JP2018059735A - Lamination boundary detection apparatus, lamination boundary detection method, lamination boundary detection program, and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a lamination boundary detection apparatus, a lamination boundary detection method, a lamination boundary detection program, and a recording medium therefor which are resistant to noise fluctuation and which are not required for a correlation analysis.SOLUTION: A lamination boundary detecting apparatus Ma of the present invention comprises an image acquiring unit 11 for acquiring an image data of a cross section of the lamination; a brightness projection processing unit 122 for obtaining a brightness projection along a laminate direction in an orthogonal direction orthogonal to the laminate direction with respect to the acquired image data; a differential absolute value processing unit 123 for obtaining an absolute value of a differential value as a differential absolute value with respect to the obtained brightness projection; a noise elimination processing unit 124 for extracting a differential absolute value exceeding a predetermined ratio of the maximum value of the differential absolute value or a large differential absolute value from the obtained differential absolute value; and a boundary extraction processing unit 125a for determining the position of a boundary surface of the lamination in the cross section on the basis of the extracted differential absolute value.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a laminated body boundary detecting device, a laminated body boundary detecting method, a laminated body boundary detecting program, and a recording medium thereof for detecting a boundary between the different materials in a laminated body that is a member in which different materials are laminated.

  It occupies a cross section including the stacking direction intersecting the boundary surface in a laminate of metal members obtained by laminating a plurality of different types of metal materials or a laminate of resin members obtained by laminating a plurality of different types of resin materials. A technique for measuring the ratio of each layer (cladding rate, stacking rate, (thickness of the layer) / (total thickness of each layer) × 100 [%]) has been developed, and is disclosed in Patent Document 1, for example.

  The cladding ratio measuring apparatus disclosed in Patent Document 1 is based on image acquisition means for capturing a cross-sectional image related to a measurement target member in which a cladding layer is formed on the surface of a core material, and a change in pixel value in the thickness direction of the cross-sectional image. Boundary determining means for obtaining two opposing boundaries of the cladding layer, and calculating means for calculating a cladding ratio of the cladding layer by calculating a distance between the boundaries. The boundary determination unit converts the cross-sectional image into a binary image based on a comparison result between each pixel value of the cross-sectional image and a set value, and a boundary between one pixel region and the other pixel region of the binary image. Is determined as the boundary on the outer edge side of the cladding layer.

  As described above, the cladding ratio measuring apparatus disclosed in Patent Document 1 binarizes an image with pixel values (luminance values) and recognizes the boundary between the core material and the cladding layer. However, in a binarized binary image, for example, the boundary is often not clearly shown due to, for example, luminance variations, and therefore it is difficult to recognize the boundary based on the binary image. In the binary image of the image obtained by imaging the alloy layer, adverse effects such as recognizing the boundary of precipitates due to the influence of precipitates contained in the alloy layer occur, so it corresponds to a complex matrix in the alloy layer etc. A boundary recognition method is required. For this reason, a boundary detection method is proposed in Patent Document 2.

  The boundary detection method disclosed in Patent Document 2 is a boundary detection method for detecting a stacked boundary that is a boundary of the plurality of different materials with respect to a stacked material in which a plurality of different materials are stacked. A feature amount that expresses a variation in luminance due to a difference in particle size distribution of the laminated material is detected from an imaging step of capturing an image of the laminated material including a boundary and a luminance distribution of the image captured in the imaging step. A feature amount detection step; and a layer boundary detection step of detecting the layer boundary in the captured image based on the feature amount detected in the feature amount detection step. The feature amount detection step detects, as the feature amount, a standard deviation and / or peak count value of the luminance distribution in a direction along the stacking boundary, and a direction in which the standard deviation and / or peak count value is detected. The similarity of the feature amounts is obtained in the vertical direction, and the stack boundary is detected based on the obtained difference in similarity.

  “A and / or B” means “at least one of A and B”.

