JP2023050842A - Image inspection method, and image inspection device - Google Patents

Abstract

To provide an image inspection method using an image inspection device (for example, an X-ray inspection device) that, in transferring an image of a faulty part acquired in the past to a transfer destination in an image of a non-defective product, if the resolution of the transfer source image is higher than the resolution of the transfer destination, reduces the resolution of the transfer source image to bring the resolution of the transfer source image closer to the resolution of the transfer destination, and thereby can create a more natural transfer image.SOLUTION: An image inspection method using an image inspection device has a transfer step of transferring one or more images of faulty parts among images discriminated as defective products in the past to a transfer destination in an image discriminated as a non-defective product in the past to create a transfer image, as a part of all steps, and in the transfer step, in transferring the images of the faulty parts to the transfer destination, if the resolution of the images of the faulty parts is higher than the resolution of the transfer destination, reduces the resolution of the images of the faulty parts to bring the resolution of the images of the faulty parts closer to the resolution of the transfer destination.SELECTED DRAWING: Figure 6

Description

According to the present invention, when an image of a defective portion of a certain component is transferred to a transfer destination of an image of another non-defective product, the resolution of the image of the defective portion is brought close to the resolution of the transfer destination to create a more natural transfer image. The present invention relates to an image inspection method and an image inspection apparatus.

Conventionally, it is known to create a pseudo defective image based on a non-defective product image when the number of defective images required for quality control cannot be obtained in an actual manufacturing process. For example, there is known an automatic pseudo-defective-image creating apparatus capable of automatically creating a large number of pseudo-defective images required for initial learning. The pseudo-defective image automatic creation device receives a non-defective product image from the non-defective product input unit as learning data for the neural network, and learns the image. Also, the defective image is input from the defective image input unit and learned. Further, the defect image extraction unit extracts the difference data from the non-defective product, and the pseudo data condition setting unit creates various defect creation conditions such as the defect synthesis position from the random values of the random number generation unit. Based on this, the pseudo-defective image creation unit synthesizes the difference data with the non-defective image to create various pseudo-defective images, which are input as learning data for the neural network (see, for example, Patent Document 1).

Also known is a learning device capable of accurately learning a model for discriminating non-defective products from defective products. The learning device includes an intermediate image generation unit, an intermediate image display unit, a boundary reception unit, and a teacher image identification unit. The intermediate image generation unit generates a plurality of intermediate images from a non-defective product image representing a non-defective product and a defective product image representing a defective product. The intermediate image display unit arranges a plurality of intermediate images between the non-defective product image and the defective product image and displays them on the display device. The boundary reception unit receives from the user a designation of a boundary of the intermediate image as a boundary between the non-defective product image and the defective product image. The teacher image identifying unit identifies the non-defective product image and the defective product image based on the specified boundaries (see Patent Document 2, for example). An image processing apparatus capable of easily generating a non-defective product image from a defective product image is also known (for example, see Patent Document 3).

According to the content of the invention as disclosed in the above Patent Document 1, a pseudo-defective image is created based on data relating to a non-defective image and data relating to a defective image, and the pseudo-defective image is input as training data for a neural network. By doing so, it is possible to improve the accuracy of quality judgment as an inspection device of the automatic pseudo-defect image creation device. However, it cannot be said that pseudo defective images and their measured values can be efficiently created using defective images that have been accumulated from the past. For example, when a pseudo-defective image is generated, the pseudo-defective image may have a different resolution than the actual defective image, resulting in an unnatural image.

Further, according to the content of the invention as disclosed in Patent Document 2, when generating a plurality of intermediate images based on the non-defective product image and the defective product image, the boundary reception unit receives the non-defective product image and the defective product image. As the boundary, the specification of the boundary of the intermediate image is received from the user. However, if a sufficient number of defective images cannot be obtained when specifying the boundary of the intermediate image, the boundary of the intermediate image must be specified based only on the non-defective image, and specification that relies on experience is required. There is fear. Designation that relies on experience is not preferable because there is a risk that the quality of inspection by the learning device will vary. Further, as in the case of Japanese Patent Application Laid-Open No. 2002-200005, it cannot be said that a new pseudo-defective image can be efficiently created using the defective images that have been accumulated from the past.

JP 2005-156334 A Japanese Patent No. 6780769 JP 2020-008488 A

The present invention has been made in view of the above-described problems. When the resolution of the image is higher than that of the transfer destination, lowering the resolution of the transfer source image brings the resolution of the transfer source image closer to the resolution of the transfer destination, creating a more natural transferred image. A final object is to provide an image inspection method and an image inspection apparatus characterized by:

The present invention for solving the above problems is
An image inspection method that makes it possible to determine whether an object to be inspected is good or bad using an image inspection device,
an image display step of displaying an image that has been determined to be non-defective in the past;
a defective image display step of displaying an image that has been determined to be defective in the past;
a selection step of selecting a transfer destination of the image that was determined to be non-defective in the past, and an image of one or more defective portions of the image that was determined to be defective in the past corresponding to the transfer destination;
Transferring one or more images of defective portions from the image selected in the selection step to the transfer destination of the image determined to be non-defective in the past to generate a transfer image. and a transfer step of
In the transfer step, when the image of the defective portion is transferred to the transfer destination, if the resolution of the image of the defective portion is higher than the resolution of the transfer destination, the resolution of the image of the defective portion is set to a low resolution. The image inspection method is characterized in that the resolution of the image of the defective portion is brought close to the resolution of the transfer destination by converting the image.

