JP2010127810A - X-ray inspection apparatus and x-ray inspection method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an apparatus capable of inspecting an object accurately, in a short time. <P>SOLUTION: An X-ray inspection apparatus 100 includes a scanning type X-ray source 10, a plurality of X-ray detectors 23.1 to 23.N, an image acquisition control mechanism 30, a computation section 70, and a memory 90. The scanning type X-ray source 10 outputs X rays to an inspection object 20. The X-ray detectors 23.1 to 23.N detect the X rays transmitted through the inspection object 20. The image acquisition control mechanism 30 stores image data 98 detected by the X-ray detectors 23.1 to 23.N in the memory 90. A reconstruction section 76, included in the computation section 70 based on the image data 98, generates reconstructed images of a horizontal cross-section by using an analytical method and generates the reconstructed images of a vertical cross-section by an iteration method. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to an X-ray inspection apparatus and an X-ray inspection method. In particular, the present invention relates to a method for generating a reconstructed image for inspecting an object using X-ray irradiation, and relates to a technique applicable to an X-ray inspection method and an X-ray inspection apparatus.

  Currently, X-ray CT (Computed Tomography) is used in examinations in various fields. CT is a technique for generating (reconstructing) three-dimensional data of an object from two-dimensional data (projected image) of the imaged object. In an inspection using CT, information that cannot be obtained from two-dimensional data, such as a three-dimensional shape of an object and images of the object viewed from various angles, can be obtained.

  In recent years, X-ray CT has been used for inspection of a printed circuit board (PCB). The PCB inspection will be described with reference to FIG. FIG. 21 is a diagram for explaining the PCB inspection.

  FIG. 21A is a perspective view of a substrate on which electronic components are mounted. A first component 1602 and a second component 1603 are mounted on the substrate 1601. The second component 1603 is physically and electrically connected to the substrate 1601 by a BGA (Ball Grid Array) 1604 or the like.

  FIG. 21B is a cross-sectional view in which a connection portion between the substrate 1601 and the second component 1603 is cut by a cross section perpendicular to the surface of the substrate 1601. The BGA 1604 connects the second component 1603 and the surface layer 1605 of the substrate 1601. The BGA 1604 is heated and deformed to a heated state 1606. However, a void 1607 may occur in the state 1606 after heating. In some cases, a plurality of solder balls are combined to form a bridge 1608.

  In the PCB inspection, the wettability of the state 1606 after heating, the presence / absence of voids 1607 and bridges 1608, the presence / absence of foreign matter, and the like are inspected. For the inspection of the solder joint surface, as a known technique, a technique mounted on various apparatuses can be used. Here, a general method for determining by the solder area in the fault is shown.

  The inspection apparatus generates three-dimensional data, cuts out the three-dimensional data, and creates a tomographic image. The inspection apparatus binarizes the created tomographic image, and acquires a binarized image obtained by separating the image into solder and other parts. For this binarization process, a general binarization process such as a discriminant analysis method can be used. The inspection apparatus labels the white (or 1) portion from the binarized image, and acquires a labeling image in which solder is distinguished. For this labeling process, it is possible to use a general labeling process that determines the presence or absence of connection by raster scanning.

  An example of a cross section parallel to the surface of the substrate 1601 is shown in FIG. 21C. FIG. 21C is a cross-sectional view of a connection location cut along a cross-section indicated by a broken line in FIG. 21B. In FIG. 21C, the solder is indicated by white and the parts other than the solder are indicated by hatching. Here, three types of states, normal, void, and bridge, are shown. Referring to FIG. 21C, when there is void 1607, a portion without solder is generated in the solder. When the bridge 1608 is present, solder is observed in a wider area than in the normal state.

  The inspection apparatus counts the area of each solder (the number of white or one pixel) from the labeling image, and obtains the area of the solder. The inspection device determines the quality of the solder joint surface by determining that the product is a good product if the area is within a certain range and that the product is defective if the area is not. In general, the predetermined range of threshold values is set in advance by the user.

  For example, Patent Document 1 (Japanese Patent Laid-Open No. 2008-026344) discloses a printed circuit board inspection apparatus. An outline of the printed circuit board inspection apparatus is shown in FIGS. FIG. 22 is a diagram schematically showing the inspection apparatus described in Patent Document 1. As shown in FIG. FIG. 23 is a schematic diagram of an inspection apparatus that inspects a printed circuit board by moving an X-ray source and an X-ray detector. As shown in FIGS. 22 and 23, the printed circuit board inspection apparatus irradiates the printed circuit board with X-rays from an oblique direction with respect to the surface on which the printed circuit board is conveyed.

  Incidentally, in general, in order to obtain a high-quality image by X-ray CT, it is desirable to perform fluoroscopic imaging of X-rays from all directions. Patent Document 2 (Japanese Patent Laid-Open No. 2006-25868) discloses an X-ray CT system for medical use that images a subject from all directions around it.

On the other hand, Patent Document 3 (Japanese Patent Laid-Open No. 2007-139620) discloses an effective data reconstruction method for a case where the cone beam angle is large. The X-ray cone beam CT apparatus described in Patent Document 3 supplements shadow zone data in Radon space by calculation.
JP 2008-026344 A JP 2006-25868 A JP 2007-139620 A

  In the inspection apparatus described in Patent Document 1, the X-ray CT inspection of a thin object such as a PCB cannot be accurately performed in a short time. Therefore, it has been difficult to realize an in-line inspection of a PCB or the like. This is because in X-ray CT examination, there is a trade-off relationship between the number of imaging data and the resolution of reconstruction data. In the oblique CT imaging, data on the PCB viewed from the long axis direction (hereinafter sometimes referred to as vertical section data) is insufficient, so that the resolution of the vertical section image of the reconstruction data is deteriorated.

  However, in the inspection of the PCB, it is not realistic from the viewpoint of the inspection time to image the object from all the surrounding directions as in the medical inspection described in Patent Document 2. In addition, it is difficult to image a very small electronic component mounted at a high density in the long axis direction of the PCB (hereinafter sometimes referred to as the horizontal direction) with high resolution.

  The method described in Patent Document 3 relates to a wide angle cone beam CT. It is not easy to apply the complementing method described in Patent Document 3 to PCB inspection.

  The present invention has been made to solve the above-described problems, and an object of the present invention is to provide an apparatus capable of accurately inspecting an object in a short time.

  The present invention according to one aspect is an X-ray inspection apparatus that inspects an object using X-rays, and outputs X-rays so that the X-rays enter the object from a plurality of directions. And X-ray detection means for capturing X-rays at positions where X-rays incident on the object and transmitted through the object from each direction, and intensity distribution data of the X-rays captured at each position, Reconstructing means for generating a plurality of reconstructed image data, and the reconstructing means generates a first reconstructed image data corresponding to a plane whose normal direction is the first direction. The first reconstruction means uses an analysis method in the first reconstructed image data generation process, and corresponds to a plane whose normal direction is a second direction different from the first direction. And further comprising second reconstruction means for generating second reconstructed image data of the object, Reconstruction means, in the generation process of the second reconstruction image data, using the iterative method.

  Preferably, the reconstruction unit performs generation of the first reconstructed image data and generation of the second reconstructed image data in parallel.

  Preferably, the second constructing unit generates the second reconstructed image data using the first reconstructed image data as the initial image of the iterative method.

  Preferably, the reconstruction means further includes a determination means for determining a region in which the second reconstruction means generates the second reconstructed image data based on the reconstructed three-dimensional data obtained by the analysis method. The reconstruction unit generates second reconstructed image data for the determined region.

  Preferably, the X-ray output means outputs the X-ray so that the X-ray enters the object from a direction other than the second direction.

  Preferably, the apparatus further includes a transport mechanism that transports the object along a transport surface whose normal direction is substantially the second direction, and the X-ray output means outputs X-rays from the first side of the transport surface. The X-ray detection means images X-rays on a second side different from the first side of the transport surface.

  Preferably, the object is a planar object, the second direction is substantially the normal direction of the surface of the object, and the X-ray output means is at a predetermined angle with respect to the surface of the object. X-rays are output so that the X-rays enter the object from the intersecting direction.

  More preferably, the object is a planar object, the second direction is substantially the normal direction of the surface of the object, and the determining means is the object of the reconstructed three-dimensional data by the analysis method The region for generating the second reconstructed image data is limited in the second direction based on the pixel distribution of a plurality of cross sections substantially parallel to the surface of the object.

  More preferably, it further includes a transport mechanism that transports the object along a transport surface whose normal direction is substantially the second direction, and the transport mechanism includes a pair of rails that sandwich the object.

  The present invention according to another aspect is an X-ray inspection method for inspecting an object using X-rays, the step of outputting X-rays so that X-rays enter the object from a plurality of directions, A step of imaging X-rays at positions where X-rays incident on the object from the direction and transmitted through the object reach, and a plurality of reconstructed image data of the object based on the intensity distribution data of the X-rays captured at each position And generating the reconstructed image data includes generating first reconstructed image data corresponding to a plane having the first direction as a normal direction, and The step of generating the configuration image data includes a step of using an analysis method in the generation processing of the first reconstruction image data, and the step of generating the reconstruction image data is a second different from the first direction. On a plane whose direction is normal Generating a second reconstructed image data of the corresponding object, wherein the second reconstructed image data is generated by using an iterative method in the generation process of the second reconstructed image data. Have

  According to the present invention, the reconstructed image data corresponding to the surface having the first direction as the normal direction is generated using the analysis method, and the surface having the second direction different from the first direction as the normal direction Reconstructed image data corresponding to is generated using an iterative method. As a result, the object can be inspected with high accuracy in a short time.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the following description, the same parts are denoted by the same reference numerals. Their names and functions are also the same. Therefore, detailed description thereof will not be repeated.

[First Embodiment]
(1. About X-ray inspection equipment)
FIG. 1 is a schematic block diagram of an X-ray inspection apparatus 100 according to the present invention. An X-ray inspection apparatus 100 according to the present invention will be described with reference to FIG. However, the configurations, dimensions, shapes, and other relative arrangements described below are not intended to limit the scope of the present invention only to those unless otherwise specified.

  The X-ray inspection apparatus 100 includes a scanning X-ray source 10 that outputs X-rays with a central axis as an axis 21, and a plurality of X-ray detectors 23.1 to 23. N is attached and each X-ray detector 23.1-23. And an X-ray detector driving unit 22 for driving N to a designated position. Further, the scanning X-ray source 10 and the X-ray detectors 23.1 to 23. An inspection object 20 is arranged between N and N. In other words, the X-ray detectors 23.1 to 23. N is at a position facing the scanning X-ray source 10 so as to sandwich the inspection object 20.