JP 2007-178375 A Japanese Unexamined Patent Publication No. 2015-200543

  By the way, the boundary detection method disclosed in Patent Document 2 uses a standard deviation and / or a statistic of a peak count value as a feature quantity expressing luminance variation. However, although it is strong against noise fluctuations in the image due to the statistical feature amount, the specific position and coordinates such as boundary detection and the statistical feature amount are not directly linked. For this reason, in order to use it practically, it is necessary to collect data of these statistical feature values and the actual position and coordinates of the boundary line and perform correlation analysis. For this reason, since it is not intuitive, it takes time to set correlation analysis and appropriate feature amount calculation, and a setter having specialized techniques is required.

  The present invention is an invention made in view of the above-described circumstances, and its object is a laminate boundary detection device, a laminate boundary detection method, a laminate boundary detection program that is resistant to noise fluctuation and does not require correlation analysis, and It is to provide the recording medium.

  As a result of various studies, the present inventor has found that the above object is achieved by the present invention described below. That is, the laminate boundary detection device according to one aspect of the present invention is an image acquisition unit that acquires image data of a cross section including a stacking direction intersecting a boundary surface in a stack that is a member in which a plurality of different materials are stacked. A luminance projection processing unit that performs luminance projection processing for obtaining a luminance projection along the stacking direction in an orthogonal direction orthogonal to the stacking direction with respect to the image data acquired by the image acquisition unit, and the luminance projection processing unit. A differential absolute value processing unit that obtains a differential value for the luminance projection obtained along the stacking direction and performs a differential absolute value process for obtaining an absolute value of the obtained differential value as a differential absolute value, and the differential absolute value processing unit The differential absolute value that is equal to or greater than a predetermined ratio with respect to the maximum value of the differential absolute value from the differential absolute value obtained in A noise removal processing unit that performs noise removal processing that extracts a differential absolute value that is larger than a predetermined ratio with respect to a maximum absolute value, and a differential absolute value extracted by the noise removal processing unit, based on the absolute value of the boundary surface in the cross section And a boundary extraction processing unit that performs a boundary extraction process for obtaining a position. Preferably, in the above-described stacked body boundary detection device, the image acquisition unit is a camera that captures a cross section of the stacked body. Preferably, in the above-described laminate boundary detection device, the image acquisition unit is a communication interface unit that communicates with a server device that manages the image data. Preferably, in the above-described laminated body boundary detection device, the image acquisition unit is an interface unit that exchanges (inputs / outputs) data with a recording medium that records the image data.

  In general, since the luminance (reflectance) varies depending on the material, the differential value for the luminance projection in the orthogonal direction obtained from the cross-sectional image along the stacking direction is the boundary line that appears in the cross-section based on the boundary surface. It changes a lot. The laminate boundary detection apparatus performs a noise removal process for extracting a differential absolute value that is equal to or greater than a predetermined ratio with respect to the maximum value of the differential absolute value or a differential absolute value that is greater than a predetermined ratio with respect to the maximum value of the differential absolute value. Since the boundary extraction processing for obtaining the position of the boundary surface in the cross section is performed based on the differential absolute value after the noise removal processing, the above-described correlation analysis is not required.

  According to another aspect, in the above-described laminate boundary detection device, the boundary extraction processing unit performs a predetermined range (predetermined) on the differential absolute value extracted by the noise removal processing unit as the boundary extraction processing. Peaks located within the first range) are grouped as a peak group, and the barycentric position of the peak within the grouped peak group is obtained as the position of the boundary surface. Preferably, the above-described laminated body boundary detection device further includes a range input unit that receives the predetermined range.

  Depending on the laminate to be measured, the boundary line that appears in the cross section based on the boundary surface is unclear. Therefore, in the differential absolute value extracted by the noise removal process, there are multiple points near (in the vicinity of) the actual boundary surface position. Peak may occur. The laminated body boundary detection apparatus obtains the position of the center of gravity of the peak in the peak group as the position of the boundary surface. Even in such a case, the position of the boundary surface in the cross section can be obtained more appropriately.

