JP2011095227A - Radiation inspection device of inspection object - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a radiation inspection device capable of acquiring a sectional image suitable for observation of an inspection object. <P>SOLUTION: An imaging part irradiates a plurality of spots of the inspection object including a substrate and an electronic component with a radiation from a plurality of directions, and images a plurality of radiation transmission images. A sectional image generation part 38 reconstitutes a plurality of sectional images of the inspection object based on the plurality of radiation transmission images. An inspection image specification part 74 acquires an inspection image based on the plurality of sectional images. An auxiliary image specification part 76 acquires an auxiliary image on which at least a part of the inspection object is reflected. A frequency analysis part 42 specifies a frequency component included in common in the inspection image and in the auxiliary image. A frequency component removal part 44 reduces an influence of the frequency component included in common from the inspection image. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to an inspection apparatus for an inspection object, and more particularly to an inspection apparatus for inspecting an inspection object using a transmission image of the inspection object obtained by irradiating radiation.

  Some substrates on which electronic components are mounted (hereinafter referred to as “substrates”) have BGA (Ball Grid Array) and LGA (Land Grid Array) mounted thereon. In these boards, the terminal, which is the electrical connection part of the component, is located between the board and the component, and it is difficult to observe the state of solder that joins the component and the substrate with a conventional visual inspection apparatus using a camera. It is. Therefore, the substrate is irradiated with X-rays from a plurality of different directions, the three-dimensional shape of the solder at the joint is reconstructed into an image based on the transmission image, and an inspection is performed using a cross-sectional image obtained by cutting an arbitrary cross-section from the image. A technique has been proposed (see Patent Document 1).

  In the technology as described above, it is desirable to install an inspection device in the production line. From the viewpoint of inspection efficiency, it is difficult to take time to capture a transmission image, and it is difficult to install a large-scale apparatus in the production line. For this reason, the direction in which X-rays can be irradiated is within a relatively narrow range with respect to the direction perpendicular to the substrate surface, and it is difficult to irradiate X-rays from the direction of the side surface of the substrate and capture a transmission image. When reconstructing the three-dimensional shape of the solder at the joint portion into an image, if a transmission image in which the X-ray irradiation angle is limited to a narrow range in a direction perpendicular to the substrate surface is used, the resolution in the thickness direction of the substrate is lowered. As a result, in the cross-sectional image of the front surface of the substrate on which the solder joint used for inspection is imaged, information unnecessary for the inspection, such as components and patterns on the back surface of the substrate, may be left in the image. In order to solve this problem, a method of extracting an image of a solder part by taking a difference between two transmission images in which a voltage applied to generate X-rays is changed (see Patent Document 2), By taking an X-ray transmission image of the board with the components soldered on the front and back surfaces, and taking the difference from the X-ray transmission image of the board with the components soldered only on the back surface, A method for extracting a transmission image (see Patent Document 3) has been proposed.

JP 2006-226875 A JP-A-10-253550 JP 2002-158500 A

  However, the above-described technology sets a specific target object as unnecessary information such as parts and patterns on the back surface, and it is necessary to capture an X-ray transmission image corresponding to the target object. Since unnecessary information from a plurality of objects remains in the image used for inspection, when trying to deal with all of them, it is necessary to take an X-ray transmission image according to each object, and the number of inspection steps increases. Will be.

  The present invention has been made in view of such circumstances, and an object of the present invention is to provide a radiation inspection apparatus capable of obtaining an image suitable for observation of an object to be inspected while suppressing an increase in inspection man-hours.

  One embodiment of the present invention relates to a radiation inspection apparatus. The apparatus irradiates a plurality of locations of an object to be inspected including a substrate and an electronic component from a plurality of directions, and captures a plurality of radiation transmission images, and the inspection based on the plurality of radiation transmission images. A cross-sectional image generating unit for reconstructing a plurality of cross-sectional images of the body, an inspection image specifying unit for acquiring an inspection image based on the plurality of cross-sectional images, and an auxiliary image on which at least a part of the inspection object is projected An auxiliary image specifying unit for acquiring a frequency component, a frequency analyzing unit for specifying a frequency component included in common in the inspection image and the auxiliary image, and a frequency component removing unit for reducing the influence of the frequency component included in common from the inspection image Including.

  Note that any combination of the above-described constituent elements and the expression of the present invention converted between a method, an apparatus, a system, a computer program, a data structure, a recording medium, and the like are also effective as an aspect of the present invention.

  ADVANTAGE OF THE INVENTION According to this invention, the radiation inspection apparatus which can obtain the image suitable for observation of a to-be-inspected object can be provided, suppressing the increase in an inspection man-hour.

It is a figure which shows typically the structure of the radiation inspection apparatus which concerns on embodiment of this invention. FIG. 2 illustrates each functional block processed by the PC of FIG. 1. It is the flowchart which showed the flow from the imaging of a transmission image until it stores the image which reduced the influence of the noise component. It is a figure explaining the principle which reduces the frequency component which concerns on Embodiment 1. FIG. 6 is a flowchart for explaining a flow of processing for reducing frequency components according to the first embodiment. It is a figure explaining the principle which reduces the frequency component which concerns on Embodiment 2. FIG. 10 is a flowchart illustrating a flow of processing for reducing frequency components according to the second embodiment. It is a figure which shows typically the attention image acquisition part 40 which concerns on Embodiment 3. FIG.