  That is, the scanning X-ray source 10 is disposed on the first side of the surface on which the inspection object 20 is disposed, and outputs X-rays from the first side toward the inspection object 20. Also, each X-ray detector 23.1-23. N is arranged on a second side different from the first side of the surface on which the inspection object 20 is arranged (that is, on the opposite side).

  Further, the X-ray inspection apparatus 100 is configured to drive the X-ray detector 23 by the X-ray detector driving unit 22 and to detect the X-ray detectors 23.1 to 23. An image acquisition control mechanism 30 for controlling acquisition of image data from N, an input unit 40 for receiving an instruction input from a user, and an output unit 50 for outputting measurement results and the like to the outside. .

  The X-ray inspection apparatus 100 further includes a scanning X-ray source control mechanism 60, a calculation unit 70, and a memory 90. In such a configuration, the calculation unit 70 executes a program 96 stored in the memory 90 to control each unit, and performs predetermined calculation processing.

  The scanning X-ray source 10 is controlled by the scanning X-ray source control mechanism 60 and irradiates the inspection target 20 with X-rays.

  FIG. 2 is a cross-sectional view showing the configuration of the scanning X-ray source 10. Referring to FIG. 2, in scanning X-ray source 10, electron beam 16 is irradiated onto target 11 such as tungsten from electron gun 19 controlled by electron beam control unit 62. Then, X-rays 18 are generated and emitted (output) from the place where the electron beam 16 collides with the target (X-ray focal position 17). The electron beam system is housed in the vacuum vessel 9. The inside of the vacuum vessel 9 is kept in a vacuum by a vacuum pump 15, and an electron beam 16 accelerated by a high voltage power source 14 is emitted from an electron gun 19.

  In the scanning X-ray source 10, the electron beam 16 is converged by the electron beam converging coil 13, and then deflected by the deflection yoke 12, so that the location where the electron beam 16 collides with the target 11 is arbitrarily determined. Can be changed. For example, the electron beam 16a deflected by the deflection yoke 12 collides with the target 11, and the X-ray 18a is output from the X-ray focal position 17a. Similarly, the electron beam 16b deflected by the deflection yoke 12 collides with the target 11, and an X-ray 18b is output from the X-ray focal position 17b. Unless otherwise specified, in the present invention, the scanning X-ray source 10 is a transmissive type. In addition, when X-rays are generated from a position (hereinafter referred to as “X-ray emission starting position”) that should be the starting point of X-ray emission set according to the inspection target portion of the inspection object, that position. It is desirable that the target is not a ring but a continuous surface so that the degree of freedom of setting can be increased. Further, in the following description, when the positions are not particularly distinguished and described, they are simply indicated as the X-ray focal position 17 as a generic name.

  In order to move the X-ray focal position to the above-described X-ray emission starting positions, for example, a fixed-focus X-ray source is used as the X-ray source, and the position of the X-ray source itself is changed each time. It can also be moved mechanically. However, in the configuration as shown in FIG. 2, when the X-ray focal point position is moved to the X-ray emission starting position, the X-ray source is mechanically moved within a certain range. An X-ray inspection apparatus excellent in maintainability and reliability can be realized without necessity. It is also possible to provide a plurality of X-ray sources and switch them according to the starting position. The X-ray source according to the present invention may be used for oblique CT imaging.

  In other words, the “starting position of X-ray radiation” means that the spatial position of the X-ray detector 23 used for imaging and the spatial position of the inspection target portion of the inspection target 20 are specified. It means a spatial position that can be specified, and the X-ray focal position means a position on the target where X-rays are actually output. Therefore, in order to bring the X-ray focal position to the “X-ray emission starting position”, it is possible to scan the electron beam with a scanning X-ray source, or mechanically move the X-ray source itself. It may be moved.

  Returning to FIG. 1, the scanning X-ray source control mechanism 60 includes an electron beam control unit 62 that controls the output of the electron beam. The electron beam control unit 62 receives an X-ray focal position and X-ray energy (tube voltage, tube current) designation from the calculation unit 70. X-ray energy varies depending on the configuration of the inspection object.

  The inspection object 20 is a scanning X-ray source 10 and an X-ray detector 23 (hereinafter, when “X-ray detectors 23.1 to 23.N” are collectively referred to as “X-ray detector 23”). It is arranged between. In moving the position of the inspection object 20, in this embodiment, the inspection object is arranged at a position for inspection by moving in one direction like a belt conveyor. However, it may be moved to an arbitrary position on the XYZ stage.

  If the inspection target is small like a printed circuit board, the scanning X-ray source 10 and the X-ray detector 23 may be fixed and the inspection target may be moved as described above. When the object has a large area and it is difficult to arbitrarily move the inspection object side, the relative position between the scanning X-ray source 10 and the X-ray detector 23 is fixed and the scanning X-ray source 10 is fixed. The X-ray detector 23 may be moved.

  The X-ray detector 23 is a two-dimensional X-ray detector that detects and images the X-rays output from the scanning X-ray source 10 and transmitted through the inspection object 20. For example, I.I. I. (Image Intensifier) tube. In this embodiment, since a plurality of X-ray detectors are arranged in the X-ray detector driving unit 22, a space-efficient FPD (flat panel detector) is desirable. Further, it is desirable to have high sensitivity so that it can be used for in-line inspection, and it is particularly desirable to use a direct conversion FPD using CdTe.

  The image acquisition control mechanism 30 includes a detector drive control unit 32 for controlling the X-ray detector drive unit 22 to move the X-ray detector 23 to a position specified by the calculation unit 70, and a calculation unit 70. And an image data acquisition unit 34 for acquiring the image data 98 of the designated X-ray detector 23. The image data acquisition unit 34 stores the acquired image data 98 in the memory 90.

  The position of the X-ray detector 23 driven by the X-ray detector drive unit 22 can be known by a position sensor (not shown), and can be taken into the calculation unit 70 via the detector drive control unit 32.

  Moreover, it is desirable that the X-ray detector driving unit 22 can be moved up and down in order to adjust the enlargement ratio. In this case, the vertical position of the X-ray detector drive unit 22 can be known by a sensor (not shown), and can be taken into the calculation unit 70 via the detector drive control unit 32.

  In the present embodiment, X-ray detectors 23.1 to 23. N, the X-ray detector drive unit 22, and the image acquisition control mechanism 30 are combined to detect X-rays that have passed through the inspection object 20, but the X-ray detection method is not limited to this. Absent. The X-ray inspection apparatus 100 only needs to have a mechanism capable of detecting X-rays at a position where X-rays incident on the inspection object 20 from various directions and transmitted through the inspection object 20 reach. For example, the X-ray inspection apparatus 100 may detect X-rays with more fixed X-ray detectors. A single X-ray detector may be used instead of the plurality of X-ray detectors.

  The input unit 40 is an operation input device for receiving user input. The output unit 50 is a display for displaying an X-ray image or the like configured by the calculation unit 70. That is, the user can execute various inputs via the input unit 40, and various calculation results obtained by the processing of the calculation unit 70 are displayed on the output unit 50. The image displayed on the output unit 50 may be output for visual quality judgment by the user, or may be output as a quality judgment result of a quality judgment unit 78 described later.

  The calculation unit 70 includes a scanning X-ray source control unit 72, an image acquisition control unit 74, a reconstruction unit 76, a pass / fail determination unit 78, an inspection target position control unit 80, and an X-ray focal position calculation unit 82. And an imaging condition setting unit 84. The arithmetic unit 70 corresponding to the processor executes the program 96 stored in the memory 90 and realizes the functions of the above-described units. However, part or all of the functions of the arithmetic unit 70 may be realized by hardware.

  The scanning X-ray source control unit 72 determines the X-ray focal position and X-ray energy, and sends a command to the scanning X-ray source control mechanism 60.

  The image acquisition control unit 74 determines the X-ray detector 23 that acquires an image from among the X-ray detectors 23 driven to the specified position by the X-ray detector driving unit 22, and issues a command to the image acquisition control mechanism 30. send. Further, image data is acquired from the image acquisition control mechanism 30.

  The reconstruction unit 76 reconstructs three-dimensional data from a plurality of image data acquired by the image acquisition control unit 74.

  The quality determination unit 78 determines the quality of the inspection target based on the three-dimensional data or the perspective data reconstructed by the reconstruction unit 76. For example, the quality is determined by recognizing the shape of the solder ball and determining whether or not the shape is within a predetermined allowable range. Note that the algorithm for performing pass / fail judgment or the input information to the algorithm is obtained from the imaging condition information 94 because it differs depending on the inspection target.

  The inspection target position control unit 80 controls the inspection target position driving mechanism 110 that moves the inspection target 20.

  When inspecting an inspection area where the inspection object 20 is present, the X-ray focal position calculation unit 82 calculates an X-ray focal position and an irradiation angle with respect to the inspection area.

  The imaging condition setting unit 84 sets conditions for outputting X-rays from the scanning X-ray source 10 according to the inspection object 20. For example, the applied voltage to the X-ray source and the imaging time.

  The memory 90 relates to X-ray focal position information 92 in which the X-ray focal position calculated by the X-ray focal position calculator 82 is stored, imaging conditions set by the imaging condition setting unit 84, and an algorithm for determining pass / fail. In addition to the imaging condition information 94 in which information and the like are stored, a program for realizing each function executed by the arithmetic unit 70 described above is included. The memory 90 only needs to be able to store data, and is configured by a storage device such as a RAM (Random Access Memory), an EEPROM (Electrically Erasable and Programmable Read-Only Memory), an HDD (Hard Disc Drive), or the like.

  FIG. 3 is a diagram for explaining the configuration of the X-ray inspection apparatus 100. The same parts as those in FIG. 1 are denoted by the same reference numerals, and the description is made among the parts directly related to the control of the X-ray focal position, the control of the X-ray detector position, the control of the inspection target position, and the like. The necessary parts are extracted and described.

  Referring to FIG. 3, the X-ray detector driving unit 22 has an XYθ operation mechanism capable of driving the X-ray detectors 23.1 and 23.2 with XYθ degrees of freedom. For example, a scanning X-ray source is used. The X-ray detector driving unit 22 is not limited to the arm type driving mechanism shown in FIG. Any X-ray detector driving unit 22 can be used as long as it plays a role of moving and rotating each X-ray detector 23 to an arbitrary position.

  In the configuration shown in FIG. 3, an inspection target position drive mechanism 110 and an inspection target position control unit 80 are provided to move the position of the inspection target.

  In FIG. 3, two independently movable X-ray detectors are used. However, two or more X-ray detectors may be provided. A plurality of fixed X-ray detectors may be used. In this case, each X-ray detector shall be arrange | positioned so that each imaging data of the direction required for a test | inspection can be imaged with any X-ray detector. However, an example using two independently movable X-ray detectors will be described below.