According to the present invention, in the transfer step, by bringing the resolution of the image of the defective portion closer to the resolution of the transfer destination, it is possible to easily create a more natural transferred image. This transferred image can be used to determine the difference between non-defective and defective products in the manufacturing process, thereby facilitating the teaching work of registering the difference between non-defective and defective products in the image inspection device. This transfer image can also be provided as a criterion for distinguishing between good and bad products. In addition, the more defective images are transferred in the transfer step, the more pseudo defective images are obtained.

Further, in the present invention, the image inspection method may further include a transferred image display step of displaying the transferred image. According to this, by displaying the transfer image in the transfer image display step, it is possible to visually distinguish the difference between the non-defective product and the defective product. become easier.

Further, in the present invention, in the transfer step, the resolution of the image of the defective portion may be reduced by compressing the size of the image of the defective portion so as to be equivalent to the size of the transfer destination. It may be an image inspection method, which is possible. According to this, it is possible to easily lower the resolution of the image of the defective portion simply by changing the size of the image of the defective portion to be transferred.

Further, in the present invention, a feature amount acquisition step of acquiring a predetermined feature amount in the transfer image and other inspection objects, and based on the measurement result of the predetermined feature amount obtained in the feature amount acquisition step, an automatic calculation step of automatically calculating a threshold value for newly classifying the inspection object into non-defective products and defective products after measurement; The image inspection method may further include a measurement result display step of displaying the threshold value obtained in the automatic calculation step. According to this, by displaying the threshold value in the measurement result display step, it is possible to visually and numerically distinguish between non-defective products and defective products. becomes even easier.

Further, in the present invention, in the result display step, the defective product is displayed such that the transfer image and the inspection object including the defective portion among the other inspection objects can be further identified. The image inspection method may be characterized by: A transfer image is, so to speak, a defective image that is intentionally created. According to this, defects can be managed by distinguishing between defects derived from the transferred image and defects actually occurring in the manufacturing process.

Further, in the present invention, the threshold value is an intermediate value between an average value of the predetermined feature amount in the non-defective item and an average value of the predetermined feature amount in the defective item. It can be used as a method. According to this, the threshold value can be quickly calculated by a simple method.

Further, in the present invention, the image inspection method may be characterized in that the threshold value is the maximum value of the predetermined feature amount in the image that has been determined to be defective in the past and the transfer image. According to this, the threshold value can be quickly calculated by a simple method.

Further, in the present invention, the image inspection method may be characterized in that, in the result display step, the measurement result is displayed in the form of a histogram. The histogram display makes it easy to check the frequency of defective products according to the measured values. You can easily check the distribution of measured values.

Further, in the present invention, in the image display step and the defective image display step, the image that has been determined to be a non-defective product in the past and the image that has been determined to be a defective product in the past are tomographic images from three directions of XYZ, respectively. It is good also as an image inspection method characterized by displaying as . According to this, when the defective product is three-dimensional, it is possible to prevent the risk of overlooking the defective portion of the defective product due to blind spots.

Further, in the present invention, the object to be inspected is a circuit board on which an electronic element is mounted, and the transfer destination of the image that has been determined to be non-defective in the past is a connection portion between the electronic element and the circuit board. The image inspection method may be characterized in that the boundary in the XY direction of the connecting portion in the image that has been determined to be a non-defective product in the past is automatically determined by binarization. According to this, even if the image before transfer is unclear, the contour of the transfer destination can be automatically clarified, and the transfer destination is clearly determined.

Further, in the present invention, the image inspection method may be characterized in that the boundary of the connection portion in the Z direction is set manually or automatically. According to this, in the Z direction, even if the image before transfer is unclear, the transfer destination can be clearly determined by the user's sense.

Further, in the present invention, in the image inspection method described above, there is provided a method for managing images of a plurality of defective locations, wherein the images of the defective locations are stored in a database, and for each of the images of the defective locations, The type of the object to be inspected, the type of the defect, and the defect level represented by a numerical value representing the degree of defect are linked, and the defect level represented by the numerical value is based on the measurement result measured in the feature value acquisition step in the past. Alternatively, the image management method of the defective portion may be characterized by being a numerical value determined based on the result of visual judgment in the past. According to this, it is possible to grasp more detailed information about the defective portion. It is also possible to provide more detailed information about this faulty location.