  According to another aspect, in the above-described laminate boundary detection device, the boundary extraction processing unit performs a predetermined range (predetermined) on the differential absolute value extracted by the noise removal processing unit as the boundary extraction processing. Peaks located within the first range) are grouped as a peak group, and the position of the maximum peak in the grouped peak group is obtained as the position of the boundary surface.

  Depending on the laminate to be measured, as described above, the boundary line that appears in the cross section based on the boundary surface is unclear, so in the differential absolute value extracted by the noise removal process, the vicinity at the actual boundary surface position A plurality of peaks may occur at (around). Since the laminate boundary detection apparatus obtains the position of the maximum peak in the peak group as the position of the boundary surface, even in such a case, the most likely position can be obtained as the position of the boundary surface in the cross section. .

  In another aspect, in the laminate boundary detection device described above, the boundary extraction processing unit, as the boundary extraction processing, is within a predetermined second range with respect to the differential absolute value extracted by the noise removal processing unit. The peaks located in the group are collected as a peak group, and the average position of the position of the maximum peak and the position of the next maximum peak in the collected peak group is obtained as the position of the boundary surface.

  Depending on the laminate to be measured, the boundary line appearing in the cross section may be wavy based on the boundary surface. In such a case, in the differential absolute value extracted by the noise removal processing unit, a plurality of peaks occur within a predetermined second range corresponding to the degree of undulation. The laminate boundary detection apparatus obtains the average position of the position of the maximum peak (first largest peak) and the position of the next maximum peak (second largest peak) in the peak group as the position of the boundary surface. Even in such a case, the most probable position can be obtained as the position of the boundary surface in the cross section.

  Here, as described above, a plurality of peaks that appear unclear in the boundary line are concentrated in the vicinity of the boundary line, whereas a plurality of peaks that appear in the boundary line undulation depend on the degree of the undulation. The predetermined second range is wider than the predetermined range (predetermined first range) (the predetermined first range is narrower than the predetermined second range (the predetermined second range)> ( The predetermined first range)).

  Further, in another aspect, in the above-described laminate boundary detection device, the cross section centered on a predetermined position in the cross section image until the number of positions of the boundary surface reaches a predetermined number set in advance. , The luminance projection processing unit performs the luminance projection processing, the differential absolute value processing unit performs differential absolute value processing, and the noise removal processing unit performs the noise removal processing. And a rotation processing unit that causes the boundary extraction processing unit to perform the boundary extraction processing. Preferably, the laminated body boundary detection device further includes an input unit that receives the predetermined number. Preferably, the above-described laminated body boundary detection device further includes a storage unit that receives the predetermined number.

  Depending on the image of the cross section, the boundary line that appears in the cross section based on the boundary surface may not be coincident with the horizontal direction of the image and is inclined with respect to the horizontal direction of the image. The laminated body boundary detection device is configured to detect the cross section of the cross section around a predetermined position (for example, a center position) in the cross section image until the number of positions of the boundary surface in the cross section reaches a predetermined number. Since the position of the boundary surface in the cross section is obtained while rotating the image by a predetermined angle, the position of the boundary surface in the cross section can be obtained even in such a case.

  The laminate boundary detection method according to another aspect of the present invention is an image acquisition step of acquiring cross-sectional image data including a stacking direction intersecting a boundary surface in a laminate that is a member in which a plurality of different materials are stacked. A luminance projection processing step for obtaining a luminance projection along the stacking direction in an orthogonal direction orthogonal to the stacking direction with respect to the image data acquired in the image acquisition step, and along the stacking direction in the luminance projection processing step. A differential absolute value processing step for obtaining a differential value for the luminance projection obtained in the above, a differential absolute value processing step for obtaining the absolute value of the obtained differential value as a differential absolute value, and the differential absolute value obtained in the differential absolute value processing step, A differential absolute value that is greater than or equal to a predetermined percentage of the maximum value of the value or greater than a predetermined percentage of the maximum value of the differential absolute value Min and absolute value noise removing process step of extracting the noise removal processing based on the extracted differential absolute value in step, characterized in that it comprises a boundary extraction processing step of determining the position of the boundary surface in the cross section.