  The X-ray detector 23.1 and the X-ray detector 23.2 are capable of XY-θ operation independently. Note that the θ rotation mechanism does not necessarily have to be provided depending on how the X-ray detector 23 is driven.

  The X-ray detector drive unit 22 includes an orthogonal type biaxial robot arm 22.1 and a detector support unit 22.2 having a rotation axis, and moves and rotates the X-ray detector 23. However, such a configuration that enables such movement in the XY direction or θ rotation in the XY plane and has the same function for the movement of the X-ray detector 23 is not necessary. It is also possible to use mechanisms other than.

  In addition, the visual field of the inspection object is made XY independently from the X-ray detector 23.1 or 23.2 by the inspection object position driving mechanism 110 controlled by the inspection object position control unit 80 in the calculation unit 70. -Z operation is possible. Furthermore, as described above, the scanning X-ray source of the X-ray source 10 can move the X-ray focal point position 17 to an arbitrary position on the X-ray target at high speed.

  The calculation unit 70 sends a command to the detector drive control unit 32, the image data acquisition unit (X-ray detector controller) 34, and the scanning X-ray source control mechanism 60, and is a flowchart for inspection processing as will be described later. Run the indicated program. Further, the operation of the inspection apparatus can be controlled by the input from the input unit 40, and the state of each unit or the inspection result can be output from the output unit 50.

  The inspection target position driving mechanism 110 includes a pair of rails 112 that fix the inspection target 20 and a displacement meter 114. Each rail 112 sandwiches the left end or right end of the inspection target 20 from above and below, and fixes the substrate. Here, left and right are defined along the Y axis, and upper and lower are defined along the Z axis. The rail 112 can move the inspection target in the horizontal (XY) and vertical (Z) directions according to a command from the inspection target position control unit 80. In the present embodiment, the inspection object position driving mechanism 110 conveys the inspection object 20 along a conveyance surface that is substantially parallel to the XY plane (horizontal plane), that is, the normal direction is substantially the Z direction. And Here, “the normal direction is substantially the Z direction” means that the conveyance surface may be inclined from the horizontal plane within a range that does not hinder the inspection (for example, 0 to 5 degrees).

  The displacement meter 114 measures the warp of the inspection target 20. In particular, mounting a component on a printed circuit board includes a heat treatment step such as reflow, so that the printed circuit board is generally deformed by heat and warps. In the case of a normal appearance inspection, the warpage of the substrate can be reduced by a support bar called a backup pin. However, the backup pin cannot be used in the X-ray inspection because the backup pin appears in the image. Therefore, when measuring a plurality of inspection objects 20, the X-ray inspection apparatus 100 corrects the inspection region in the Z direction based on the measurement result of the displacement meter 114.

  The inspection target position driving mechanism 110 does not cover the upper and lower surfaces of the inspection target 20. Therefore, X-rays can pass through the inspection object 20. However, the configuration of the inspection target position driving mechanism 110 is not limited to the above. For example, an XY stage that transmits X-rays may be used as the inspection target position driving mechanism 110. However, in order to inspect many inspection objects 20, the illustrated inspection object position drive mechanism 110 is suitable.

  The inspection target 20 is assumed to be a planar object. Further, it is assumed that the surface of the inspection object 20 is arranged on the plane in which the surface of the inspection object 20 is substantially parallel to the XY plane, that is, the normal direction is substantially the Z direction. Therefore, the major axis direction of the inspection object 20 is substantially orthogonal to the Z direction. Here, “the major axis direction is substantially perpendicular to the Z direction” means that the angle between the major axis direction and the Z direction may be different from 90 degrees within a range that does not hinder the inspection (for example, , May be 85 to 95 degrees).

  The X-ray detector drive unit 22 includes an orthogonal type biaxial robot arm 22.1 and a detector support unit 22.2 having a rotation axis, and is supplied from the calculation unit 70 through the detector drive control unit 32. The X-ray detector 23 is moved to a designated position by an instruction. Further, the detector drive control unit 32 sends the position information of the X-ray detector 23 at that time to the calculation unit 70.

  The computing unit 70 acquires an X-ray fluoroscopic image and transfers imaging data at a timing instructed by an instruction passed through the detector drive control unit 32.

  The scanning X-ray source 10 generates an electron beam in accordance with a command from the calculation unit 70 through the scanning X-ray source control mechanism 60 and is designated on the target by the electron beam converging coil 13 and the deflection yoke 12. The electron beam is converged to the position, and the X-ray focal position 17 is moved at high speed.

  The scanning X-ray source 10 outputs X-rays so that X-rays enter the inspection object 20 from a plurality of directions. The incident angle of the X-rays on the surface of the inspection object 20 is within an angle limited by the X-ray scanning range and the X-ray detection position. However, the X-ray incident direction of the inspection object 20 does not include the major axis direction of the inspection object 20 because of the positional relationship between the inspection object 20 and the scanning X-ray source 10.

(2. Reconstruction method)
Here, a general description of X-ray data reconstruction will be given. X-ray data reconstruction is the X-ray absorption coefficient inside the inspection object by measuring how much X-rays irradiated from the outside of the inspection object are absorbed (attenuated) by the inspection object from a plurality of angles. This is a technique for obtaining the distribution of.

  In the following description, it is assumed that measurement is performed using a so-called scanning X-ray source as the X-ray source.

  FIG. 4 is a diagram for explaining the data reconstruction method. Referring to FIG. 4, X-rays emitted from an X-ray focal point Fa corresponding to the X-ray detector Da pass through an inspection target (not shown) and reach a pixel Pa of the X-ray detector Da. As the X-rays pass through the inspection object, the X-ray dose (X-ray intensity) is attenuated by an amount corresponding to the inherent X-ray absorption coefficient of each of the components constituting the inspection object. The attenuation amount of the X-ray intensity is recorded as the pixel value of the detector pixel Pa.

The X-ray intensity emitted from the X-ray focal point Fa is I, the path through which X-rays pass from the X-ray focal point Fa to the detector pixel Pa is t, and the X-ray absorption coefficient distribution in the inspection object is f (x, y, z)
Then, the intensity Ia of the X-ray that has reached the detector pixel Pa is expressed by the following equation (1).

  Taking the logarithm of both sides of this equation, the X-ray absorption coefficient distribution along the path t is represented by a line integral value as in the following equation (2). A value obtained by measuring this X-ray absorption coefficient distribution with an X-ray detector is referred to as projection data. That is, the X-ray detector detects an X-ray attenuation distribution (which may be replaced with an X-ray intensity distribution).

  An algorithm for generating three-dimensional data of an object from projection data in CT is roughly classified into two methods, an analytical method (analysis method) and an iterative method (iteration method).

  The analysis method is to obtain three-dimensional data by superimposing projection images. The analysis method includes a Fourier transform method, an FBP method (Filtered Back-Projection method, filter-corrected back projection method), and the like.

  The iterative method compares the actual projected image with the projected image of the reconstructed data. If they do not match, repeat the procedure of correcting the reconstructed data many times, and reconstruct the object closer to the object. Obtain configuration data. Examples of iterative methods include algebraic methods and statistical methods. The iterative method is classified into a simultaneous iterative method in which reconstruction data is updated with reference to projection images in all directions and a sequential iteration method in which reconstruction data is updated with reference to projection images in only one direction. The

Hereinafter, each of the analysis method and the iterative method will be described in detail.
(Explanation of analytical method)
As shown in FIG. 4, in using the analytical method, X is arranged at a position different from the arrangement of the X-ray detector Da with respect to one inspection object (or one part of the inspection object). Projection data for the X-ray intensity Ib emitted from the focal point Fb and reaching the line detector Db is detected. By actually detecting such projection data for a plurality of arrangements with respect to one inspection object (or one part of the inspection object), the three-dimensional of the inspection object is obtained from these projection data. Data will be reconstructed.

  FIG. 5 is a top view of the arrangement of the reconstructed pixel V, the X-ray focal points Fa and Fb, and the X-ray detectors Da and Db of the field of view FOV and the field of view FOV of the inspection object shown in FIG. It is the figure seen from. X-rays that have passed through the reconstructed pixel V form (distance from the focal point to the reconstructed pixel V) pair (distance from the focal point to the X-ray detector) when forming an image on the X-ray detectors Da and Db. An enlarged image is formed according to the ratio of.

  Feldkamp et al. Proposed a reconstruction algorithm for performing three-dimensional data reconstruction based on the equation (2). This algorithm (the so-called Feldkamp method) A. Feldkamp, L.M. C. Davis and J.M. W. Kress, “Practical cone-beam algorithm”, Journal of the Optical Society of Famerica. A, 612-619 (1984), which is a known technique and will not be described in detail here. Hereinafter, the FBP method (Filtered Back-Projection method), which is one of general methods, will be briefly described.

  An operation for adding the projection data along the path t through which the X-rays pass from the projection data to obtain the X-ray absorption coefficient distribution f (x, y, z) is called back projection. However, if projection data is simply added, blurring occurs due to the point spread function of the imaging system, and thus the projection data is filtered. As this filter, for example, a high-frequency emphasis filter such as a Shepp-Logan filter is used. The direction of filtering is preferably perpendicular to the direction of the X-ray transmission path, but the Feldkamp method performs filtering by approximating the transmission path direction of the projection data to be all the same. Images can be reconstructed.

  The procedure for data reconstruction in this embodiment will be described below. First, the value pa ′ obtained by filtering the projection data pa of the detector pixel Pa of the X-ray detector Da is added to the pixel value v of the reconstructed pixel V. Further, the value pb ′ obtained by filtering the projection data pb of the detector pixel Pb of the X-ray detector Db is added to the pixel value v of the reconstructed pixel V. Then, v = pa ′ + pb ′. By performing this back projection operation on all X-ray detectors or a part of X-ray detectors, the final pixel value v of the reconstructed pixel V is expressed according to the following equation (3).

  By performing this operation on all the reconstruction pixels V in the reconstruction area (field of view) FOV, the X-ray absorption coefficient distribution to be inspected is obtained and reconstruction data is obtained.

FIG. 6 is a flowchart showing the processing procedure of such a filter-corrected back projection method.
Referring to FIG. 6, when processing by an analytical method is started (step S5002), first, projection data to be processed is selected from a plurality of captured projection data (step S5004). Next, a process for filtering the selected projection data is performed (step S5006).

  Further, an unprocessed reconstruction pixel V in the reconstruction field FOV is selected (step S5008), and a detector pixel corresponding to the reconstruction pixel V is obtained (step S5010).

  Subsequently, the filtered pixel value is added to the reconstructed pixel V (step S5012), and it is determined whether all reconstructed pixels have been added (step S5014). If the process has not been completed for all the reconstructed pixels, the process returns to step S5008. If the process has been completed, the process proceeds to step S5016.