Moreover, in the present invention,
An image inspection device capable of determining whether an object to be inspected is good or bad,
a database that accumulates data of images that have been determined to be non-defective in the past;
a defect database that accumulates image data that has been determined to be defective in the past;
One or more defective portions of the image determined as defective in the past obtained from the defective database corresponding to the transfer destination in the image determined as non-defective in the past obtained from the database. a transfer unit that transfers the image of and generates a transferred image;
a display unit that displays the image that was previously determined as a non-defective product that was acquired from the database, and the image that was determined as a defective product in the past, which was acquired from the defective database;
with
When transferring the image of the defective portion to the transfer destination, the transfer unit reduces the resolution of the image of the defective portion to a low resolution when the resolution of the image of the defective portion is higher than the resolution of the transfer destination. The image inspection apparatus may be characterized in that the resolution of the image of the defective portion is brought close to the resolution of the transfer destination by converting the image.

According to the present invention, it is possible to easily create a more natural transfer image with a simple device configuration, and the criteria for distinguishing non-defective products from non-defective products are clarified. This discriminant criterion can be used to discriminate between good and bad products in the manufacturing process, thereby facilitating the teaching task.

Further, in the present invention, the image inspection apparatus may be characterized in that the display section displays the transfer image obtained from the transfer section. According to this, by displaying the transfer image on the display unit, it is possible to visually distinguish between non-defective products and defective products. .

Further, in the present invention, the transfer unit may reduce the resolution of the image of the defective portion by compressing the size of the image of the defective portion so as to be equivalent to the size of the transfer destination. It may be an image inspection device, which is possible. According to this, it is possible to easily reduce the resolution of the image of the defective portion simply by measuring the size of the transfer destination and the size of the image of the defective portion.

Further, in the present invention, a measurement unit for measuring a predetermined feature amount in the transfer image and other inspection objects, and based on the measurement result of the predetermined feature amount acquired from the measurement unit, the above new measurement is performed after the measurement. a calculation unit that automatically calculates a threshold for classifying the inspection object into non-defective products and defective products, wherein the display unit displays the measurement result of the predetermined feature amount acquired from the measurement unit and/or The image inspection method may be characterized by displaying the threshold value obtained from the calculation unit. According to this, by displaying the threshold value on the display unit, it is possible to visually and numerically distinguish the difference between non-defective products and defective products. Become.

Further, in the present invention, the defective product is displayed on the display unit so that the inspection object including the defective portion can be further identified from the transferred image and the other inspection objects. It is good also as an image inspection apparatus. According to this, the defects can be managed by distinguishing between defects originating in the transferred image and defects actually occurring in the manufacturing process.

Further, in the present invention, the threshold value is an intermediate value between an average value of the predetermined feature amount in the non-defective item and an average value of the predetermined feature amount in the defective item. It may be used as a device. According to this, the threshold value can be quickly calculated by a simple method.

Further, in the present invention, the image inspection apparatus may be characterized in that the threshold value is the maximum value of the predetermined feature amount in the image that has been determined to be defective in the past and the transfer image. According to this, the threshold value can be quickly calculated by a simple method.

Further, in the present invention, the image inspection apparatus may be characterized in that the measurement result is displayed in the form of a histogram on the display section. The histogram display makes it easy to check the frequency of defective products according to the measured values. You can easily check the distribution of measured values. Also, if the transfer image and the histogram can be displayed simultaneously on the display section, it becomes easier to visually and numerically distinguish between non-defective products and defective products.

Further, in the present invention, the image that has been determined to be a non-defective product in the past and the image that has been determined to be a defective product in the past are displayed as tomographic images from three directions of XYZ, respectively, on the display unit. It is good also as an image inspection apparatus. According to this, when the defective product is three-dimensional, it is possible to prevent the risk of overlooking the defective portion of the defective product due to blind spots. Since the display unit can display images of non-defective products and images of defective products as tomographic images from all of the XY, YZ, and XZ directions, it is easy to grasp the overall image of non-defective products and defective parts.

Further, in the present invention, the object to be inspected is a circuit board on which an electronic element is mounted, and the transfer destination of the image that has been determined to be non-defective in the past is a connection portion between the electronic element and the circuit board. The image inspection apparatus may be characterized in that a boundary in the XY direction of the connecting portion in the image that was determined to be a non-defective product in the past is automatically determined by binarization. According to this, even if the image before transfer is unclear, the contour of the transfer destination can be automatically clarified, and the transfer destination is clearly determined.

Further, in the present invention, the image inspection apparatus may be characterized in that the boundary of the connection portion in the Z direction is set manually or automatically. According to this, in the Z direction, even if the image before transfer is unclear, the transfer destination can be clearly determined by the user's sense.

Further, in the present invention, the images of the defective locations are stored in a database, and for each of the images of the defective locations, the types of the inspection objects, the types of defects, and the numerical values representing the degree of defects are displayed. The stage is linked, and the defective stage by the numerical value is a numerical value determined based on the measurement result measured by the measuring unit in the past or based on the result visually judged in the past. It is good also as an image inspection apparatus. According to this, it is possible to grasp more detailed information related to the defective portion.

