inspected, An image generation unit generates a first image and a second image, which are X-ray images with different irradiation angles, based on the X-rays irradiated by the X-ray irradiation unit and detected by the X-ray detection unit while changing the irradiation angle by the angle changing unit. The system includes an image analysis unit that calculates the three-dimensional position of the target portion by triangulation based on the position of the target portion included in the object to be inspected in each of the first and second images generated by the image generation unit, and the angle difference between the irradiation angle of the X-rays for generating the first image and the irradiation angle of the X-rays for generating the second image. The image analysis unit determines whether the target portion is located on the front or back surface of the substrate, and is an X-ray imaging system.

2. The angle changing unit changes the irradiation angle by changing at least one of the following: the positioning angle of the object to be inspected relative to the X-ray irradiation unit, the position of the object to be inspected, the positioning angle of the X-ray irradiation unit relative to the object to be inspected, and the position of the X-ray irradiation unit. The X-ray imaging system according to claim 1, wherein the image analysis unit calculates the three-dimensional position of the target portion by triangulation without performing reconstruction processing on the first image and the second image.

3. An X-ray irradiation unit that irradiates an object to be inspected, including the target portion, An X-ray detection unit that emits X-rays irradiated from the aforementioned X-ray irradiation unit, An angle changing unit for changing the irradiation angle of X-rays irradiated onto the object to be inspected, An image generation unit generates a first image and a second image, which are X-ray images with different irradiation angles, based on the X-rays irradiated by the X-ray irradiation unit and detected by the X-ray detection unit while changing the irradiation angle by the angle changing unit. The system includes an image analysis unit that calculates the three-dimensional position of the target portion by triangulation based on the position of the target portion included in the object to be inspected in each of the first and second images generated by the image generation unit, and the angle difference between the irradiation angle of the X-rays for generating the first image and the irradiation angle of the X-rays for generating the second image. The aforementioned target portion includes a first target portion and a second target portion that are different from each other. The image analysis unit calculates the three-dimensional position of the first target portion and the three-dimensional position of the second target portion by triangulation, based on the positions of the first target portion and the second target portion in each of the first image and the second image generated by the image generation unit, and the angular difference between the irradiation angle of the X-rays for generating the first image and the irradiation angle of the X-rays for generating the second image.

4. The first target portion includes a plurality of first solder balls arranged in a grid pattern on the surface of the substrate, The second target portion includes a plurality of second solder balls arranged in a grid pattern on the back surface of the substrate, The X-ray imaging system according to claim 3, wherein the image analysis unit calculates the three-dimensional position of each of the plurality of first solder balls and the plurality of second solder balls by triangulation based on the positions of each of the plurality of first solder balls and the plurality of second solder balls in each of the first image and the second image, and the angular difference between the irradiation angle of the first image and the irradiation angle of the second image.

5. The X-ray imaging system according to claim 1 or 2, wherein the image analysis unit detects the region of the target portion in each of the first and second images, and acquires the centroid position of the detected region of the target portion as the position of the target portion in each of the first and second images.

6. An X-ray irradiation unit that irradiates an object to be inspected, including the target portion, An X-ray detection unit that emits X-rays irradiated from the aforementioned X-ray irradiation unit, An angle changing unit for changing the irradiation angle of X-rays irradiated onto the object to be inspected, An image generation unit generates a first image and a second image, which are X-ray images with different irradiation angles, based on the X-rays irradiated by the X-ray irradiation unit and detected by the X-ray detection unit while changing the irradiation angle by the angle changing unit. The system includes an image analysis unit that calculates the three-dimensional position of the target portion by triangulation based on the position of the target portion included in the object to be inspected in each of the first and second images generated by the image generation unit, and the angle difference between the irradiation angle of the X-rays for generating the first image and the irradiation angle of the X-rays for generating the second image. The X-ray imaging system includes an image analysis unit that uses a trained model generated by machine learning to detect the region of the target portion in each of the first and second images, and obtains the position of the target portion in each of the first and second images based on the detected region of the target portion.

7. The aforementioned target portion includes a first target portion and a second target portion that are different from each other. The first target portion and the second target portion have portions that overlap each other when viewed from the irradiation direction, which is the direction from the X-ray irradiation unit toward the X-ray detection unit, in at least one of the first image and the second image. The X-ray imaging system according to claim 6, wherein the image analysis unit detects the region of the first target portion and the region of the second target portion in a separated state by using the trained model generated by machine learning, and calculates the three-dimensional position of the first target portion and the three-dimensional position of the second target portion by obtaining the positions of the first target portion and the second target portion in the first image and the second image, respectively, based on the detected regions of the first target portion and the second target portion.

8. The X-ray imaging system according to claim 7, wherein the image analysis unit uses the trained model, which has been trained to separately identify overlapping portions and non-overlapping portions of the first target portion and the second target portion that overlap each other when viewed from the irradiation direction, and detects the region of the first target portion and the region of the second target portion in each of the first and second images in a separated state based on the separately identified overlapping portions and non-overlapping portions.

9. The X-ray imaging system according to claim 7, wherein the image analysis unit identifies the target region in each of the first and second images by using the trained model which has been trained to identify a target region that includes both the overlapping portion and the non-overlapping portion of the first target region and the second target region, without distinguishing between the overlapping portion and the non-overlapping portion of the first target region and the non-overlapping portion when viewed from the irradiation direction, and by detecting as the overlapping portion a region with a brightness value smaller than a predetermined identification threshold in the identified target region, thereby detecting the region of the first target region and the region of the second target region in each of the first and second images in a separated state.

10. An X-ray irradiation unit that irradiates an object to be inspected, which includes a first target part and a second target part, which are different target parts from each other, An X-ray detection unit that emits X-rays irradiated from the aforementioned X-ray irradiation unit, An image generation unit generates an X-ray image based on the X-rays irradiated by the X-ray irradiation unit and detected by the X-ray detection unit, An X-ray imaging system comprising: an image analysis unit that detects the region of the first target portion and the region of the second target portion in the X-ray image in a separated state based on the separately identified overlapping portion and non-overlapping portion, by using a trained model that has been trained to separately identify the overlapping portion and non-overlapping portion of the first target portion and the second target portion in the X-ray image generated by the image generation unit, when viewed from the irradiation direction which is the direction from the X-ray irradiation unit to the X-ray detection unit.

11. A step of acquiring a first image and a second image, which are X-ray images with different irradiation angles, generated based on X-rays detected by irradiating an object to be inspected, including a target portion placed on a substrate, while changing the irradiation angle, A step of calculating the three-dimensional position of the target portion by triangulation based on the position of the target portion included in the object to be inspected in each of the first and second images, and the angle difference between the irradiation angle of the X-rays for generating the first image and the irradiation angle of the X-rays for generating the second image. An X-ray image analysis method comprising the step of determining whether the target portion is located on the front surface or the back surface of the substrate.

Images