  In step S5016, it is determined whether all projection data have been processed. If all the projection data has not been completed, the process returns to step S5004. On the other hand, if all projection data have been completed, reconstruction data generation ends (step S5018).

(Description of iterative method)
In the iterative method, the X-ray absorption coefficient distribution f (x, y, z) to be examined and the projection data ln (I / Ia) are reconstructed as an equation.

  FIG. 7 is a conceptual diagram showing the concept of processing in an iterative method when a scanning X-ray source is used. On the other hand, FIG. 8 is a top view of the conceptual diagram of FIG. 7 viewed from above.

  Hereinafter, a procedure for reconfiguration by an iterative method will be described with reference to FIGS. 7 and 8. A vector ν in which the pixel values of the reconstructed data are arranged in a line (at the head to represent the vector: → attached to the head: hereinafter referred to as “ν”) and a vector p in which the projection data are arranged in a line (vector) Is represented by the following formula (4) and formula (5).

  In the following, for example, the pixel for the image calculated when the X-ray from the X-ray focal point Fa is connected to the X-ray detector Da when the value of the reconstructed pixel V is assumed to be a certain value is referred to as the intermediate projection pixel Qa. A pixel actually observed on the X-ray detector Da is called a detector pixel Pa. The X-ray detector Db is also referred to as an intermediate projection pixel Qb and a detector pixel Pb, respectively.

  In an iterative approach, for the assumed reconstructed pixel vector ν and the corresponding intermediate projection data vector q, the intermediate projection data vector q is the actual measured detector pixel value, as described below. The solution ν is obtained by an iterative operation that updates the assumed vector ν until Pa or Pb can be regarded as coincident with the projection data.

  Here, J is the number of pixels in the reconstruction area (field of view), and I is the number of pixels in the projection data. T represents transposition. A projection operation relating ν and p is represented by an I × J coefficient matrix of the following equation (6).

  At this time, the data reconstruction by the iterative method can be formulated as a problem of obtaining ν by solving the following linear equation (7).

  In other words, the contribution of vj to pi is wij. W represents how much the pixel value ν of the reconstruction data contributes to the pixel value p of the projection data, and can be obtained from the geometric position of the X-ray focal point and the X-ray detector, Sometimes called detection probability or weight.

  As the iterative method, a method of algebraically solving an equation, a method considering statistical noise, and the like have been devised. ). For details, see A. H. Anderson and A.M. C. Kak, “SIMULTANEOUS ALGEBRAIC RECONSTRUCTIONTECHNIQUE (SART): A SUPERRIOR IMPLEMENTATION OF THE ALGORITHM”, ULTRASONIC IMAGEING 6, 81-94 (1984).

In SART, first, initial data ν ⁰ represented by the following equation (8) is assumed (indicated by “ν ⁰ ” in the text body).

The initial data ν ⁰ may be all zero data, or data obtained from CAD (Computer Aided Design) data or the like may be assumed.

Next, using the projection operation W, intermediate projection data q ⁰ represented by the following formula (9) (attached to the head to represent a vector: hereinafter referred to as “q ⁰ ” in the text body): Generate.

The generation of the intermediate projection data q ⁰ may be performed for one projection data or a plurality of projection data. In the following description, it is assumed that the process is performed on one piece of projection data.

The generated intermediate projection data q ⁰ is compared with the projection data p acquired from the X-ray detector. The comparison method includes a method of taking a difference and a method of dividing, but in SART, a difference (p−q ⁰ ) is taken.

Update initial data ν ⁰ . The formula used for updating (iteration formula) is as shown in formula (10).

  Note that the following calculation formulas (11) and (12) in the formula (10) can be calculated in advance to reduce the update calculation time.

  The reconstruction data generated by the above calculation is substituted as initial data, and the same processing is repeated a plurality of times to obtain reconstruction data.

FIG. 9 is a flowchart for explaining the process of the iterative method.
Referring to FIG. 9, first, when processing by an iterative method is started (step S5102), initial data is set (step S5104). As described above, the initial data may be all zero values, for example. Next, projection data to be processed is selected from a plurality of projection data corresponding to a plurality of X-ray detector positions (step S5106).

  Intermediate projection data is generated. The method for generating the intermediate projection data is as described above. (Step S5108).

  Further, an unprocessed reconstruction pixel V in the reconstruction field of view FOV is selected (step S5110), and a detector pixel corresponding to the reconstruction pixel V is obtained (step S5112).

Based on the iterative formula, the value of the reconstructed pixel V is updated (step S5114).
Next, it is determined whether or not all the reconstructed pixels have been updated (step S5116). If the process has not been completed for all the reconstructed pixels, the process returns to step S5110. The process moves to step S5118.

  In step S5118, it is determined whether all projection data have been processed. If all the projection data has not been completed, the process returns to step S5106. On the other hand, if the processing has been completed for all projection data, the process proceeds to step S5120.

  In step S5120, it is determined whether the process has been performed the specified number of iterations. If the process has not been repeated, the process returns to step S5104, the current reconstructed pixel value is adopted as initial data, and the process is repeated to define the process. If it has been repeated the number of times, the reconstruction data generation ends (step S5122).

  As described above, a three-dimensional image of the inspection object can be reconstructed from the projection data acquired by the X-ray detector.

  However, in the analytical method, in order to obtain each of the plurality of projection data for reasons such as the ease of calculation when the filtering process is performed on each pixel of the X-ray detector, the X-ray detector and the focus are acquired. Even when the relative position between the object and the object is changed, it is desirable that the relative arrangement of the X-ray focal point and the X-ray detector maintain a certain relationship. In other words, when the X-ray detector is viewed from the focal point, even if the angle of the portion included in the field of view of the target object within the solid angle to be seen, the position in the target object, and the like change, It is desirable that the positional relationship of be kept constant. Moreover, when performing the back projection method as described above, in order to reduce artifacts and the like, it is desirable that a plurality of projection data be acquired for each equiangular portion of the portion included in the field of view of the object.

  In contrast, in the iterative method, there is no such restriction on the relative arrangement of the X-ray focus and the X-ray detector.

(Characteristics of analysis method and iteration method)
In general, analysis methods and iterative methods have the following characteristics. The analysis method requires a large number of projection images in order to obtain accurate reconstruction data, but the calculation time for reconstruction is short. On the other hand, the iterative method can reconstruct highly accurate three-dimensional data from a small number of projection images, but requires a long calculation time.

Furthermore, from the viewpoint of PCB inspection, the analysis method and the iterative method have the following characteristics.
In the analysis method, the projection image is filtered and back-projected, so that edges are emphasized in a two-dimensional cross-sectional image (referred to as a reconstructed image) obtained by cutting out reconstruction data obtained by the analysis method. Therefore, the analysis method is suitable for an inspection in which the absorption coefficient is different, for example, a void and a foreign matter contamination inspection. On the other hand, in the analysis method, due to oblique CT imaging, a reconstructed image (hereinafter referred to as a vertical cross-sectional image) corresponding to a surface having the major axis (horizontal) direction of the PCB substantially in the normal direction is unclear. There is a problem of becoming. In order to prevent this, it is necessary to increase the number of projections (for example, 32 sheets). Therefore, X-ray CT inspection of a thin object using an analysis method requires a lot of imaging time, and as a result, takes time.

  On the other hand, using reconstructed data by an iterative method, unlike the analysis method, a high-quality reconstructed image can be generated regardless of the imaging position. Therefore, even when there is a shortage of imaging data (vertical cross-section data) from the long axis direction, a high-quality image can be generated. Therefore, the iterative method is suitable for wet inspection and the like that require a vertical cross-sectional shape. However, since the calculation amount of the iterative calculation is large, X-ray CT inspection of a thin object using the iterative method takes a lot of calculation time. In addition, the iterative method has a problem that it is not suitable for void inspection. This is because the edges of the reconstructed image obtained by iterative calculation are blurred.

  The characteristics of the analysis method and the iterative method will be summarized with reference to FIG. FIG. 10 is a diagram for explaining the characteristics of the analysis method and the iterative method. The horizontal axis in FIG. 10 indicates the number of projections, that is, the number of projection images. However, the number of projections shown in FIG. 10 is a guide, and the number of projections is not limited to this.

  In the region 1 shown in FIG. 10, an accurate inspection can be performed by using any of the analysis method and the iterative method. That is, when the number of projections is large, either the analysis method or the iterative method may be used. Moreover, when performing the foreign substance bridge inspection which is one of the horizontal cross-section inspections, either the analysis method or the iterative method may be used regardless of the number of projections. However, in consideration of the calculation time, it is preferable to use the analysis method in the region 1.

  In the region 2 shown in FIG. 10, it is preferable to perform an inspection using an analysis method. That is, it is preferable to use an analysis method when performing a void inspection which is one of the horizontal cross-sectional inspections. This is because, as described above, according to the analysis method, a reconstructed image with enhanced edges is obtained.

  In region 3 shown in FIG. 10, it is preferable to perform an inspection using an iterative method. That is, it is preferable to use an iterative method when performing a wetness inspection which is a vertical cross-sectional inspection. This is because, as described above, the reconstructed image of the vertical section obtained by the analysis method when the number of projections is small is unclear.

  As described above, which of the analysis method and the iterative method is preferably used depends on the inspection type and the number of projections.

  Specific examples of reconstructed images obtained by the analysis method and the iterative method are shown in FIG. FIG. 11A is a diagram illustrating a specific example of a reconstructed image obtained by an analysis method, and FIG. 11B is a diagram illustrating a specific example of a reconstructed image obtained by an iterative method.

  Referring to FIG. 11A, reconstructed images 5a to 5i are horizontal cross-sectional images reconstructed from a plurality of projection images of a spherical object by a computer. The reconstructed images 5a to 5c were calculated based on the eight projection images by the analysis method, the sequential iteration method, and the simultaneous iteration method, respectively. The reconstructed images 5d to 5f were calculated based on 16 projection images by an analysis method, a sequential iteration method, and a simultaneous iteration method, respectively. The reconstructed images 5e to 5i were calculated based on the 32 projection images by an analysis method, a sequential iteration method, and a simultaneous iteration method, respectively. FIG. 11A also shows a projected image 4 of a horizontal section by a computer for reference.

  Referring to FIG. 11B, reconstructed images 7a to 7i are vertical cross-sectional images reconstructed from a plurality of projection images of a spherical object by a computer. The reconstructed images 7a to 7c were calculated based on the eight projected images by the analysis method, the sequential iteration method, and the simultaneous iteration method, respectively. The reconstructed images 7d to 7f were calculated based on 16 projection images by an analysis method, a sequential iteration method, and a simultaneous iteration method, respectively. The reconstructed images 7e to 7i were calculated based on the 32 projection images by the analysis method, the sequential iteration method, and the simultaneous iteration method, respectively. FIG. 11B also shows a projected image 6 of a horizontal section by a computer for reference. Further, in the reconstructed images 7a to 7i, portions corresponding to the outer periphery of the object in the projection image 6 of the horizontal section are indicated by dotted lines.