The means for solving the above problems can be used in combination with each other as much as possible.

According to the present invention, when an image of a defective portion acquired in the past (image of a transfer source) is transferred to a transfer destination of a non-defective image, the resolution of the transfer source image is higher than the resolution of the transfer destination. In this case, by lowering the resolution of the original image, it is possible to create a more natural transfer image with both resolutions. As a result, it becomes possible to visually discriminate the difference between non-defective products and defective products in the inspection object and use it for quality control. Also, teaching can be easily performed based on this transferred image.

FIG. 1 is a functional block diagram showing an example of an X-ray inspection apparatus according to an embodiment of the invention. FIG. 2 is an explanatory diagram showing an example of a method of managing images of a plurality of defective locations in an image inspection method using an X-ray inspection apparatus according to an embodiment of the present invention. FIG. 3A is a first diagram illustrating the flow of an example of contents displayed by the display unit in the image inspection method using the X-ray inspection apparatus according to the embodiment of the present invention. FIG. 3B is a second diagram illustrating the flow of an example of contents displayed by the display unit in the image inspection method using the X-ray inspection apparatus according to the embodiment of the present invention. FIG. 4A is a third diagram illustrating the flow of an example of contents displayed by the display unit in the image inspection method using the X-ray inspection apparatus according to the embodiment of the present invention. FIG. 4B is a fourth diagram illustrating the flow of an example of contents displayed by the display unit in the image inspection method using the X-ray inspection apparatus according to the embodiment of the present invention. FIG. 5 is a fifth diagram for explaining the flow of an example of contents displayed by the display section in the image inspection method using the X-ray inspection apparatus according to the embodiment of the present invention. 6A, 6B, and 6C are supplementary diagrams explaining in detail the transfer flow shown in FIG. 4B. FIG. 7 is a flow chart showing the procedure of the image inspection method using the X-ray inspection apparatus according to the embodiment of the present invention.

[Example of application]
An outline of an application example of the present invention will be described below with reference to some of the drawings. The present invention can be applied to an X-ray inspection apparatus 1 as shown in FIG. Moreover, the present invention can be applied to the processing shown in the flowchart of FIG. 7 by using the X-ray inspection apparatus 1 .

FIG. 1 is a functional block diagram showing an example of an X-ray inspection apparatus 1 to which the present invention is applicable. As shown in FIG. 1 , the X-ray inspection apparatus 1 is roughly divided into an imaging section 10 , a computing section 20 and a display section 30 . The imaging unit 10 includes a stage 11 , an imaging condition storage unit 12 , an X-ray generator 13 and an X-ray detector 14 . An object to be inspected by the X-ray inspection apparatus 1 (for example, a circuit board to which electronic components are soldered) is placed on the stage 11, and the imaging conditions read from the imaging condition storage unit 12 (for example, imaging distance, (brightness, etc.) is captured by an imaging camera (not shown). The three-dimensional structure and the like of the object to be inspected placed on the stage 11 are analyzed by X-rays generated from the X-ray generator 13 . The X-ray detector 14 detects the intensity of the X-ray emitted from the X-ray generator 13 and transmitted through the inspection object.

The calculation unit 20 includes a three-dimensional data creation unit 21 , a database 22 , a defect database 23 , a transfer unit 24 , a measurement unit 25 and a calculation unit 26 . The three-dimensional data creation unit 21 creates three-dimensional data based on the image of the inspection object captured by the imaging camera. Note that three-dimensional data is a tomographic image viewed from three directions of XYZ. Here, it is often difficult to obtain three-dimensional data of defective products from the inspection object because the frequency of occurrence of defects in the manufacturing process of the inspection object is low. Therefore, the three-dimensional data of the object to be inspected is often created as non-defective data. Also, the created three-dimensional data is accumulated in the database 22 . For example, if the object to be inspected is a planar structure, or if it is desired to acquire data only on the surface of the object to be inspected, the computing unit 20 may be configured to generate two-dimensional data instead of the three-dimensional data generating unit 21. may be provided. The defect database 23 accumulates three-dimensional data of a plurality of defect points in parts of the same type as the object to be inspected. In addition, regarding the data accumulated in the defect database 23,
It may be two-dimensional data.

The display unit 30 displays the three-dimensional data created by the three-dimensional data creation unit 21 (hereinafter, the three-dimensional data will be referred to as “non-defective image” in order to clearly distinguish it from the three-dimensional data of the defective part), and the defect database 23. displays the accumulated three-dimensional data of the defective portion (hereinafter referred to as "defective portion image"). In the defect database 23, images of defect locations acquired in advance in past inspections are stored as a database. By selecting a partial region of the non-defective image and the image of the defective portion on the display unit 30, the transfer unit 24 transfers the image of the defective portion to the partial region of the non-defective image as a transfer destination. As a result, a non-defective image is processed to generate a transfer image including a pseudo-defective portion. Here, the transfer means a process of cutting and pasting an image by combining a general image processing method (for example, a binarization method) so as not to look unnatural (as described in the examples below). The same is true for transcription). Further, the transfer destination means an area in which processing for cutting and pasting an image is executed (the same applies to the transfer destination described in the following examples). In the present invention, an image that has been determined to be a non-defective product in the past and an image that has been determined to be a defective product in the past are hereinafter simply referred to as a non-defective image and a defective portion image.