  Prior to the description of embodiments of the present invention (hereinafter referred to as “embodiments”), an outline will be described first. In the embodiment, attention is paid to an image including a substrate surface on which solder for joining the substrate and a component mounted on the substrate is projected (hereinafter referred to as “surface image”), and the image is displayed on the surface image. The noise component is specified and removed. There are four forms, Embodiment 1, Embodiment 2, Embodiment 3, and Embodiment 4, depending on the difference in the image that is mainly used to specify the noise component, and the outline of each will be described below.

  In the first embodiment, the noise component is specified using an image including the back surface of the substrate (hereinafter referred to as “back surface image”) and a cross-sectional image inside the substrate (hereinafter referred to as “internal image”). The second embodiment specifies a noise component from the surface image itself without using an image other than the surface image for specifying the noise component. In the third embodiment, a noise component is generated using a pseudo transmission image (hereinafter referred to as “pseudo transmission image”) generated by projecting a plurality of back images and internal images in a direction perpendicular to the substrate surface. Identify. In the fourth embodiment, a noise component is specified using an image captured by a single component before being mounted on a substrate.

  The “noise component” refers to an image component that is not necessary for the inspection existing in the surface image, and refers to an image including a predetermined frequency component such as a wiring pattern of a component mounted on a substrate. The “predetermined frequency” is a spatial frequency peculiar to a noise component, for example, a spatial frequency possessed by a wiring pattern or the like on the back surface of the substrate where the BGA is mounted. Usually, since the solder which joins a component and a BGA board is arrange | positioned periodically, they have a specific spatial frequency component. Moreover, since the mounting positions and patterns of components are different for each board, the board has a specific frequency. Hereinafter, a component mounting position, a wiring pattern, and a land for placing solder are also simply referred to as a “pattern”.

(Embodiment 1)
FIG. 1 is a diagram schematically illustrating a radiation inspection apparatus 100 according to the first embodiment. The radiation inspection apparatus 100 includes a PC (Personal Computer) 10, a monitor 12, a radiation quality changing unit 14, a radiation generator driving unit 16, a substrate holding unit driving unit 18, a detector driving unit 20, a radiation generator 22, and a substrate holding unit. 24, including the detector 26;

  The radiation generator 22 is a part that generates radiation such as X-rays, and this generates radiation by causing accelerated electrons to collide with a target such as tungsten or diamond.

  The substrate holding unit 24 holds a substrate that is an object to be inspected. The substrate held by the substrate holder 24 is irradiated with radiation generated by the radiation generator 22, and the radiation transmitted through the substrate is captured as an image by the detector 26. Hereinafter, the radiation transmission image of the substrate imaged by the detector 26 is referred to as a “transmission image”.

  The transmission image is sent to the PC 10 and reconstructed into an image including the three-dimensional shape of the solder at the joint portion using a known technique such as a filtered back projection method (Filtered-Backprojection method, FBP method). The reconstructed image and the transparent image are stored in a storage in the PC 10 or an external storage (not shown). Hereinafter, an image reconstructed into a three-dimensional image including the three-dimensional shape of the solder at the joint portion based on the transmission image is referred to as a “reconstructed image”. An image obtained by cutting an arbitrary cross section from the reconstructed image is referred to as a “cross section image”. The reconstructed image and the cross-sectional image are output to the monitor 12.

  The radiation quality changing unit 14 changes the radiation quality of the radiation generated by the radiation generator 22. The radiation quality is determined by a voltage (hereinafter referred to as “tube voltage”) applied to accelerate electrons colliding with the target and a current (hereinafter referred to as “tube current”) that determines the number of electrons. The quality changing unit 14 controls the tube voltage and the tube current. This can be achieved using known techniques such as transformers and rectifiers.

  Here, the radiation quality is determined by the brightness and hardness (radiation spectrum distribution) of the radiation. Increasing the tube current increases the number of electrons colliding with the target and the number of radiation photons generated. As a result, the brightness of radiation increases. For example, some components such as capacitors are thicker than other components, and it is necessary to irradiate radiation with high brightness in order to capture transmission images of these components. In such a case, the brightness of the radiation is adjusted by adjusting the tube current. Further, when the tube voltage is increased, the energy of electrons colliding with the target is increased, and the energy (spectrum) of the generated radiation is increased. In general, the greater the energy of radiation, the greater the penetration force of the substance and the less it is absorbed by the substance. A transmitted image captured using such radiation has a low contrast. Therefore, the tube voltage can be used to adjust the contrast of the transmission image.

  The radiation generator drive unit 16 has a drive mechanism such as a motor (not shown), and can move the radiation generator 22 up and down along an axis passing through its focal point. This makes it possible to change the irradiation field by changing the distance between the radiation generator 22 and the object to be inspected held by the substrate holder 24, and to change the magnification of the transmission image picked up by the detector 26. .