  A laminate boundary detection program according to another embodiment of the present invention acquires, in a computer, image data of a cross section including a stacking direction intersecting a boundary surface in a laminate that is a member in which a plurality of different materials are stacked. An image acquisition step, a luminance projection processing step for obtaining a luminance projection along the stacking direction in a direction orthogonal to the stacking direction with respect to the image data acquired in the image acquisition step, and the stacking in the brightness projection processing step. A differential value is obtained for the luminance projection obtained along the direction, a differential absolute value processing step for obtaining the absolute value of the obtained differential value as a differential absolute value, and a differential absolute value obtained in the differential absolute value processing step, The differential absolute value that is equal to or greater than a predetermined ratio with respect to the maximum value of the differential absolute value or the maximum value of the differential absolute value A noise removal processing step for extracting a differential absolute value larger than a predetermined ratio, and a boundary extraction processing step for obtaining the position of the boundary surface in the cross section based on the differential absolute value extracted in the noise removal processing step. It is a program to make it.

  Such a laminate boundary detection method and a laminate boundary detection program have a differential absolute value that is greater than or equal to a predetermined ratio with respect to the maximum value of the differential absolute value or a differential absolute value that is greater than a predetermined ratio with respect to the maximum value of the differential absolute value. Since the noise removal processing step for extracting is performed, the boundary extraction processing step for obtaining the position of the boundary surface in the cross section is performed based on the absolute value of the differential after the noise removal processing step because of being resistant to noise fluctuations. There is no need for correlation analysis.

  The recording medium according to another aspect of the present invention is an image acquisition step of acquiring, on a computer, image data of a cross section including a stacking direction intersecting a boundary surface in a stacked body that is a member in which a plurality of different materials are stacked. A luminance projection processing step for obtaining a luminance projection along the stacking direction in an orthogonal direction orthogonal to the stacking direction with respect to the image data acquired in the image acquisition step, and along the stacking direction in the luminance projection processing step. A differential absolute value processing step for obtaining a differential value for the luminance projection obtained in the above, a differential absolute value processing step for obtaining the absolute value of the obtained differential value as a differential absolute value, and the differential absolute value obtained in the differential absolute value processing step, The differential absolute value that is equal to or greater than a predetermined ratio to the maximum value of the value or a predetermined ratio with respect to the maximum value of the differential absolute value A noise removal processing step for extracting a differential absolute value, and a boundary extraction processing step for obtaining a position of the boundary surface in the cross section based on the differential absolute value extracted in the noise removal processing step. It is a computer-readable recording medium on which a laminate boundary detection program is recorded.

  According to this, it is possible to provide a computer-readable recording medium that records the above-described laminate boundary detection program that is resistant to noise fluctuation and does not require correlation analysis.

  The laminate boundary detection apparatus, laminate boundary detection method, and laminate boundary detection program according to the present invention are resistant to noise fluctuation and do not require correlation analysis. According to the present invention, it is possible to provide a computer-readable recording medium that records the above-described laminate boundary detection program that is resistant to noise fluctuation and does not require correlation analysis.

It is a figure which shows the structure of the laminated body boundary detection apparatus in 1st thru | or 3rd embodiment. It is a flowchart which shows operation | movement of the laminated body boundary detection apparatus in 1st and 2nd embodiment. In the laminated body boundary detection apparatus of 1st Embodiment, it is a figure for demonstrating an example of a boundary detection. It is a figure for demonstrating another example of a boundary detection in the laminated body boundary detection apparatus of 1st Embodiment. It is a figure for demonstrating an example of a boundary detection in the laminated body boundary detection apparatus of 2nd Embodiment. It is a flowchart which shows operation | movement of the laminated body boundary detection apparatus in 3rd Embodiment. It is a figure for demonstrating an example of a boundary detection in the laminated body boundary detection apparatus of 3rd Embodiment.

  Hereinafter, an embodiment according to the present invention will be described with reference to the drawings. In addition, the structure which attached | subjected the same code | symbol in each figure shows that it is the same structure, The description is abbreviate | omitted suitably. In the present specification, when referring generically, it is indicated by a reference symbol without a suffix, and when referring to an individual configuration, it is indicated by a reference symbol with a suffix.