The display unit 30 displays this transfer image, and the measurement unit 25 measures predetermined feature amounts in the displayed transfer image and other inspection objects. After the measurement is completed, the calculator 26 automatically calculates the threshold value based on the measurement result (not shown). After completing the automatic calculation, the display unit 30 displays the measurement result and the threshold value based on the measurement result. Based on this measurement result, it is possible to confirm the frequency of occurrence of defects in the inspection object. In addition, this threshold value clarifies the criteria for distinguishing non-defective products from defective products. Further, the calculation unit 26 automatically calculates the threshold value, so that, for example, when the predetermined feature amount in the non-defective image before the measurement by the measurement unit 25 is less than the threshold value, the predetermined feature amount is measured. The image of the non-defective product is later detected as the image of the defective product.

Details of the arithmetic processing executed by the arithmetic unit 20 and the contents displayed by the display unit 30 will be described in the following embodiments with reference to FIGS. 3A, 3B, 4A, 4B, and 5. FIG.

FIG. 7 is a flow chart showing the procedure of an image inspection method using the X-ray inspection apparatus 1 to which the present invention is applicable. As for the flow, only the outline will be described in this application example, and the details will be described in the following embodiments.

In the image inspection method using the X-ray inspection apparatus 1 in this application example, as described above, the image of the non-defective product created by the three-dimensional data creation unit 21 is first displayed on the display unit 30, and then acquired from the defect database 23. The image of the defective portion is also displayed on the display unit 30 . Next, an arbitrary point is selected as a transfer destination from the non-defective image displayed on the display unit 30, and one or more images of defective portions are selected from the plurality of images of defective portions displayed on the display unit 30. do. When the transfer destination is determined, the image of the selected defective portion is transferred to the transfer destination. As a result, a non-defective image is processed to generate a transfer image including a pseudo-defective portion. Next, this transfer image is displayed on the display unit 30, the measurement unit 25 measures the predetermined feature amount of this transfer image and other inspection objects, and the calculation unit 26 automatically sets the threshold value based on the measurement result. Calculated by Finally, the measurement result and the threshold value based on the measurement result are displayed on the display unit 30 .

As described above, according to the image inspection method using the X-ray inspection apparatus 1 in this application example, no defect occurs in the manufacturing process of the inspection object, and the three-dimensional data generation unit 21 determines the defect location in the inspection object. Even in the case where the data cannot be acquired, the transfer unit 24 can generate a transfer image including a pseudo-defective portion. Furthermore, the measurement unit 25 measures a predetermined feature amount of the inspection object including the transferred image, and the calculation unit 26 automatically calculates a threshold value based on the measurement result, thereby visually and numerically showing the difference between good products and defective products. can be discriminated. It is also possible to provide a measurement result including a threshold value as a criterion for determining the difference between non-defective products and defective products.

〔Example〕
Hereinafter, the image inspection method using the X-ray inspection apparatus 1 according to the embodiment of the present invention will be described in more detail with reference to the drawings (including the drawings once described in the application example above). Note that the X-ray inspection apparatus according to the present invention is not intended to be limited to the following configuration. Also, in the embodiments, the X-ray inspection apparatus 1 is illustrated as an example of the image inspection apparatus, but the present invention is not limited to this, and other types of apparatus may be used.

<Device configuration>
Now, let us return to the description of FIG. Since the X-ray inspection apparatus 1 according to the present embodiment has the same configuration as the X-ray inspection apparatus 1 described in the application example, detailed description of the contents described in the application example will be omitted. Also, in this specification, the same components are described using the same reference numerals.

The three-dimensional data creation unit 21 can acquire from the database 22 images of non-defective products that have been acquired in past inspections and accumulated in the database 22 . Therefore, even if the inspection object is not placed on the stage 11, the transfer unit 24 can transfer the image of the defective portion to the image of the non-defective product of the inspection object. Therefore, the three-dimensional data creation unit 21 can create a transferred image as a function independent of the imaging function of the X-ray inspection apparatus 1 . Before the calculation unit 26 automatically calculates the threshold value, the image of the non-defective item and the image of the defective portion should be strictly referred to as the image of the candidate of the non-defective item and the image of the candidate of the defective portion, respectively. , hereinafter simply referred to as an image of a non-defective item and an image of a defective portion.