  The detector drive unit 20 also has a drive mechanism such as a motor (not shown), and rotates the detector 26 along the detector rotation path 30. The substrate holding unit driving unit 18 also has a drive mechanism such as a motor (not shown), and translates the substrate holding unit 24 on a plane on which the substrate rotation track 28 is stretched. The substrate holding unit 24 is configured to rotate and move on the substrate rotation track 28 in conjunction with the rotational movement of the detector 26. Thereby, it is possible to capture a transmission image while changing the relative positional relationship between the substrate held by the substrate holding unit 24 and the radiation generator 22.

  Here, the rotation radius of the substrate rotation track 28 and the detector rotation track 30 is not fixed, but can be freely changed. Thereby, it becomes possible to change the irradiation angle of the radiation irradiated to the components arranged on the substrate.

  The PC 10 controls all operations of the radiation inspection apparatus 100 described above. Hereinafter, various functions of the PC 10 will be described with reference to the drawings. Although not shown, the PC 10 is connected to an input device such as a keyboard and a mouse.

  FIG. 2 illustrates each functional block processed by the PC 10. The PC 10 includes an imaging control unit 34, an information storage unit 36, a cross-sectional image generation unit 38, an attention image acquisition unit 40, a frequency analysis unit 42, and a frequency component removal unit 44. Each of these functional blocks is realized by cooperation of a CPU that executes various arithmetic processes, hardware such as a RAM used as a work area for data storage and program execution, and software. Therefore, these functional blocks can be realized in various forms by a combination of hardware and software.

  The information storage unit 36 stores imaging conditions for capturing a transmission image of the substrate, information such as the design of the substrate that is the object to be inspected, and design information of the component itself mounted on the substrate. The information storage unit 36 also stores a transmission image, a reconstructed image, and a cross-sectional image of the substrate.

  Here, the “board design information” refers to Gerber data and CAD (Computer Aided Design) data. Information that records the coordinates of the solder joint that joins the substrate and the component is called Gerber data, and information that records the type of the mounted component and the coordinates of the mounting position is called CAD data. The coordinates of the solder joint and component mounting position are described as coordinates in the XY coordinate system set on the board. By referring to the Gerber data and CAD data, it is possible to obtain the type and size of the component existing on the substrate, the position of the component and the solder joint portion, and the wiring pattern on the substrate surface. The “imaging condition” refers to a parameter set in the radiation inspection apparatus 100 when a transmission image necessary for calculating a reconstructed image of a solder that joins a certain component and a substrate is captured. This is, for example, the elevation angle of the radiation generator 22 viewed from the components mounted on the substrate, the number of transmission images required for reconstruction, the magnification of the transmission images, the radiation quality of the radiation to be irradiated, and the like.

  The “component design information” is information on a component unit to be mounted on a substrate, and refers to information such as a solder ball pitch of a BGA, an IC lead pitch, and a wiring pattern inside the component. The pitch of BGA solder balls is about 1 mm (for example, 0.8 mm or 1.4 mm). The lead pitch of the IC is around 0.5 mm.

  The area of the substrate included in the reconstructed image and the resolution of the reconstructed image are determined by the imaging conditions. For example, if the enlargement ratio of the transmission image is increased, the area of the substrate included in the reconstructed image is narrowed. If the angle through which the radiation is transmitted is narrow, the resolution in the direction perpendicular to the substrate is reduced.

  The imaging control unit 34 instructs the radiation quality changing unit 14 to change the tube voltage and the tube current based on the imaging conditions acquired from the information storage unit 36, and the radiation generator driving unit 16, the substrate holding unit driving unit 18, The detector driving unit 20 is controlled to capture a transmission image necessary for generating a reconstructed image. The captured transmission image is stored in the information storage unit 36.

  The cross-sectional image generation unit 38 includes a reconstructed image generation unit 46 and an arbitrary cross-section generation unit 48.

  The reconstructed image generation unit 46 reads the transparent image stored in the information storage unit 36 and generates a reconstructed image. This can be realized using known techniques such as the FBP method and the maximum likelihood estimation method. If the reconstruction algorithm is different, the properties of the obtained reconstructed image and the time required for reconstruction also differ. Therefore, a configuration may be adopted in which a plurality of reconstruction algorithms and parameters used for the algorithm are prepared in advance and the user selects them. Accordingly, it is possible to provide the user with a degree of freedom of selection such as giving priority to shortening the time required for reconstruction or giving priority to good image quality even if time is required.

  The arbitrary cross-section generating unit 48 generates a cross-sectional image obtained by cutting out an arbitrary cross-section from the reconstructed image generated by the reconstructed image generating unit 46. Examples of the cross section include a cross section parallel to the substrate surface. The arbitrary cross-section generator 48 can freely change the interval between cross-sectional images to be generated (hereinafter referred to as “slice interval”) and the thickness of one cross-sectional image (hereinafter referred to as “slice thickness”). . These settings can be changed by the user using an input device such as a keyboard or a mouse (not shown), and the arbitrary section generation unit 48 determines the algorithm based on the imaging conditions stored in the information storage unit 36. You can also. For example, the slice interval is adjusted to the resolution calculated based on the imaging conditions.