  Referring to FIGS. 11A and 11B, the resolution of the reconstructed image obtained by the analysis method when the object is viewed from the vertical direction is high, but the reconstructed image obtained by the analysis method when the object is viewed from the horizontal direction is extended. I understand. On the other hand, in the reconstructed image by the iterative method, a spherical object is also reproduced in the horizontal direction.

  As described above, the method using either the analysis method or the iterative method cannot accurately inspect an object in a short time. In particular, with this method, both the horizontal inspection and the vertical inspection cannot be accurately performed in a short time. As will be described in detail below, X-ray inspection apparatus 100 according to the present embodiment can inspect an object with high accuracy in a short time by combining an analysis method and an iterative method.

(3. Outline of X-ray inspection)
With reference to FIG. 12, the outline of the X-ray inspection by the X-ray inspection apparatus 100 is demonstrated. FIG. 12 is a diagram showing an outline of the X-ray inspection by the X-ray inspection apparatus 100 in the form of a flowchart.

  Referring to FIG. 12, first, when the process is started (step S1200), the inspection target position control mechanism 110, in accordance with a command from the inspection target position control unit 80 of the arithmetic unit 70, causes the inspection target (field of view) to be inspected. ) Is moved to a position where it can be imaged (step S1202). That is, in order to capture a fluoroscopic image, the stage on which the inspection object is placed and the X-ray detector are moved to a predetermined position. The movement of the position of each component is usually determined in advance and can be read with a magnescale or the like, but a laser displacement meter or the like may be used. In general, in an inspection, an optical camera (not shown) is mounted for specifying an inspection position, so that the position can be determined based on an image of the optical camera. As another method, it may be automatically determined based on CAD data to be inspected, or may be performed visually by an operator.

  In addition, first, a fluoroscopic image is captured (step S1204). Specifically, the X-ray inspection apparatus 100 starts exposure of the X-ray detector by moving the X-ray focal point to a predetermined position with the start of X-ray irradiation. The exposure is terminated after a certain time, and the obtained image is transferred to the inspection apparatus control mechanism.

  The quality determination unit 78 of the calculation unit 70 inspects the fluoroscopic image and determines the quality of the visual field to be inspected (the range captured by the fluoroscopic image) from the acquired fluoroscopic image (step S1206). Various methods have been proposed as the pass / fail judgment method, and since they are publicly known, details are not described here. For example, as the most basic inspection, the fluoroscopic image is binarized with a constant value and compared with design information such as CAD data, and it is determined from the area whether there is a part at a predetermined position on the fluoroscopic image.

  Subsequently, the arithmetic unit 70 determines whether or not inspection by the reconstructed image is necessary (step S1208). Judgment criteria can be set in advance based on design information such as CAD data. In addition, the calculation unit 70 can also determine whether or not the inspection with the reconstructed image is necessary from the result of pass / fail determination of the fluoroscopic image. For example, in the inspection of the mounting board, when a component is mounted only on one side, it may not be necessary to perform pass / fail determination using a reconstructed image because it can be determined pass / fail with a fluoroscopic image.

  If the inspection using the reconstructed image is not necessary, the arithmetic unit 70 ends the inspection (S1218).

  On the other hand, when the inspection by the reconstructed image is necessary, the arithmetic unit 70 subsequently performs CT imaging for one field of view (S1210). Here, the “field of view” is a range that is reconstructed and displayed on the monitor. The X-ray inspection apparatus 100 sets an area similar to the fluoroscopic image imaging range in step S1204 and another area necessary for data reconstruction as a visual field. In CT imaging, the visual field within the examination object is imaged from a plurality of directions.

  Note that imaging from a plurality of directions is preferably realized by moving the focal position of the scanning X-ray source 10. This is because the moving time of the focal point is about 1/100 shorter than the time required for mechanically moving the X-ray detector 23 and the stage.

  Next, the reconstruction unit 76 of the calculation unit 70 generates reconstruction data by an analysis method and reconstruction data by an iterative method from captured images in a plurality of directions (step S1212). In step S1212, the reconstruction unit 76 generates a horizontal cross-sectional image using the reconstruction data obtained by the analysis method. The reconstruction unit 76 generates a vertical slice image using reconstruction data obtained by an iterative method. Details of the processing performed by the reconstruction unit 76 in step S1212 will be described later.

  Here, the vertical cross-sectional image is a reconstructed image corresponding to a vertical cross-section in which the major axis direction of the object is substantially the normal direction. Here, “the major axis direction is substantially the normal direction” means that the obtained vertical cross-sectional image can be used for the vertical cross-sectional inspection, and the normal direction of the vertical cross-section coincides with the major axis direction. It means that you do n’t have to.

Further, the horizontal cross-sectional image is a reconstructed image corresponding to a horizontal cross-section substantially parallel to the surface of the inspection target 20 with the direction substantially perpendicular to the major axis direction of the object as the normal direction. Here, “the direction substantially perpendicular to the major axis direction is the normal direction” and “substantially parallel to the surface of the inspection object 20” are ranges in which the obtained horizontal cross-sectional image can be used for the horizontal cross-sectional inspection. Thus, it means that the normal direction of the horizontal section does not have to coincide with the major axis direction.

  Therefore, according to the X-ray inspection apparatus 100 according to the present embodiment, an object can be inspected with high accuracy in a short time. Compared to the case where either the iterative method or the analysis method is used, a reconstructed image suitable for each of the vertical direction inspection and the horizontal direction inspection can be obtained based on a small number of pieces of imaging data.

  Subsequently, the quality determination unit 78 of the calculation unit 70 performs quality determination of the object based on the reconstructed image (vertical cross-sectional image and horizontal cross-sectional image) (step S1214). Since the quality determination method is well known, a quality determination method suitable for the inspection item may be used, and detailed description thereof will not be repeated here. For example, the quality determination unit 78 determines the quality of the mounting board based on the solder area in the binarized image.

  Further, the calculation unit 70 determines whether or not the inspection of the entire field of view has been completed (step S1216). If not completed, the calculation unit 70 controls the inspection target position control mechanism to determine the inspection part (field of view) to be inspected. Change (step S1220). Thereafter, the arithmetic unit 70 returns the process to step S1204. On the other hand, if the inspection has been completed for the entire visual field, the arithmetic unit 70 ends the inspection (step S1218). In the in-line inspection, since a range to be inspected (a collection of a plurality of visual fields) is determined, the arithmetic unit 70 can easily determine whether or not the inspection of all the visual fields has been completed. Alternatively, the operator may determine whether the inspection of the entire visual field is completed and give an instruction to the X-ray inspection apparatus 100.

  In FIG. 12, the inspection is performed using the fluoroscopic image and the reconstructed image, but it is also possible to perform only the inspection using the reconstructed image without performing the inspection using the fluoroscopic image. However, since the reconstruction process usually takes a relatively long time, the entire inspection time can be shortened by performing pass / fail judgment on the fluoroscopic image before the inspection using the reconstructed image.

(4. Data reconstruction processing according to the first embodiment)
The reconstruction unit 76 of the X-ray inspection apparatus 100 according to the first embodiment performs the reconstruction calculation by the analysis method and the reconstruction calculation by the iterative method in parallel in step S1212 (data reconstruction processing) described above. To do. Since two calculations are performed in parallel, the inspection time can be shortened compared to the case where two calculations are performed in series. It is because it is not necessary to wait for the end of one calculation and start the other calculation by performing calculations in parallel.

  The reconstruction unit 76 needs to be able to execute the calculation of the analysis method and the calculation of the iterative method in parallel. For example, a plurality of computers may be used as the reconstruction unit 76. However, in recent years, computers equipped with a plurality of CPUs have become widespread. One such computer may be used as the reconstruction unit 76.

  With reference to FIG. 13, the process of the reconstruction part 76 which concerns on 1st Embodiment is demonstrated. FIG. 13 is a diagram illustrating the processing of the reconstruction unit 76 according to the first embodiment in a flowchart format.

  As illustrated in FIG. 13, the reconstruction unit 76 performs a reconstruction process based on an analysis method (steps S1301 to S1313) and a reconstruction process based on an iterative method (steps S1315 to S1329) in parallel.

First, the reconstruction process by the analysis method will be described.
In step S1301, the reconstruction unit 76 selects one projection data (or set of projection data) to be used for reconstruction. Normally, the reconstruction unit 76 selects the projection data in order based on the number previously assigned to the projection data. The reconstruction unit 76 sequentially selects projection data by incrementing the number to be selected.

  In step S1303, the reconstruction unit 76 performs filtering processing on the selected projection data using a high frequency enhancement filter.

  In step S1305, the reconstruction unit 76 selects one reconstruction pixel to be backprojected. Usually, the reconstruction unit 76 selects the reconstruction pixels in order based on the numbers assigned in advance to the reconstruction pixels. The reconstruction unit 76 sequentially selects reconstruction pixels by incrementing the number to be selected.

  In step S1307, the reconstruction unit 76 calculates detector pixels that are backprojected to the selected reconstruction pixel. That is, the reconstruction unit 76 calculates which sensor pixel of the detector the X-ray passing through the reconstruction pixel has entered. The X-ray path can be geometrically calculated from the X-ray focal point position and the reconstructed pixel position, and the detector position is known. Therefore, the reconstruction unit 76 can calculate which pixel of the detector the X-ray is incident on.

  In step S1309, the reconstruction unit 76 adds the pixel value of the sensor pixel corresponding to the reconstruction pixel calculated in step S1307 to the pixel value of the reconstruction pixel. At this time, the value to be added is the value after filtering in step S1303.

  Note that there may be a plurality of pixels on which X-rays enter due to sampling problems. In that case, linear interpolation at four points around the detector pixel may be used. As can be easily imagined, there is a trade-off between interpolation accuracy and calculation time.

  In step S1311, the reconstruction unit 76 determines whether the processing from step S1305 to step S1309 has been performed on all the reconstruction pixels. For example, the reconstruction unit 76 increments the number of reconstructed reprojected pixels to determine whether or not all the reconstructed pixels have been processed, and whether or not the predetermined total number of reconstructed pixels has been exceeded. Judgment by.

  If the process has not been completed for all the reconstructed pixels (No in step S1311), the process returns to step S1305. If the process has been completed (Yes in step S1311), the process proceeds to step S1313.

  In step S1313, the reconstruction unit 76 determines whether or not the processing from step S1301 to step S1311 has been performed for all projection data. For example, the reconstruction unit 76 determines whether or not it has been performed for all the projection data by incrementing the number of selected projection data and determining whether or not the total number of captured projection data has been exceeded.