For example, as shown in FIG. 2, the defect database 23 stores a plurality of images of defective portions obtained in advance in past inspections. The type of product, the type of defect, and the defect level by numerical values representing the degree of defect are linked. An ID is assigned to each set of four items: an image of the defective portion, the type of the part, the type of the defect, and the level of the defect represented by the numerical value. The display unit 30 may display the four items of information. In addition, it is possible to search for one of the four items by using the filter function and the sort function. If the inspection object is, for example, a circuit board to which electronic elements are soldered, the types of defects include, for example, non-wetting and voids (in FIG. (illustrating an image of a bad pin 51 (described in detail below)). The numerical failure level is a numerical value that is determined based on the results of past measurement by the measuring unit 25 or based on the results of visual judgment in the past. Information in which the above four items are managed as one set can also be provided as detailed information related to the image of the defective portion. The management method shown in FIG. 2 corresponds to the image management method of the defective portion in the present invention.

<Image inspection method using X-ray inspection device>
3A, 3B, 4A, 4B, and 5, an example flow of contents displayed by the display unit 30 in the image inspection method using the X-ray inspection apparatus 1 according to the present embodiment will be described below. explain. The content displayed by the display unit 30 is based on the arithmetic processing executed by the arithmetic unit 20 . In the following, as an example, the inspection object is assumed to be a plurality of pins 50 in a chip 40, which is an electronic element on a circuit board.

FIG. 3A is an example of the content displayed by the display unit 30 when the three-dimensional data creation unit 21 creates an image of a non-defective chip 40 . The display unit 30 displays tomographic images of one chip 40 when viewed from the XY directions (directions in plan view), YZ directions (directions in side view), and XZ directions (directions in front view or rear view), respectively. display as In order to clarify from which direction the image is viewed, the respective images may be given notations such as "XY", "YZ", and "XZ". In the case of a non-defective product, the pin 50 has a circular shape with high roundness when viewed from the XY directions, and the pin 50 has an elliptical shape when viewed from the YZ and XZ directions. . Further, the display unit 30 displays buttons indicating commands to the calculation unit 20, such as "transfer setting", "transfer execution", and "measurement", and when the user presses these buttons, the calculation unit 20 starts calculation processing. .

FIG. 3B is an example of the content displayed by the display unit 30 when images of a plurality of defective locations (hereinafter referred to as "defective pins 51") are obtained from the defect database 23. FIG. Images of a plurality of defective pins 51 are listed and displayed when the user presses a "transfer setting" button. Images of the defective pin 51 as viewed in the XY, YZ, and XZ directions are also displayed as a set of tomographic images. Also, the defective pins 51 may be labeled with "XY", "YZ", and "XZ". Incidentally, in this embodiment, the defective pin 51 means a pin whose shape, for example, a perfect circle or an ellipse is broken.

FIG. 4A is an example of contents displayed by the display unit 30 when selecting a part of the image of the chip 40 to be transferred and the image of the defective pin 51 to be transferred. When determining the transfer destination and the transfer source, an arbitrary point is selected from the image of the chip 40 displayed in FIG. to select. Each selected portion is displayed as a black dot or the like, for example, as shown in FIG. 4A. Also, when an arbitrary point is selected from the image of the chip 40, the transfer destination is determined based on that point. Specifically, for the transfer destination pin 50, the boundary in the XY directions is automatically determined by binarization. Alternatively, after manually setting the area, the image within the area is automatically determined by binarization. Boundaries in the Z direction are set by reading boundary information manually entered in a separate file. As a result, the image portion of the chip 40 that is the transfer destination and the image of the defective pin 51 that is the transfer source corresponding to the transfer destination are determined. In addition, in the image of the chip 40 , a region may be selected so that a plurality of images of the pins 50 are included, and a plurality of images of the defective pins 51 corresponding to each of the plurality of pins 50 may be selected. Here, the transfer destination corresponds to the pin 50 as a connecting portion between the electronic element and the circuit board.

FIG. 4B is an example of contents displayed by the display unit 30 when the image of the defective pin 51 is transferred to a part of the image of the chip 40 . When the user presses the "execute transfer" button, the transfer unit 24 transfers the image of the defective pin 51 selected in FIG. 4A to part of the image of the chip 40. FIG. By this transfer, the pins 50 in the chip 40 are replaced with defective pins 51 (hereinafter, the pins 50 replaced by this transfer are referred to as "transfer pins 52"), and a transfer image is generated. In FIG. 4A, for example, when a transfer destination is selected for a portion of the image on the chip 40 viewed from the XY directions, the transfer destination is determined not only in the XY directions but also in the YZ and XZ directions. In FIG. 4B, when the transfer image is generated, the transfer pins 52 are displayed in all of the XY, YZ, and XZ directions (that is, the transfer pins 52 are displayed when viewed from the XY directions). , the transfer pins 52 when viewed from the YZ and XZ directions are also displayed corresponding to the transfer pins 52). The transfer pins 52 are displayed by enclosing them in a square, for example, as shown in FIG. 4B. Further, as described above, in the image of the chip 40 in FIG. 4A , an area is selected so that a plurality of images of the pins 50 are included, and a plurality of images of the defective pins 51 corresponding to each of the plurality of pins 50 are generated. In this case, a plurality of defective pins 51 can be transferred to obtain images of more transfer pins 52 (that is, images of defective portions). Details of the flow of transfer shown in FIG. 4B are shown in FIGS. 6A, 6B, and 6C below.