  The attention image acquisition unit 40 includes an inspection image specifying unit 74 that specifies a front image to be noted as an inspection target, and an auxiliary image specifying unit 76 that specifies a back image or an internal image to be noticed for extracting a noise component. .

  Since the solder for joining the substrate and the component is on the surface of the substrate, it can be determined whether or not the solder is appropriately joining the substrate and the component by inspecting the surface image. However, as described above, information unnecessary for inspection, such as parts and patterns on the back side of the substrate, may remain in the image on the front surface image where the solder joints used for inspection are imaged. . Therefore, information unnecessary for the inspection is reduced by specifying information such as parts and solder patterns displayed on the back image and subtracting them from the front image. The frequency analysis unit 42 specifies a noise component to be removed from the front surface image based on the back surface image.

  The frequency component removing unit 44 reduces the influence of the noise component specified by the frequency analyzing unit 42 from the surface image specified by the inspection image specifying unit 74. The frequency component removal unit 44 further stores the image in which the influence of the noise component is reduced in the information storage unit 36.

  FIG. 3 is a flowchart showing a flow from capturing a transmission image to storing a cross-sectional image in which the influence of noise components is reduced. The processing in this flowchart starts when, for example, a substrate is set on the substrate holding unit 24 of the radiation inspection apparatus 100 in the production line.

  Under the control of the imaging control unit 34, the imaging unit 32 sets the quality of the radiation by the quality change unit 14, moves the radiation generator 22, the substrate holding unit 24, and the detector 26, and is necessary for generating a reconstructed image. A plurality of transmission images are taken (S10). The reconstructed image generation unit 46 generates a reconstructed image based on the transmission image captured by the imaging unit 32 and stored in the information storage unit 36 (S12). The arbitrary cross section generator 48 generates a cross section image by cutting out a cross section from the reconstructed image generated by the reconstructed image generator 46 (S14). The inspection image specifying unit 74 specifies a surface image from the cross section generated by the arbitrary cross section generating unit 48 (S16). The frequency analyzing unit 42 performs frequency analysis on the back image specified by the auxiliary image specifying unit 76 to extract an image including a noise component (S18). The frequency component removal unit 44 reduces the influence of the noise component extracted by the frequency analysis unit 42 from the surface image (S20). Thereafter, the frequency component removal unit 44 stores the image in which the influence of the noise component is reduced in the information storage unit 36 (S22). When the image in which the influence of the noise component is reduced is stored in the information storage unit 36, the process is finished.

  FIG. 4 is a diagram for explaining the principle of reducing frequency components according to the first embodiment.

  The inspection image specifying unit 74 specifies the front surface image 50, and the auxiliary image specifying unit 76 specifies the back image 56 used for specifying noise. In order to specify the surface image 50, for example, a pattern is acquired from the design information of the substrate stored in the information storage unit 36, and a cross-sectional image that best matches the pattern is acquired using a pattern matching technique. Can be realized. Alternatively, it can also be realized by obtaining a pattern of the substrate surface from the design information of the substrate stored in the information storage unit 36, obtaining the spatial frequency thereof, and obtaining a cross-sectional image having the largest number of the spatial frequency components. For the analysis of the spatial frequency, for example, a two-dimensional fast Fourier transform (FFT) algorithm can be used. Hereinafter, the two-dimensional fast Fourier transform is simply referred to as a two-dimensional Fourier transform.

  In order to specify the back surface image 56, a pattern similar to that in the case of acquiring the pattern of the back surface of the substrate from the design information of the substrate and specifying the front surface image 50 can be used. Alternatively, if the front image 50 can be specified, the thickness of the substrate can be acquired from the design information of the substrate, and the back image 56 can be specified in consideration of the slice interval. Specifically, if the thickness of the substrate is 2 mm and the slice interval is 0.5 mm, for example, the fourth cross-sectional image counted from the front image 50 is the back image 56. In this case, the specific order of the front image 50 and the back image 56 can be reversed. Further, the front image 50, the back image 56, and the spatial frequency can be manually set by the user.

  The frequency analysis unit 42 extracts the frequency component of the pattern on the back surface of the substrate from the back surface image 56 specified by the auxiliary image specifying unit 76. Specifically, it can be realized by acquiring a pattern on the back surface of the substrate from the design information of the substrate and applying the component extraction filter 58 that passes only the spatial frequency of the pattern to the back surface image 56. Alternatively, it can be realized by acquiring a pattern on the back side of the substrate from the design information of the component. If the back surface image is subjected to two-dimensional Fourier transform, and the two-dimensional inverse Fourier transform is performed while leaving only the frequency component of the pattern, an image of the frequency component of the pattern on the back surface of the substrate (hereinafter referred to as “component image”) is obtained.

  The frequency component removing unit 44 subtracts the component image acquired by the frequency analyzing unit 42 from the surface image 50, thereby obtaining a noise component reduced image 54 in which the influence of the component image unnecessary for the inspection is reduced.

  FIG. 5 is a flowchart for explaining the flow of processing for reducing frequency components according to the first embodiment. The processing in this flowchart describes in detail an example of the processing in steps S16, S18, and S20 in FIG.