  The laminated body boundary detection device according to the embodiment is the position of the boundary surface in the cross section of the laminated body that is a member in which a plurality of different materials are laminated, that is, the boundary line that appears in the cross section of the laminated body based on the boundary surface. It is a device that detects the position. The material constituting the laminate is arbitrary, and is, for example, a metal (including an alloy) or a resin. The laminate is, for example, a member obtained by laminating a plurality of types of metals (including alloys), a member obtained by laminating a plurality of types of resins, a composite member (composite member) obtained by laminating a resin on a metal, or the like. Such a laminate boundary detection device, for example, in the laminate, an image acquisition unit that acquires image data of a cross section including a stacking direction intersecting the boundary surface, and the image data acquired by the image acquisition unit, A luminance projection processing unit for performing a luminance projection process for obtaining a luminance projection in the orthogonal direction perpendicular to the stacking direction along the stacking direction; and a differential value for the luminance projection obtained in the stacking direction by the luminance projection processing unit. A differential absolute value processing unit for performing differential absolute value processing to obtain an absolute value of the obtained differential value as a differential absolute value, and a maximum value of the differential absolute value from the differential absolute value obtained by the differential absolute value processing unit. A differential absolute value that is greater than or equal to a predetermined ratio with respect to or a differential absolute value greater than a predetermined ratio with respect to a maximum value of the differential absolute value A noise removal processing unit that performs noise removal processing to be output; and a boundary extraction processing unit that performs boundary extraction processing for obtaining the position of the boundary surface in the cross section based on the differential absolute value extracted by the noise removal processing unit. . Hereinafter, such a laminated body boundary detection apparatus will be described more specifically with reference to the first to third embodiments.

(First embodiment)
FIG. 1 is a diagram illustrating a configuration of a stacked body boundary detection device according to the first to third embodiments. In FIG. 1, configurations common to the stacked body boundary detection devices Ma, Mb, and Mc in the first to third embodiments are indicated by reference numerals without subscripts, and the stacked body boundary detection device in the first embodiment. The configuration relating to Ma is indicated by a symbol having a subscript a, the configuration relating to the stacked body boundary detection device Mb in the second embodiment is indicated by a symbol having a subscript b, and the third The configuration related to the stack boundary detection device Mc in the embodiment is indicated by a symbol having a subscript c.

  In FIG. 1, the laminated body boundary detection apparatus Ma in 1st Embodiment is provided with the image acquisition part 11, the control process part 12a, the memory | storage part 13, the input part 14, and the output part 15, for example.

  The image acquisition unit 11 is connected to the control processing unit 12a, acquires cross-sectional image data including the stacking direction in the stacked body according to the control of the control processing unit 12a, and outputs the acquired image data to the control processing unit 12a. Device.

  Such an image acquisition unit 11 includes, for example, a camera that captures a cross section of the stacked body. The camera is, for example, an imaging optical system that forms an optical image of an imaging target on a predetermined imaging surface, and a light receiving surface that is aligned with the imaging surface, and the optical image of the imaging target is electrically For example, a CCD (Charged Coupled Device) type or a CMOS (Complementary Metal Oxide Semiconductor) type area image sensor that converts to an image signal and data representing the image to be imaged by processing the output of the area image sensor. An image processing unit for generating image data is provided. In the present embodiment, the imaging target is a cross section of the stacked body obtained by cutting the stacked body along the stacking direction. The cross-sectional image data of the laminate is acquired by, for example, the camera attached to a microscope that can observe the enlarged cross-section of the laminate.

  Further, for example, the image acquisition unit 11 is a communication interface unit (communication IF unit) that communicates with a server device that manages image data of a cross section of the laminate via a communication network. The communication IF unit generates a communication signal containing data to be transferred input from the control processing unit 12a in accordance with a communication protocol used in the communication network, and generates the generated communication signal via the communication network. Send to server device. The communication IF unit receives a communication signal from the server device via the communication network, extracts data from the received communication signal, and converts the extracted data into data that can be processed by the control processing unit 12a. And output to the control processing unit 12a. The communication IF unit 21 includes, for example, a communication card according to the IEEE 802.11 standard or the like.