FIG. 5 shows the display unit 30 when the transfer image and the predetermined feature amount of the other chip 40 are measured.
is an example of the content displayed by . The transfer image has pins 50 and transfer pins 52 , and when the user presses the “measurement” button, the measurement unit 25 measures all the pins 50 and transfer pins 52 on the transfer image and on other chips 40 . Predetermined feature amounts are measured for all pins 50 and defective pins 51 . When the measurement is executed, the threshold value is automatically calculated based on the measurement result 60, and the display unit 30 displays the measurement result 60 and the threshold value (dotted line shown in the measurement result 60 in FIG. 5). The measurement result 60 is displayed as a histogram, for example, as shown in FIG. 5, with the horizontal axis representing the measured value and the vertical axis representing the frequency. Further, the predetermined feature amount is, for example, the roundness or area of the surfaces of the pin 50, the defective pin 51, and the transfer pin 52. FIG. The chips 40 are sorted into non-defective products and non-defective products based on the threshold values of the measurement results. Here, the defective product in the measurement result 60 can be displayed so that the transferred image and the other chip 40 including the defective pin 51 can be identified (for example, displayed in a different color). good. In addition, as an image inspection method, for example, an intermediate value between the average value of a predetermined feature value for a good product and the average value of a predetermined feature value for a defective product, or a predetermined value for an image that has been determined as a defective product in the past and a transferred image. The maximum value of the feature amount may be calculated as the threshold.

6A, 6B, and 6C are supplementary diagrams explaining in detail the transfer flow shown in FIG. 4B. Note that FIGS. 6A, 6B, and 6C are diagrams schematically showing the flow of transfer, and are not an example of what the display unit 30 actually displays. As shown in FIG. 6A, the transfer unit 24 transfers the image of the defective pin 51 selected in FIG. 4A to the transfer destination in the image of the chip 40, but as shown in FIG. , to be distinguished from the defective pin 51) is different from the resolution of the image on the transfer destination, an unnatural transfer image is generated after the transfer. Note that, in this embodiment, that the two images have different resolutions means that the two images have different sizes. Specifically, the size of the image on the transfer pin 52 is many times larger than the size of the image on the transfer destination, so the resolution of the image on the transfer pin 52 is many times larger than the resolution of the image on the transfer destination. is also expensive. Therefore, as shown in FIG. 6C, the size of the image of the defective pin 51 (for example, 60×60 pixels) is compressed before transfer to bring it closer to the size of the image on the transfer destination (for example, 20×20 pixels). As a result, the resolutions of the two become close to each other, and a natural transferred image (an image including defects that can actually occur in the process) is generated after transfer. As another example of the method for reducing the resolution of a high-resolution image, when the resolution of a defective image is high, the resolution may be reduced by thinning out the pixels of the defective image and regenerating the image.

<Flowchart>
Now, return to the description of FIG. The procedure of the image inspection method using the X-ray inspection apparatus 1 according to this embodiment will be described in detail below with reference to FIG. In this flowchart, first, the three-dimensional data creation unit 21 creates an image of the tip 40 when viewed from the XY, YZ, and XZ directions, and the display unit 30 displays the image of the tip 40 as a tomographic image. (S101). Here, S101 corresponds to the image display step in the present invention and FIG. 3A in this embodiment. Furthermore, the display unit 30 also displays images of a plurality of defective pins 51 viewed from the XY, YZ, and XZ directions acquired from the defect database 23 side by side with the image of the chip 40 (S102). Here, S102 corresponds to the defective image display step in the present invention and FIG. 3B in this embodiment. Next, an arbitrary point is selected from the image of the chip 40 displayed on the display unit 30, and one or more images of the defective pin 51 are selected from the images of the defective pins 51 corresponding to the arbitrary point. When an arbitrary point is selected from the image of the chip 40, the transfer destination is determined by determining the boundary in the XY direction and the boundary in the Z direction of the pin 50 based on the arbitrary point (S103). Here, S103 corresponds to the selection step in the present invention and FIG. 4A in this embodiment. When the transfer destination is determined, the image of the selected defective pin 51 is transferred to the image of the transfer destination. As a result, the image of the chip 40 is processed, and a transferred image including the transfer pin 52 as a pseudo defective portion is generated (S104). As described above, when a transfer image is generated, the resolution of the image of the defective pin 51 selected in S103 before transfer is brought close to the resolution of the transfer destination. Here, S104 corresponds to the transfer step in the present invention and FIG. 4B in this embodiment. The display unit 30 displays the transfer image (S105), and the measurement unit 25 measures all the pins 50 and transfer pins 52 on the transfer image and all the pins 50 and the defective pins 51 on the other chip 40. For example, the circularity and area of those surfaces are measured (S106). Here, S105 corresponds to the transferred image display step in the present invention and FIG. 5 in this embodiment. S106 also corresponds to FIG. 5 in this embodiment. The calculator 26 automatically calculates the threshold based on the measurement result 60 obtained in S106 (S107). Here, S107 also corresponds to FIG. 5 in this embodiment. The measurement result 60, including the threshold value obtained in S107, is displayed on the display unit 30, for example, as a histogram (S108). Here, S108 also corresponds to FIG. 5 in this embodiment.