  The inspection image specifying unit 74 specifies the front image (S24), and the auxiliary image specifying unit 76 specifies the back image (S26). The frequency analysis unit 42 specifies the frequency component of the pattern on the back surface of the substrate from the back image specified by the auxiliary image specifying unit 76 based on the design information of the substrate (S28). The frequency analysis unit 42 acquires a component image from the back image (S30). The frequency component removing unit 44 subtracts the component image from the surface image, thereby acquiring a noise component reduced image obtained by removing the component image unnecessary for the inspection from the surface image (S20).

  As described above, according to the first embodiment, it is possible to reduce information unnecessary for the inspection existing on the back image remaining in the front image. The information is extracted by analyzing the back image, which is advantageous in that the extraction accuracy is increased.

(Embodiment 2)
In the second embodiment, a surface image is specified and information unnecessary for inspection is reduced based on the surface image. Unlike the first embodiment, there is a difference that the back image is not used to specify the noise component. The following description is centered on differences from the first embodiment.

  Since the second embodiment does not use the back image, the auxiliary image specifying unit 76 is not included. The operations of the frequency analysis unit 42 and the frequency component removal unit 44 are different from those of the first embodiment.

  FIG. 6 is a diagram for explaining the principle of reducing frequency components according to the second embodiment.

  The inspection image specifying unit 74 specifies the surface image 50 by the same method as in the first embodiment. The frequency analysis unit 42 acquires the pattern on the back surface of the substrate from the design information stored in the information storage unit 36, and calculates the frequency component of the pattern. Next, the frequency analysis unit 42 generates a Gaussian-Laplacian pyramid of the surface image 50.

  Here, the Gaussian-Laplacian pyramid refers to a set of a reference image, an image obtained by applying a low-pass filter of different strength to the reference image, and a difference image of these images. For example, in FIG. 6, the surface image 50 is a reference image. The reference image is blurred by down-sampling by applying a Gaussian filter which is a kind of low-pass filter of appropriate strength to the reference image, and the first low-resolution image 62a is obtained. The first low-resolution image 62a is further blurred by applying a Gaussian filter, and then down-sampled to obtain a second low-resolution image 62b. Similarly, a third low resolution image 62c is obtained. In this case, since the second low resolution image 62b is obtained by applying a Gaussian filter to the first low resolution image 62a, the Gaussian filter is applied twice to the reference image. Therefore, when the first low-resolution image 62a and the second low-resolution image 62b are compared, it can be said that both are images obtained by applying low-pass filters having different strengths to the reference image. That is, the second low resolution image 62b is subjected to a stronger low-pass filter than the first low resolution image 62a, resulting in a more blurred image (low frequency image).

  Now, the first difference image 64a is obtained by subtracting the first low resolution image 62a from the reference image. Here, considering that the first low-resolution image 62a is an image obtained by applying a Gaussian filter that is a low-pass filter to the reference image, the first difference image 64a includes a high-frequency component of the reference image. Think of it as an image. When subtracting the first low-resolution image 62a from the reference image, a measure is taken to make the first low-resolution image 62a the same number of pixels as the reference image, such as upsampling.

  The second difference image 64b is obtained by subtracting the second low resolution image 62b from the first low resolution image 62a. The frequency component included in the first low-resolution image 62a is lower than the frequency component included in the reference image. The frequency component included in the second low resolution image 62b is lower than the frequency component included in the first low resolution image 62a. Accordingly, the second difference image 64b is an image including a high frequency component lower than that of the first difference image 64a.

  Similarly, the third difference image 64c is obtained by subtracting the third low resolution image 62c from the second low resolution image 62b. The frequency component included in the third difference image 64c is lower than the frequency component included in the second difference image 64b.

  Thus, a series of images obtained by applying a Gaussian filter, such as the reference image, the first low resolution image 62a, the second low resolution image 62b, and the third low resolution image 62c, is a “Gaussian pyramid”. Sometimes called. Also, a first difference image 64a, which is a difference image between the reference image and the first low resolution image 62a, a difference image 64b between the first low resolution image 62a and the second low resolution image 62b, and a second low image. A series of images including different frequency components generated from the difference image of the pyramid image, such as the difference image 64c between the resolution image 62b and the third low resolution image 62c, may be referred to as a “Laplacian pyramid”. The Gaussian pyramid and the Laplacian pyramid are sometimes referred to as the “Gaussian-Laplacian pyramid”.

  Note that the reference image is restored by sequentially adding the third difference image 64c, the second difference image 64b, and the first difference image 64a to the third low-resolution image 62c that is the lowest-resolution image. Can do. Therefore, the third low-resolution image 62c, the third difference image 64c, the second difference image 64b, and the first difference image 64a include all the information of the reference image, and all the frequency components included in the respective images Different. In this sense, the third low-resolution image 62c, the third difference image 64c, the second difference image 64b, and the first difference image 64a can be said to be images obtained by decomposing the reference image into images including different frequency components. Hereinafter, such decomposition of an image into a plurality of frequency components may be referred to as “multiple frequency analysis”.