  For example, the image acquisition unit 11 is an interface unit (IF unit) that exchanges (inputs / outputs) data with a recording medium that records (stores) image data of a cross section of the laminate. The recording medium is, for example, a USB memory, and the IF unit is an interface circuit using a USB (Universal Serial Bus) standard. Further, for example, the recording medium may be an SD card (miniSD card, microSD card), and the IF unit is an interface circuit using the SD standard. For example, the recording medium may be a CD-R (Compact Disc Recordable), a DVD-RW (Digital Versatile Disc Read Write), or the like, and the IF unit is a CD drive device, a DVD drive device, or the like.

  The image acquisition unit 11 may include a plurality of these cameras, communication IF units, and IF units.

  The input unit 14 is connected to the control processing unit 12a and, for example, various commands such as a command for instructing start of boundary detection, and a peak range (predetermined first range) to be collected as a peak group, for example, and a boundary detection target Is a device for inputting various data necessary for detecting a boundary, such as an identifier in the stacked body, to the stacked body boundary detecting device Ma, for example, a plurality of input switches assigned with predetermined functions, a keyboard, Such as a mouse. The output unit 15 is connected to the control processing unit 12a. Under the control of the control processing unit 12a, the command and data input from the input unit 14, the cross-sectional image of the stacked body acquired by the image acquiring unit 11, and the stacking It is a device that outputs the position of the boundary surface (position of the boundary line) in the cross section of the laminate detected by the body boundary detection device Ma, such as a display device such as a CRT display, LCD (liquid crystal display), and organic EL display, A printing device such as a printer.

  A touch panel may be configured from the input unit 14 and the output unit 15. In the case of configuring this touch panel, the input unit 14 is a position input device that detects and inputs an operation position such as a resistive film method or a capacitance method, and the output unit 15 is a display device. In this touch panel, a position input device is provided on the display surface of the display device, one or more input content candidates that can be input to the display device are displayed, and the user touches the display position where the input content to be input is displayed. The position is detected by the position input device, and the display content displayed at the detected position is input to the laminate boundary detection device Ma as the operation input content of the user. In such a touch panel, since the user can easily understand the input operation intuitively, the multilayer boundary detection device Ma that is easy for the user to handle is provided.

  The storage unit 13 is a circuit that is connected to the control processing unit 12a and stores various predetermined programs and various predetermined data in accordance with the control of the control processing unit 12a. The various predetermined programs include, for example, control programs for controlling the units 11 and 13 to 15 of the laminate boundary detection device Ma according to functions of the units, and image data acquired by the image acquisition unit 11. A luminance projection processing program for performing a luminance projection process for obtaining a luminance projection in the orthogonal direction perpendicular to the stacking direction along the stacking direction, and a derivative of the brightness projection obtained along the stacking direction by the luminance projection processing program. A differential absolute value processing program for performing differential absolute value processing for obtaining a value and obtaining the absolute value of the obtained differential value as a differential absolute value, or from the differential absolute value obtained by the differential absolute value processing program, The differential absolute value that is greater than or equal to a predetermined ratio to the maximum value or the maximum of the differential absolute value A noise removal processing program for performing noise removal processing for extracting a differential absolute value larger than a predetermined ratio with respect to a value, or a boundary for obtaining the position of the boundary surface in the cross section based on the differential absolute value extracted by the noise removal processing program A control processing program such as a boundary extraction processing program for performing the extraction processing is included. In the various predetermined data, for example, the position of the boundary surface (the position of the boundary line) in the cross section of the laminate, such as the predetermined ratio used in the noise removal processing program and the identifier to be detected, is obtained. The necessary data is included. Such a storage unit 13 includes, for example, a ROM (Read Only Memory) that is a nonvolatile storage element, an EEPROM (Electrically Erasable Programmable Read Only Memory) that is a rewritable nonvolatile storage element, and the like. The storage unit 13 includes a RAM (Random Access Memory) that serves as a working memory of a so-called control processing unit 12a that stores data generated during execution of the predetermined program.