<Appendix 1>
An image inspection method that enables determination of good or bad in an inspection object (40) using an image inspection device (1),
an image display step (S101) of displaying an image (50) that has been determined to be non-defective in the past;
a defective image display step (S102) for displaying an image (51) that has been determined to be defective in the past;
a selection step (S103) of selecting a transfer destination of the image that has been determined to be non-defective in the past, and one or more defective images of the image that has been determined to be defective in the past corresponding to the transfer destination; ,
Transferring one or more images of defective portions from the image selected in the selection step to the transfer destination of the image determined to be non-defective in the past to generate a transfer image. and a transfer step (S104),
In the transfer step, when the image of the defective portion is transferred to the transfer destination, if the resolution of the image of the defective portion is higher than the resolution of the transfer destination, the resolution of the image of the defective portion is set to a low resolution. an image inspection method, characterized in that the resolution of the image of the defective portion is brought close to the resolution of the transfer destination by converting the image.

<Appendix 2>
An image inspection device (1) capable of determining whether an inspection object (40) is good or bad,
a database (22) that accumulates data of images (50) that have been determined to be non-defective in the past;
a defect database (23) that accumulates data of images (51) that have been determined to be defective in the past;
One or more defective portions of the image determined as defective in the past obtained from the defective database corresponding to the transfer destination in the image determined as non-defective in the past obtained from the database. a transfer unit (24) for transferring the image of and generating a transferred image;
a display unit (30) for displaying the image previously determined as a non-defective product obtained from the database and the image previously determined as a defective product obtained from the defective database;
with
When transferring the image of the defective portion to the transfer destination, the transfer unit reduces the resolution of the image of the defective portion to a low resolution when the resolution of the image of the defective portion is higher than the resolution of the transfer destination. An image inspection apparatus (1), characterized in that the resolution of the image of the defective portion is brought close to the resolution of the transfer destination by converting the image.

Reference Signs List 1: X-ray inspection apparatus 10: Imaging unit 11: Stage 12: Imaging condition storage unit 13: X-ray generator 14: X-ray detector 20: Calculation unit 21: Three-dimensional data creation unit 22: Database 23: Defect database 24 : transfer unit 25 : measurement unit 26 : calculation unit 30 : display unit 40 : chip 50 : pin 51 : defective pin 52 : transfer pin 60 : measurement result

Claims (6)

An image inspection method that makes it possible to determine whether an object to be inspected is good or bad using an image inspection device,
an image display step of displaying an image that has been determined to be non-defective in the past;
a defective image display step of displaying an image that has been determined to be defective in the past;
a selection step of selecting a transfer destination of the image that was determined to be non-defective in the past, and an image of one or more defective portions of the image that was determined to be defective in the past corresponding to the transfer destination;
Transferring one or more images of defective portions from the image selected in the selection step to the transfer destination of the image determined to be non-defective in the past to generate a transfer image. and a transfer step of
In the transfer step, when the image of the defective portion is transferred to the transfer destination, if the resolution of the image of the defective portion is higher than the resolution of the transfer destination, the resolution of the image of the defective portion is set to a low resolution. an image inspection method, characterized in that the resolution of the image of the defective portion is brought close to the resolution of the transfer destination by converting the image. 2. The image inspection method according to claim 1, further comprising a transferred image display step of displaying said transferred image. 2. In said transferring step, the resolution of said image of said defective portion can be reduced by compressing the size of said image of said defective portion so as to be equal to the size of said transfer destination. Or the image inspection method according to 2. An image inspection device capable of determining whether an object to be inspected is good or bad,
a database that accumulates data of images that have been determined to be non-defective in the past;
a defect database that accumulates image data that has been determined to be defective in the past;
One or more defective portions of the image determined as defective in the past obtained from the defective database corresponding to the transfer destination in the image determined as non-defective in the past obtained from the database. a transfer unit that transfers the image of and generates a transferred image;
a display unit that displays the image that was previously determined as a non-defective product that was acquired from the database, and the image that was determined as a defective product in the past, which was acquired from the defective database;
with
When transferring the image of the defective portion to the transfer destination, the transfer unit reduces the resolution of the image of the defective portion to a low resolution when the resolution of the image of the defective portion is higher than the resolution of the transfer destination. an image inspection apparatus, wherein the resolution of the image of the defective portion is brought close to the resolution of the transfer destination by converting the image. 5. The image inspection apparatus according to claim 4, wherein the display section displays the transferred image obtained from the transfer section. 5. The transfer unit is capable of reducing the resolution of the image of the defective portion by compressing the size of the image of the defective portion so as to be equal to the size of the transfer destination. 6. The image inspection apparatus according to 5.

Images