  If the processing has not been completed for all projection data (No in step S1313), the processing returns to step S1301, and if completed (Yes in step S1313), the reconstruction unit 76 performs data reconstruction by an analysis method. End the process.

  In step S1314, the reconstruction unit 76 generates a vertical cross-sectional image by cutting out a vertical cross-section of the reconstructed data obtained by the analysis method.

Subsequently, reconstruction processing by an iterative method will be described.
In step S1315, the reconstruction unit 76 sets initial data used for reconstruction. The initial data in the first iteration may be all zero data, or data obtained from CAD data or the like may be assumed.

  For the initial data in the second and subsequent iterations, the reconstruction unit 76 sets the immediately previous reconstruction data as the initial data. That is, the initial data of the second iteration is the reconstructed data obtained by the first iteration, and the initial data of the nth iteration is the reconstruction data obtained by the (n-1) th iteration. It becomes configuration data.

  In step S1317, the reconstruction unit 76 selects one projection data (or set of projection data) to be used for reconstruction. Normally, the reconstruction unit 76 selects the projection data in order based on the number previously assigned to the projection data. The reconstruction unit 76 sequentially selects projection data by incrementing the number to be selected.

  In step S <b> 1319, the reconstruction unit 76 generates intermediate projection data using the projection calculation W. The reconstruction unit 76 may generate intermediate projection data for one projection data or a plurality of projection data. In the following description, it is assumed that the reconstruction unit 76 has generated intermediate projection data for one projection data.

  In step S1321, the reconstruction unit 76 selects one reconstruction pixel to be backprojected. Usually, the reconstruction unit 76 selects the reconstructed pixels in order based on the numbers previously assigned to the reconstructed pixels. The reconstruction unit 76 sequentially selects reconstruction pixels by incrementing the number to be selected.

  In step S1323, the reconstruction unit 76 updates the initial data. As an equation used for updating (an iterative equation), equation (10) is used.

  In step S1325, the reconstruction unit 76 determines whether or not the processing in steps S721 and S723 has been performed on all reconstruction pixels. For example, the reconstruction unit 76 increments the number of reconstructed reprojected pixels to determine whether or not all the reconstructed pixels have been processed, and whether or not the predetermined total number of reconstructed pixels has been exceeded. Judgment by.

  If the process has not been completed for all the reconstructed pixels (No in step S1325), the process returns to step S1321, and if completed (Yes in step S1325), the process proceeds to step S1327.

  In step S1327, the reconstruction unit 76 determines whether or not the processing from step S1315 to step S1325 has been performed for all projection data. For example, the reconstruction unit 76 determines whether or not it has been performed for all the projection data by incrementing the number of selected projection data and determining whether or not the total number of captured projection data has been exceeded.

  If the processing has not been completed for all projection data (No in step S1327), the process returns to step S1317. If the processing has been completed (Yes in step S1327), the process proceeds to step S1329.

  In step S1329, the reconstruction unit 76 determines whether or not a predetermined number of iterations has been performed. For example, the reconfiguration unit 76 increments the number of iterations to determine whether or not a prescribed number of iterations has been performed, and determines whether or not the prescribed number of iterations has been exceeded (for example, 10 times). If the prescribed number of iterations is not satisfied (No in step S1329), the process returns to step S1315. That is, the reconstruction unit 76 obtains reconstruction data by repeating the same process a plurality of times using the reconstruction data generated by the above calculation as initial data. If the prescribed number of iterations is satisfied (Yes in step S1329), the reconstruction unit 76 ends the data reconstruction process by the iterative method.

  In step S1330, the reconstruction unit 76 cuts out a vertical section of the reconstruction data by the iterative method and generates a vertical section image.

  As described above, according to the X-ray inspection apparatus 100 according to the present embodiment, a reconstructed image by the analysis method and a reconstructed image by the iterative method are obtained. The X-ray inspection apparatus 100 can perform horizontal section inspection such as void inspection using the reconstructed image by the analysis method and vertical section inspection such as wetness inspection using the reconstructed image by the iterative method. The number of projection data to be prepared for inspection may be smaller than that in the case of calculating a reconstructed image by either the analysis method or the iterative method. Therefore, according to the X-ray inspection apparatus 100 according to the present embodiment, the inspection time can be shortened. Moreover, since the X-ray inspection apparatus 100 performs the calculation of the analysis method and the calculation of the iterative method in parallel, the inspection time can be further shortened.

  The inspection time excluding the CT imaging time by the X-ray inspection apparatus 100 according to the first embodiment is estimated. However, the reconstruction time by the analysis method is Ta, the reconstruction time of one iteration by the iterative method is Ti, the number of iterations is n, the horizontal section inspection time is Th, and the vertical section inspection time is Tv.

  The inspection time excluding the CT imaging time in the first embodiment depends on the longer of the analysis method inspection time Ta + Th and the iterative method inspection time Ti * n + Tv. Usually, the reconstruction time for one iteration in the iterative method is Ti> 2Ta and Th≈Tv, and therefore the inspection time of the iterative method is longer. Therefore, the inspection time excluding the CT imaging time of the first embodiment is Ti * n + Tv.

  FIG. 12 shows the flow of performing data reconstruction after CT imaging, but the X-ray inspection apparatus 100 preferably performs data reconstruction in parallel with CT imaging. When data reconstruction is performed in parallel with CT imaging, the data reconstruction time can be concealed by the CT imaging time, so that high-speed inspection is possible.

  The X-ray inspection apparatus 100 according to the first embodiment further increases the total inspection time including the CT imaging time by starting the calculation of the iterative method when the number of projection images necessary for the iterative method is captured. It can be shortened.

[Second Embodiment]
The X-ray inspection apparatus 100 according to the second embodiment performs reconstruction by an iterative method using reconstruction data of the analysis method as initial data after performing reconstruction by the analysis method. Unlike the first embodiment, the X-ray inspection apparatus 100 performs serial processing, that is, sequentially calculates the analysis method and the iterative method. However, since the reconstructed data of the analysis method is used as the initial data of the iterative method, the number of iterations required to obtain a highly accurate image can be reduced, so that the inspection time can be shortened as a whole. Further, it is not necessary to use a computer capable of executing calculations in parallel as the reconfiguration unit 76, and the inspection method according to the present embodiment can be executed even when there is one computer.

  The configuration of the X-ray inspection apparatus 100 according to the second embodiment is the same as that of the first embodiment described with reference to FIGS. 1 to 3, and the description thereof will not be repeated here. The general flow of the X-ray inspection is the same as that of the first embodiment described with reference to FIG. 12, and the description thereof will not be repeated here.

  The X-ray inspection apparatus 100 according to the second embodiment performs data reconstruction processing different from that of the first embodiment. Processing of the reconstruction unit 76 of the X-ray inspection apparatus 100 according to the second embodiment will be described with reference to FIG. FIG. 14 is a diagram illustrating the processing of the reconfiguration unit 76 according to the second embodiment in the form of a flowchart. The processing of the reconstruction unit 76 according to the second embodiment is roughly divided into the following (1) to (3).

  (1) The reconstruction unit 76 first performs reconstruction by an analysis method. That is, the reconstruction unit 76 performs each process from step S1401 to step S1413, and generates reconstruction data by an analysis method. Since the processing in steps S1401 to S1413 is the same as the processing in steps S701 to S713, description thereof will not be repeated here.

  (2) In step S1415, the reconstruction unit 76 sets the image reconstructed by the analysis method as initial data in the first iterative process of the iterative method. For the initial data in the second and subsequent iterations, the reconstruction unit 76 sets the immediately previous reconstruction data as the initial reconstruction data. That is, the initial data of the second iteration is the reconstructed data obtained by the first iteration, and the initial data of the nth iteration is the reconstruction data obtained by the (n-1) th iteration. It becomes configuration data.

  (3) The reconstruction unit 76 is the same as in the first embodiment except that the image reconstructed by the analysis method is set as initial data in the first iteration of the iteration method. Processing (steps S1415 to S1429) is executed. Since the processing in steps S1415 to S1429 is the same as the processing in steps S715 to S729, description thereof will not be repeated here.

  The inspection time excluding the CT imaging time by the X-ray inspection apparatus 100 according to the first embodiment is estimated. However, the reconstruction time by the analysis method is Ta, the reconstruction time of one iteration by the iterative method is Ti, the number of iterations is n / 2, the number of iterations is Th, the horizontal section inspection time is Th, and the vertical section inspection time is Tv. . Reconstruction data by the analysis method is set as initial data of the iterative method, and the number of iterations is half that of the first embodiment in view of the fact that the number of iterations may be smaller than that of the first embodiment.

  The inspection time excluding the CT imaging time in the second embodiment is the sum of the analysis method inspection time Ta + Th and the iterative method inspection time Ti * n / 2 + Tv. Therefore, the inspection time in the second embodiment is Ta + Th + Ti * n / 2 + Tv.

  The inspection time excluding the CT imaging time of the second embodiment is compared with the inspection time excluding the CT imaging time of the first embodiment.

  (Inspection time of the first embodiment) − (Inspection time of the second embodiment) = (Ti * n + Tv) − (Ta + Th + Ti * n / 2 + Tv) = Ti * n / 2−Ta−Th.

  Here, generally, the reconstruction time for one iteration in the iterative method is typically Ti> 2Ta. Further, considering the reconstruction time and the inspection time, generally Ta >> Th. Therefore, (inspection time of the first embodiment) − (inspection time of the second embodiment)> 0, that is, (inspection time of the second embodiment) <(of the first embodiment) Inspection time).

[Third Embodiment]
The X-ray inspection apparatus 100 according to the third embodiment sets the range to be reconstructed by the iterative method based on the result of the analysis method after performing the reconstruction by the analysis method, and the iterative method for the set range Reconfiguration by The point that the calculation of the analysis method and the calculation of the iterative method are performed in series is the same as that of the second embodiment, but the range of reconstruction by the iterative method is narrowed based on the result of the analysis method, so that iterative 1 The reconstruction time can be shortened, and the reconstruction time of the iterative method can be shortened. Further, it is not necessary to use a computer capable of executing calculations in parallel as the reconfiguration unit 76, and the inspection method according to the present embodiment can be executed even when there is one computer.

  The configuration of the X-ray inspection apparatus 100 according to the third embodiment is the same as that of the first embodiment described with reference to FIGS. 1 to 3, and the description thereof will not be repeated here. The general flow of the X-ray inspection is the same as that of the first embodiment described with reference to FIG. 12, and the description thereof will not be repeated here.