  As described above, the Laplacian pyramid can be said to be an image obtained by decomposing the reference image into various frequency components. Since the pattern of solder or the like included in the back surface of the image has a specific frequency component, the pattern should appear across one or a plurality of Laplacian pyramid images. Therefore, when restoring the image, the influence of the restored image can be reduced by removing the Laplacian pyramid image in which the pattern appears or by multiplying the pixel value by a coefficient of 1 or less.

  Specifically, it is assumed that the pattern appears in the second difference image 64b in FIG. In this case, the frequency component removing unit 44 sets γ (symbol 66c), which is a coefficient to be multiplied to the third difference image 64c, to 1, and first generates a third modulation difference image 68c. γ is set to 1 because the frequency component of the pattern is not present in the third difference image 64c. The first modulated low-resolution image 70a is obtained by adding the upsampled third low-resolution image 62c to the third modulation difference image 68c.

  Next, β (symbol 66b) that is a coefficient to be multiplied to the second difference image 64b is set to a coefficient of 1 or less (for example, 0.1), and the second modulation difference image 68b is generated. The second modulated low-resolution image 70b is obtained by adding the up-sampled first modulated low-resolution image 70a to the second modulation difference image 68b. Since the value of the coefficient β is 1 or less, the second modulation difference image 68b is reduced in the information of the second difference image 64b, that is, the pattern information.

  Since the first difference image 64a is not affected by the pattern, the coefficient α (symbol 66a) to be multiplied by the first difference image 64a is set to 1 as in the case of the third modulation difference image 68c, and the first modulation is performed. A difference image 68a is created. By adding the second modulated low resolution image 70b upsampled to the first modulation difference image 68a, a noise component reduced image 72 in which frequency components peculiar to the pattern are finally reduced is obtained.

  The number of Gaussian pyramids and the intensity of the Gaussian filter can be freely changed. The number of stages and the filter strength may be determined based on the design information so that a pattern appears in a specific Laplacian image.

  FIG. 7 is a flowchart for explaining the flow of processing for reducing frequency components according to the second embodiment. The processing in this flowchart describes in detail an example of the processing in steps S16, S18, and S20 in FIG.

  The inspection image specifying unit 74 acquires the board design information from the information storage unit 36 (S32). Next, a surface image is specified based on the acquired substrate design information (S34). The frequency analysis unit 42 also first acquires board design information from the information storage unit 36 (S36). Instead of using the back surface image, the frequency component of the pattern on the back surface of the substrate is specified by calculation based on the acquired design information of the substrate (S38). Next, the frequency analysis unit 42 generates a Gaussian-Laplacian pyramid image of the surface image (S40). The frequency component removal unit 44 specifies a difference image (Laplacian image) including the frequency component of the pattern on the back surface of the substrate (S42). Next, a coefficient (restoration coefficient) for restoring the image is set (S44), and the image is restored (S46).

  As described above, according to the second embodiment, it is possible to reduce information unnecessary for the inspection existing on the back image remaining in the front image. Since the method according to the second embodiment can be performed by specifying only the front image without using the back image, the time required for extracting the back image can be omitted, and it is advantageous in that it is not affected by the extraction accuracy of the back image. is there.

(Embodiment 3)
The third embodiment specifies information unnecessary for the inspection existing in the inspection image from the auxiliary image, but differs from the first embodiment mainly in that the auxiliary image is a pseudo-transparent image.

  FIG. 8 is a diagram schematically illustrating the attention image acquisition unit 40 according to the third embodiment. The attention image acquisition unit 40 includes a pseudo transmission image generation unit 78 in addition to the inspection image specification unit 74 and the auxiliary image specification unit 76.

  As described above, the surface image to be inspected includes the structure on the back side of the substrate as noise. Therefore, the auxiliary image specifying unit 76 in Embodiment 3 specifies a plurality of cross-sections existing on the back side of the substrate based on the design information of the substrate stored in the information storage unit 36.

  Based on the plurality of cross-sectional images specified by the auxiliary image specifying unit 76, the pseudo-transparent image generating unit 78 generates a pseudo-transparent image of the structure of the board or component displayed in the cross-sectional group. Here, the pseudo-transmission image is a cross-sectional image having a thickness obtained by projecting a plurality of cross sections in a direction perpendicular to the substrate surface. The pseudo-transparent image generation unit 78 can also generate a plurality of pseudo-transparent images by further dividing the plurality of cross sections into a plurality of subsets and generating respective pseudo-transparent images. Further, when only one cross-sectional image is input, the pseudo-transparent image generation unit 78 outputs the input cross-sectional image itself as the generated pseudo-transparent image.

  In the first embodiment, the frequency analysis unit 42 extracts the frequency component of the pattern on the back surface of the substrate based on the cross-sectional image specified by the auxiliary image specifying unit 76. In the third embodiment, the frequency analyzing unit 42 uses the pseudo-transparent image generated by the pseudo-transparent image generating unit 78 based on the cross-sectional image specified by the auxiliary image specifying unit 76 as an auxiliary image. Extract frequency components. The specific extraction method is the same as in the first embodiment.