  The X-ray inspection apparatus 100 according to the third embodiment performs a data reconstruction process that is different from the first embodiment and the second embodiment. Processing of the reconstruction unit 76 of the X-ray inspection apparatus 100 according to the third embodiment will be described with reference to FIG. FIG. 15 is a diagram illustrating the processing of the reconstruction unit 76 according to the third embodiment in a flowchart format. The processing of the reconstruction unit 76 according to the third embodiment is roughly divided into the following (1) to (4).

  (1) The reconstruction unit 76 first performs reconstruction by an analysis method. That is, the reconstruction unit 76 performs each process from step S1501 to step S1513, and generates reconstruction data by an analysis method. The processes in steps S1501 to S1513 are the same as the processes in steps S701 to S713, respectively, and thus description thereof will not be repeated here. In step S1514, the reconstruction unit 76 creates a horizontal cross-sectional image from the reconstruction data obtained by the analysis method.

  (2) Next, in step S1515, the reconstruction unit 76 performs a horizontal section inspection based on the horizontal section image, and limits the range that needs to be reconstructed by an iterative method. Since a well-known method can be used for the horizontal section inspection method, detailed description will not be given here.

The reconstruction unit 76 limits the range in which reconstruction by the iterative method is necessary in the following cases.
(A) When there is no part in a certain range The range without a part is not required for inspection, so there is no need to reconfigure. Therefore, it is possible to limit the range of the iterative method by determining whether there is a part before performing the iterative method based on the horizontal cross-sectional image.

(B) When it is a defective product as a result of the horizontal cross-sectional inspection When it is determined as a defective product in the horizontal cross-sectional inspection, it can be determined as a defective product without performing another vertical inspection.

  (3) In step S1517, the reconstruction unit 76 sets the image reconstructed by the analysis method as initial data in the first iterative process of the iterative method. For the initial data in the second and subsequent iterations, the reconstruction unit 76 sets the immediately previous reconstruction data as the initial data. That is, the initial data of the second iteration is the reconstructed data obtained by the first iteration, and the initial data of the nth iteration is the reconstruction data obtained by the (n-1) th iteration. It becomes configuration data.

  Since the reconstruction data of the analysis method is used as the initial data, the number of iterations of the iteration method can be reduced as in the second embodiment.

  In the above, the example in which the reconstruction data of the analysis method is used as the initial data has been described. However, as in the first embodiment, a predetermined image can be used as the initial data. Also in this case, since the time required for one iteration can be shortened, the inspection time can be shortened. However, it is a matter of course that the inspection time can be further shortened by using the reconstructed data of the analysis method as initial data as in the above example.

  (4) Similar to the first embodiment, the reconstruction unit 76 sets the image reconstructed by the analysis method as initial data in the first iteration of the iterative method. Processing (step S1517 to step S1531) is executed. Since the processes in steps S1517 to S1531 are the same as the processes in steps S715 to S729, description thereof will not be repeated here. In step S1532, the reconstruction unit 76 creates a vertical cross-sectional image from reconstruction data obtained by an iterative method.

  The inspection time excluding the CT imaging time by the X-ray inspection apparatus 100 according to the third embodiment is estimated. However, the reconstruction time by the analysis method is Ta, the reconstruction time by one iteration by the iteration method is Ti, the number of iterations is n / 2 times, the horizontal section inspection time is Th, the vertical section inspection time is Tv, and the repetition of the iterative method The time required for limiting the configuration range is defined as Tl. The reconstruction data by the analysis method is set as the initial image of the iterative method, and the number of iterations is half that of the first embodiment in view of the fact that the number of iterations may be smaller than that of the first embodiment. It is also assumed that the reconstruction range by the iterative method is limited to half. More than half of the mounting board space is air or printed board that does not need to be reconfigured, and this assumption is reasonable.

  The inspection time excluding the CT imaging time of the third embodiment includes the inspection time Ta + Th of the analysis method, the time Tl required for limiting the reconstruction range of the iterative method, and the inspection time Ti / 2 * n / 2 + Tv of the iterative method. Will be added. Therefore, the inspection time excluding the CT imaging time of the third embodiment is Ta + Th + Tl + Ti / 2 * n / 2 + Tv.

  The examination time excluding the CT imaging time of the third embodiment is compared with the examination time excluding the CT imaging time of the second embodiment.

  (Inspection time of the second embodiment) − (Inspection time of the third embodiment) = (Ta + Th + Ti * n / 2 + Tv) − (Ta + Th + Tl + Ti / 2 * n / 2 + Tv) = Ti * n / 4−Tl is there.

  Here, considering the reconstruction time and the inspection time, since generally Ti >> Ta >> Th≈Tv≈Tl, (inspection time of the third embodiment) <(in the second embodiment) Inspection time).

[Fourth Embodiment]
In the third embodiment, it has been described that the reconstruction area of the iterative method is limited from the reconstruction result of the analysis method. Here, in particular, an embodiment that limits the height direction of the reconstruction area will be described as a modified example of the third embodiment.

  The configuration of the X-ray inspection apparatus 100 according to the fourth embodiment is the same as that of the third embodiment, and the description thereof will not be repeated here. The general flow of the X-ray inspection is the same as that described with reference to FIG. 12, and the description thereof will not be repeated here.

  As described with reference to FIG. 3, the X-ray inspection apparatus 100 includes a displacement meter 114 for measuring the warpage of the substrate and performing height correction. Here, the warping of the substrate will be described in detail with reference to FIG. FIG. 16 is a cross-sectional view of a printed circuit board on which a component (BGA) is mounted as viewed from the conveyance direction.

  Referring to FIG. 16, the printed circuit board is warped. The magnitude of this warpage varies depending on the printed circuit board. When inspecting a plurality of printed boards, the X-ray inspection apparatus 100 obtains the amount of warpage of the substrate using the displacement meter 114 before each inspection, and based on the obtained amount of warpage, the periphery of the component to be inspected By limiting the reconstruction range to the above, the reconstruction time can be shortened. However, the displacement meter 114 generally has a repeatability of only about 100 um. Therefore, the X-ray inspection apparatus 100 needs to limit the reconstruction range in consideration of this error.

  On the other hand, one pixel (voxel) size of the reconstruction data is several to several tens of um. For this reason, if the reconstruction range is set in consideration of the error in repeatability, the amount of calculation increases significantly. For example, consider inspecting a BGA having a bump diameter of 300 μm. In this case, because of the repeatability of the displacement meter 114, extra reconstruction data having a height of 100 × 2 = 200 μm must be created. Therefore, a calculation time of (300 + 100 × 2) /300=1.7 times is required as compared with the case where there is no repetition error.

  Therefore, the X-ray inspection apparatus 100 according to the present embodiment obtains pixel distributions in a plurality of slices orthogonal to the height (Z) direction (that is, parallel to the XY plane), and based on the pixel distribution, the height Limit the reconstruction range to the direction.

  FIG. 17 is a diagram showing the relationship between the field of view of the analysis method (data reconstruction range) and the field of view of the iterative method in the present embodiment. Referring to FIG. 17, the visual field of the analysis method is the size in the height direction of the inspection object so that the inspection object (here, the substrate 1601, the component 1603, and the BGA 1604) is included in the visual field even if the substrate is warped. It is set larger than On the other hand, the field of view of the iterative method is set to the minimum necessary area including the BGA 1604 to be inspected.

  An example of the tomographic image of the reconstructed three-dimensional data by the analysis method is shown in FIG. As shown in FIG. 18, the reconstructed image obtained by the analysis method extends in the height direction due to the influence of oblique CT imaging. However, it is possible to infer the position where the solder ball exists from the reconstructed data obtained by the analysis method.

  In the present embodiment, the reconstruction unit 76 of the X-ray inspection apparatus 100 estimates the position where the solder ball exists based on the reconstruction data obtained by the analysis method as follows, and performs the reconstruction using the iterative method. Set.

  First, the reconstruction unit 76 obtains a variance Vz of pixel values for a plurality of horizontal cross-sectional images of reconstruction data obtained by an analysis method. That is, the reconstruction unit 76 calculates the average of all the pixel values in each horizontal slice image, and calculates the sum of the squares of the difference between the average and the pixel value.

  Next, the reconstruction unit 76 calculates a position Zmid where the variance Vz is maximized. The reconstruction unit 76 compares all the obtained variances Vz and selects the largest value. Then, the reconstruction unit 76 sets the height of the image having the largest variance Vz as Zmid.

  The position (Zmid) where the variance Vz is maximized is considered to be near the center of the part. In the board to be inspected, the components are characterized by absorbing X-rays and being densely packed at the same height. This is because, in the horizontal tomographic image at the height where the component exists, the luminance value varies greatly, and the variance is considered to increase. Therefore, the reconstruction unit 76 searches for the upper end position and the lower end position of the part starting from Zmid, and obtains the height range of the field of view.

  The reconstruction unit 76 calculates a position Zlow where the variance Vz is smaller than the threshold value Vth at a Z position lower than Zmid. The reconstruction unit 76 compares the variances Vz and Vth in order from Zmid in the direction of decreasing Z, and sets the initial height at which Vz is smaller than Vth as Zlow. Here, Vth is set in advance by the user.

  Similarly, the reconstruction unit 76 calculates a position Zhigh where the variance Vz is smaller than the threshold value Vth at a Z position higher than Zmid. The reconstruction unit 76 compares the variances Vz and Vth in order from Zmid to Z, and sets the first height at which Vz is smaller than Vth as Zhigh.

  Then, the reconstruction unit 76 sets the height range of the region (field of view) to be reconstructed by the iterative method from Zlow to Zhigh.

  This setting method will be described with reference to FIG. FIG. 19 is a diagram illustrating dispersion of pixel values of a tomographic image of reconstruction data by the analysis method illustrated in FIG. The horizontal axis indicates the Z position of the tomographic image. The vertical axis represents the variance Vz.

  Referring to FIG. 19, the dispersion Vz has a peak corresponding to the solder ball. The reconstruction unit 76 obtains this position as Zmid. Further, the variance Vz gradually decreases as Zmid moves away. Here, Vth is set to 2000. The reconstruction unit 76 obtains Z at which Vz = 2000 as Zhigh and Zlow.

  FIG. 18 also shows Zmid, Zhigh, and Zlow thus obtained. In this figure, the object appears to exist beyond Zhigh and Zlow. However, in oblique CT imaging, since the object extends in the height direction, the field of view to be reconstructed is sufficient in the obtained height range.

  Processing of the reconstruction unit 76 of the X-ray inspection apparatus 100 according to the fourth embodiment will be described with reference to FIG. FIG. 20 is a diagram illustrating the processing of the reconstruction unit 76 according to the fourth embodiment in a flowchart format.

  In step S2001, the reconstruction unit 76 reconstructs 3D data and creates a horizontal cross-sectional image by an analysis method. Specifically, the reconfiguration unit 76 executes the processing from step S1501 to step S1514 shown in FIG. Since these processes have already been described, details thereof will not be repeated.