  The inspection image specifying unit 74 according to the third embodiment can specify a plurality of cross sections existing on the surface side of the substrate. The pseudo-transparent image generating unit 78 generates a pseudo-transparent image based on the plurality of cross sections specified by the inspection image specifying unit 74 and uses it as an inspection image. Note that when the inspection image specifying unit 74 and the auxiliary image specifying unit 76 specify only one cross section, the pseudo-transparent image generating unit 78 performs the same operation as in the first embodiment.

  According to the third embodiment described above, it is possible to reduce information unnecessary for the inspection existing on the back image remaining in the front image. In the third embodiment, a noise component is extracted from one pseudo-transmission image generated based on a plurality of cross-sectional images on the back side of the substrate. This pseudo-transparent image is considered to contain noise information distributed over a plurality of sheets, and such a wide range of noise information can be extracted by one extraction operation, which is advantageous in terms of calculation cost. In the third embodiment, the information on the tomographic images on the plurality of surface sides on which the solder joints for joining the component and the substrate are projected can be handled at one time, and the crack occurred at a position away from the substrate It is also advantageous in that it is possible to inspect solder defects such as.

(Embodiment 4)
In the fourth embodiment, information unnecessary for inspection existing in the inspection image is specified from the auxiliary image. However, the auxiliary image is not a tomographic image or pseudo-transmission image of the substrate, but a component alone before being mounted on the substrate. The second embodiment is different from the first embodiment in that it is a cross-sectional image generated based on the image picked up in FIG. Therefore, the operation other than the operation of the auxiliary image specifying unit 76 is the same as that of the first embodiment.

  In the fourth embodiment, the auxiliary image specifying unit 76 specifies a component mounted on the board based on the board design information stored in the information storage unit 36. Next, a tomographic image of a single component generated based on a previously captured transmission image and stored in the information storage unit 36 is acquired. The frequency analysis unit 42 extracts the frequency component of the pattern on the back surface of the substrate from the tomographic image of the component alone. The specific extraction method is the same as in the first embodiment.

  As described above, according to the fourth embodiment, it is possible to reduce information unnecessary for the inspection reflected in the surface image. In the fourth embodiment, a noise component is extracted based on a tomographic image of a single component that has been captured and generated in advance. Since this tomographic image contains no information other than components such as a substrate, it is advantageous in that noise components can be accurately extracted.

  The four embodiments have been described above. Combinations of these embodiments are also useful as embodiments.

  When combining the first embodiment and the second embodiment and using the multiple frequency analysis similar to the second embodiment in the frequency analysis unit 42 in the first embodiment, in addition to the effects of the first embodiment, there is also an effect that the frequency components can be extracted more finely. is there.

  When combining the second embodiment and the third embodiment and using the multiple frequency analysis in the frequency analysis unit 42 in the third embodiment as in the second embodiment, in addition to the effects of the third embodiment, there is an effect that the frequency components can be extracted more finely. is there. As another combination method, when a pseudo-transparent image is adopted as the inspection image of the second embodiment, in addition to the effects of the second embodiment, it is advantageous in that more information can be handled than one cross-sectional image.

  When the third embodiment and the fourth embodiment are combined and a pseudo-transmission image based on a tomographic image of a single component that is captured and generated in advance as an auxiliary image is used, a new embodiment that is generated by the combination is each of the combined embodiments. Combined with effects.

  The present invention has been described based on the embodiments. It is to be understood by those skilled in the art that the embodiments are exemplifications, and that various modifications can be made to combinations of the respective components and processing processes, and such modifications are within the scope of the present invention.

  In the first embodiment, the case where the auxiliary image specifying unit 76 specifies the back image as the auxiliary image has been described. Here, not only one back surface image but also a plurality of back side images existing on the back surface side of the substrate may be specified. In this case, the frequency analysis unit 42 and the frequency component removal unit 44 perform the same processing for each of the plurality of back side images. This is advantageous in that noise components distributed over a plurality of sheets can be reduced.

  When the modified example of the first embodiment and the third embodiment are combined and a plurality of pseudo-transparent images are used as the auxiliary image, the noise components distributed over the plurality of images are all in cross-sectional image units. It is more advantageous in terms of calculation cost than processing. Also, using a pseudo-transparent image as the inspection image and using a plurality of tomographic images or pseudo-transparent images as the auxiliary image is advantageous in terms of calculation cost over processing all in units of cross-sectional images.

  In the above description, the substrate back surface image is mainly described as the noise component remaining in the front surface image, but there is also a wiring pattern inside the component as the noise component remaining in the front surface image. Therefore, the auxiliary image specifying unit 76 may specify an image including the internal structure of the component instead of or in addition to the back image as a target for extracting the noise component. In this case, since the types of images including noise components remaining in the surface image increase, it is advantageous in that the accuracy of noise component extraction is improved.

  In the above description, the reduction of noise components remaining in the surface image has been described. However, if the multi-frequency analysis is used, the solder arrangement pattern of the surface image necessary for the inspection can be highlighted and displayed.

  Solder for joining a board | substrate and components is arrange | positioned periodically on the lower surface of BGA or LGA. This solder placement information can be obtained from the design information. Therefore, according to the second embodiment, the frequency component of the solder arrangement is calculated based on the design information, and the number of steps of the Gaussian pyramid and the intensity of the Gaussian filter so that the frequency component appears in a specific (multiple) Laplacian image. Is adjusted, a Laplacian image in which many frequency components of the solder arrangement appear can be obtained. When the coefficient multiplied to the Laplacian image is increased during image restoration, an image in which the solder arrangement pattern of the surface image is emphasized can be obtained. The magnitude of the coefficient may be determined by experiment according to the type of component and the degree of emphasis.

  In the above description, in Embodiment 1, the back image is used for specifying the noise component. However, when the substrate is a multilayer, the noise component derived from the pattern of the intermediate layer inside the substrate is reflected in the front image. May come. In this case, a cross-sectional image inside the substrate may be used for specifying the noise component. Design information may be used to specify the noise component.

  10 PC, 12 Monitor, 14 Radiation quality changing unit, 16 Radiation generator driving unit, 18 Substrate holding unit driving unit, 20 Detector driving unit, 22 Radiation generator, 24 Substrate holding unit, 26 Detector, 28 Substrate rotation trajectory , 30 detector rotation trajectory, 32 imaging unit, 34 imaging control unit, 36 information storage unit, 38 cross-sectional image generation unit, 40 attention image acquisition unit, 42 frequency analysis unit, 44 frequency component removal unit, 46 reconstructed image generation unit 48 Arbitrary slice generator, 50 Front image, 54 Noise component reduced image, 56 Back image, 58 Component extraction filter, 62a First low resolution image, 62b Second low resolution image, 62c Third low resolution image, 64a first difference image, 64b second difference image, 64c third difference image, 68a first variation Difference image, 68b second modulation difference image, 68c third modulation difference image, 70a first modulation low resolution image, 70b second modulation low resolution image, 72 noise component reduced image, 74 inspection image specifying unit, 76 Auxiliary image specifying unit, 78 pseudo transmission image generating unit, 100 radiation inspection apparatus.

Claims (9)

An imaging unit that irradiates radiation from a plurality of directions to a plurality of locations of an object to be inspected including a substrate and an electronic component, and captures a plurality of radiation transmission images;
A cross-sectional image generator that reconstructs a plurality of cross-sectional images of the object to be inspected based on the plurality of radiation transmission images;
An inspection image specifying unit for acquiring an inspection image based on the plurality of cross-sectional images;
An auxiliary image specifying unit for acquiring an auxiliary image on which at least a part of the inspection object is projected;
A frequency analysis unit that identifies a frequency component that the inspection image and the auxiliary image include in common;
A frequency component removing unit that reduces the influence of the frequency component included in common from the inspection image;
Radiation inspection apparatus characterized by including.   The apparatus according to claim 1, wherein the auxiliary image specifying unit acquires an auxiliary image from the plurality of cross-sectional images based on a cross-sectional image at a position different from the inspection image in the vertical direction. A pseudo-transparent image generation unit that generates a pseudo-transparent image by superimposing a plurality of cross-sectional images;
The pseudo-transparent image generating unit uses the pseudo-transparent image generated from one or a plurality of cross-sectional images acquired from the plurality of cross-sectional images by the inspection image specifying unit as the inspection image, and the auxiliary image specifying unit The apparatus according to claim 1 or 2, wherein a pseudo-transparent image generated from one or a plurality of slice images acquired from the plurality of slice images is used as the auxiliary image. The pseudo-transparent image generation unit divides one or a plurality of cross-sectional images acquired from the plurality of cross-sectional images by the auxiliary image specifying unit into one or a plurality of subsets, and 1 or A plurality of pseudo-transparent images as the auxiliary image,
The frequency analysis unit specifies a frequency component that is included in common with the inspection image for each of the one or more pseudo-transmission images,
The apparatus according to claim 3, wherein the frequency component removing unit reduces the influence of each of the commonly included frequency components from the inspection image. The inspection image specifying unit acquires a cross-sectional image on one surface side of the substrate as the inspection image,
The apparatus according to claim 1, wherein the auxiliary image specifying unit acquires a cross-sectional image of the other surface side of the substrate or the inside of the substrate as the auxiliary image. The inspection image specifying unit includes one of two types of cross-sectional images: a cross-sectional image in which a solder for joining the electronic component and the substrate is projected, and a cross-sectional image in which the inside of the electronic component is projected. An image is acquired as the inspection image,
The apparatus according to claim 1, wherein the auxiliary image specifying unit acquires, as the auxiliary image, the other cross-sectional image different from the inspection image among the two types of cross-sectional images.   The apparatus according to claim 1, wherein the auxiliary image specifying unit acquires, as the auxiliary image, a cross-sectional image of a component of the same type as the electronic component that has been previously imaged alone and reconstructed.   The apparatus according to claim 1, wherein the frequency analysis unit specifies a frequency component that the inspection image and the auxiliary image include in common based on design information of the board.   The apparatus according to any one of claims 1 to 7, wherein the frequency analysis unit specifies a frequency component that the inspection image and the auxiliary image include in common based on design information of a part.

Images