  In step S2003, the reconstruction unit 76 obtains a variance Vz of pixel values in the horizontal cross-sectional image for one height position (Z position).

  In step S2005, the reconstruction unit 76 determines whether the variance Vz has been obtained for all Z positions.

  If it is determined that the variance Vz has not been obtained for all the Z positions (NO in step S2005), the reconstruction unit 76 changes the Z position and performs the process of step S2003 again.

  If it is determined that the variance Vz has been obtained for all the Z positions (YES in step S2005), the reconstruction unit 76 calculates a position Zmid where the variance Vz is maximum in step S2007.

  In step S2009, the reconstruction unit 76 compares the variances Vz and Vth in order from Zmid to decrease Z, and obtains the initial height at which Vz becomes smaller than Vth as Zlow.

  In step S2011, the reconstruction unit 76 compares the variances Vz and Vth in order from Zmid to Z, and sets the first height at which Vz is smaller than Vth as Zhigh.

  In step S2013, the reconstruction unit 76 sets the range of the visual field of the iterative method from Zlow to Zhigh.

  In step S2015, the reconstruction unit 76 reconstructs 3D data and creates a vertical cross-sectional image by an iterative method. Specifically, the reconstruction unit 76 executes the processing from step S1517 to step S1532 shown in FIG. Since these processes have already been described, details thereof will not be repeated.

  With the above processing, according to the X-ray inspection apparatus 100 according to the present embodiment, the field of view of the iterative method can be set to the minimum necessary in the height direction. Therefore, the reconstruction time of the iterative method can be shortened. For example, in the case of inspection of a BGA having a bump diameter of 300 μm, the calculation time of the iterative method can be increased by 1.7 times.

  In the above, the height at which the dispersion is maximum is assumed to be the height of the center of the solder ball, but the following methods can be considered as other methods for obtaining Zmid.

(1) Diameter of solder ball The reconstruction unit 76 has a height at which the diameter of the solder area corresponding to the solder ball (white or area having a value of 1 when binarized) is maximized in the cross-sectional image, It may be Zmid. This utilizes the property that the solder ball has a spherical shape and the diameter near the center is maximized.

(2) Area of Solder Ball The reconstruction unit 76 may set the height at which the area of the solder region is maximized to Zmid. This setting method is based on the same concept as the method of setting the solder ball diameter to Zmid. However, this method can avoid erroneous recognition due to the fact that the solder ball fault is not a perfect circle. In addition, this method has an advantage that since the area is the square of the length, it is less susceptible to noise than the comparison by length.

  In the above description, Zlow and Zhigh for determining the height of the field of view by setting a threshold value for dispersion are calculated. However, the following methods can be considered as other methods for determining Zlow and Zhigh.

(1) Use of a predetermined height range If the ball size is known in advance, it can be easily predicted how thick the mounted solder will be. Therefore, when the reconstruction unit 76 obtains the height of the center of the solder ball, the reconstruction unit 76 may set an area within a predetermined height range from the center height as the field of view of the iterative method. Good. According to this method, even if the height of the substrate is changed due to the warpage of the substrate, the field-of-view data having a certain height can be reconstructed by an iterative method following the height of the substrate.

(2) Utilizing the ratio from the maximum value The reconstruction unit 76 can also use a value obtained by multiplying the maximum value (dispersion, area, etc.) used to obtain the solder ball center by a predetermined ratio as a threshold value. It is. For example, the reconstruction unit 76 sets the position indicating the dispersion of 20% of the dispersion of the height of the ball center to Zlow and Zhigh. According to this method, the X-ray inspection apparatus 100 can accurately obtain the substrate height even if there is a change in dispersion due to noise.

[Others]
The embodiment disclosed this time should be considered as illustrative in all points and not restrictive. The scope of the present invention is defined by the terms of the claims, rather than the description above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims.

1 is a schematic block diagram of an X-ray inspection apparatus 100 according to the present invention. 2 is a cross-sectional view showing a configuration of a scanning X-ray source 10. FIG. It is a figure for demonstrating the structure of the X-ray inspection apparatus 100 of 1st Embodiment. It is a figure for demonstrating a data reconstruction method. FIG. 4 is a top view of the arrangement of the reconstruction pixel V, the X-ray focal points Fa and Fb, and the X-ray detectors Da and Db of the field of view FOV and the field of view FOV of the inspection object shown in FIG. It is. It is a flowchart which shows the process sequence of a filter correction | amendment back projection method. It is a conceptual diagram which shows the concept of the process by an iterative method at the time of using a scanning X-ray source. It is the top view which looked at the conceptual diagram of FIG. 7 from the upper surface. It is a flowchart for demonstrating the process of an iterative method. It is a figure for demonstrating the characteristic of an analysis method and an iterative method. It is a figure which shows the specific example of the reconstruction image obtained by each of an analysis method and an iterative method. It is the figure which showed the outline of the X-ray inspection by the X-ray inspection apparatus 100 in the flowchart format. It is the figure which showed the process of the reconstruction part 76 which concerns on 1st Embodiment in the flowchart format. It is the figure which showed the process of the reconstruction part 76 which concerns on 2nd Embodiment in the flowchart format. It is the figure which showed the process of the reconstruction part 76 which concerns on 3rd Embodiment in the flowchart format. It is sectional drawing which looked at the printed circuit board in which components (BGA) were mounted from the conveyance direction. It is a figure which shows the relationship between the visual field (data reconstruction range) of the analysis method, and the visual field of an iterative method in 4th Embodiment. It is a figure which shows an example of the tomographic image of the reconstruction three-dimensional data by an analysis method. It is a figure which shows dispersion | distribution of the pixel value of the tomographic image of the reconstruction data by an analysis method. It is the figure which showed the process of the reconstruction part 76 which concerns on 4th Embodiment in the flowchart format. It is a figure for demonstrating the test | inspection of PCB. It is a figure which shows typically the inspection apparatus of patent document 1. FIG. It is a schematic diagram of the inspection apparatus which moves a X-ray source and a X-ray detector, and test | inspects a printed circuit board.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 10 Scanning X-ray source, 11 Target, 12 Deflection yoke, 13 Electron beam converging coil, 14 High voltage power supply, 15 Vacuum pump, 16, 16a, 16b Electron beam, 17, 17a, 17b X-ray focal position, 18, 18a, 18b X-ray, 19 electron gun, 20 inspection object, 21 axis, 22 X-ray detector drive unit, 22.1, robot arm, 22.2 detector support unit, 23 X-ray detector, 30 image acquisition control mechanism, 32 detector drive control unit, 34 image data acquisition unit, 40 input unit, 50 output unit, 60 X-ray source control mechanism, 62 electron beam control unit, 70 calculation unit, 72 X-ray source control unit, 74 image acquisition control unit , 76 reconstruction unit, 78 pass / fail judgment unit, 80 inspection object position control unit, 82 X-ray focal position calculation unit, 84 imaging condition setting unit, 90 memory, 92 X-ray focal position information Information, 94 imaging condition information, 96 program, 98 image data, 100 X-ray inspection apparatus, 110 inspection object position drive mechanism, 1601 substrate, 1602 first part, 1603 second part, 1604 BGA, 1605 surface layer, 1606 State after heating, 1607 void, 1608 bridge.

Claims (10)

An X-ray inspection apparatus for inspecting an object using X-rays,
X-ray output means for outputting the X-ray so that the X-ray is incident on the object from a plurality of directions;
X-ray detection means for imaging the X-ray at a position where the X-ray that has entered the object from each direction and transmitted through the object reaches;
Reconstructing means for generating a plurality of reconstructed image data of the object based on the X-ray intensity distribution data imaged at each of the positions;
The reconstruction means includes
First reconstruction means for generating first reconstructed image data corresponding to a surface having a first direction as a normal direction;
The first reconstruction means uses an analysis method in the generation processing of the first reconstruction image data,
A second reconstructing means for generating second reconstructed image data of the object corresponding to a plane whose normal direction is a second direction different from the first direction;
The X-ray inspection apparatus, wherein the second reconstruction unit uses an iterative method in the generation process of the second reconstruction image data.   The X-ray examination apparatus according to claim 1, wherein the reconstruction unit performs generation of the first reconstruction image data and generation of the second reconstruction image data in parallel.   3. The X-ray examination apparatus according to claim 1, wherein the second configuration unit generates the second reconstructed image data using the first reconstructed image data as an initial image of an iterative method. The reconstruction means further includes a determination means for determining a region in which the second reconstruction means generates the second reconstructed image data based on the three-dimensional reconstruction data by the analysis method,
The X-ray inspection apparatus according to claim 1, wherein the second reconstruction unit generates the second reconstruction image data for the determined region.   5. The X-ray according to claim 1, wherein the X-ray output unit outputs the X-ray so that the X-ray is incident on the object from a direction other than the second direction. 6. Inspection device. A transport mechanism that transports the object along a transport surface whose normal direction is substantially the second direction;
The X-ray output means outputs the X-ray from the first side of the transport surface,
The X-ray inspection apparatus according to claim 1, wherein the X-ray detection unit images the X-ray on a second side different from the first side of the transport surface. The object is a planar object,
The second direction is substantially the normal direction of the surface of the object;
7. The X-ray output unit according to claim 1, wherein the X-ray output unit outputs the X-ray so that the X-ray is incident on the object from a direction intersecting the surface of the object at a predetermined angle. The X-ray inspection apparatus described. The object is a planar object,
The second direction is substantially the normal direction of the surface of the object;
The determination means is configured to generate a region for generating the second reconstructed image data based on a pixel distribution of a plurality of cross sections substantially parallel to the surface of the object of the reconstructed three-dimensional data by the analysis method. The X-ray inspection apparatus according to claim 4, wherein the second direction is limited. A transport mechanism that transports the object along a transport surface whose normal direction is substantially the second direction;
The X-ray inspection apparatus according to claim 8, wherein the transport mechanism includes a rail that sandwiches the object. An X-ray inspection method for inspecting an object using X-rays,
Outputting the X-ray so that the X-ray is incident on the object from a plurality of directions;
Imaging the X-rays at a position where the X-rays incident on the object from each direction and transmitted through the object reach;
Generating a plurality of reconstructed image data of an object based on the X-ray intensity distribution data imaged at each of the positions,
The step of generating the reconstructed image data includes the step of generating first reconstructed image data corresponding to a plane whose normal direction is the first direction,
The step of generating the first reconstructed image data includes a step of using an analysis method in the generation processing of the first reconstructed image data.
The step of generating the reconstructed image data further includes the step of generating second reconstructed image data of the object corresponding to a plane whose normal direction is a second direction different from the first direction. Including
The step of generating the second reconstructed image data includes the step of using an iterative method in the process of generating the second reconstructed image data.