  The control processing unit 12a controls each unit 11, 13 to 15 of the laminate boundary detection device Ma according to the function of each unit, and obtains the position of the boundary surface (position of the boundary line) in the cross section of the laminate. Circuit. The control processing unit 12a includes, for example, a CPU (Central Processing Unit) and its peripheral circuits. The control processing unit 12a functionally includes a control unit 121, a luminance projection processing unit 122, a differential absolute value processing unit 123, a noise removal processing unit 124, and a boundary extraction processing unit 125a by executing the control processing program.

  The control part 121 controls each part 11, 13-15 of the said laminated body boundary detection apparatus Ma according to the function of the said each part.

  The luminance projection processing unit 122 performs the luminance projection processing. The differential absolute value processing unit 123 performs the differential absolute value processing. The noise removal processing unit 124 performs the noise removal processing.

  The boundary extraction processing unit 125a performs the boundary extraction processing. More specifically, the boundary extraction processing unit 125a in the first embodiment peaks a peak located within a predetermined first range with respect to the differential absolute value extracted by the noise removal processing unit 124 as the boundary extraction processing. As a group, the position of the center of gravity of the peak in the group of collected peaks is obtained as the position of the boundary surface (position of the boundary line).

  Note that when the image acquisition unit 11 is a camera, the stacked body boundary detection device Ma, as indicated by a broken line in FIG. Further, an IF unit 17 or the like for exchanging (input / output) data with an external device may be further provided.

  Such a laminated body boundary detection device Ma can be configured by, for example, a desktop or notebook computer including the image acquisition unit 11.

  Next, the operation of the stacked body boundary detection apparatus in the first embodiment will be described. FIG. 2 is a flowchart showing the operation of the stacked body boundary detection device according to the first and second embodiments. In FIG. 2, operations common to the multilayer boundary detection devices Ma and Mb in the first and second embodiments are indicated by reference numerals without subscripts, and relate to the multilayer boundary detection device Ma in the first embodiment. The operation is indicated by a symbol having a subscript a, and the operation related to the stacked body boundary detection device Mb in the second embodiment is indicated by a symbol having a subscript b. FIG. 3 is a diagram for explaining an example of boundary detection in the stacked body boundary detection device according to the first embodiment. FIG. 3A schematically illustrates a cross-sectional image of the stack, FIG. 3B illustrates the luminance projection in the orthogonal direction x determined along the stack direction y of the stack, and FIG. The absolute value of the differential value (absolute value differential value) obtained with respect to the luminance projection in the orthogonal direction x obtained along the stacking direction y (FIG. 3B) is shown. FIG. 4 is a diagram for explaining another example of the boundary detection in the stacked body boundary detection apparatus of the first embodiment. FIG. 4A shows an image of a cross section of the laminate, and FIG. 4B shows the absolute value of the differential value obtained for the luminance projection (not shown) in the orthogonal direction x obtained along the lamination direction y. (Absolute differential value) is shown. 3 and 4, the stacking direction y is the vertical direction y in the cross-sectional image of the stacked body, and the orthogonal direction x in the cross-sectional image of the stacked body is the horizontal direction x. The same applies to the examples shown in FIGS.

  In the laminated body boundary detection device Ma having such a configuration, first, when a power switch (not shown) is turned on, the laminated body boundary detection device Ma is activated, and necessary parts are initialized by the control processing unit 12a. The control processing unit 12a is functionally controlled by executing a control processing program, so that the control unit 121, the luminance projection processing unit 122, the differential absolute value processing unit 123, the noise removal processing unit 124, and the boundary extraction processing unit 125a are functionally performed. Is configured.

  When the start of measurement is instructed from the input unit 14, in FIG. 2, the laminate boundary detection device Ma acquires the image data of the cross section of the laminate by the image acquisition unit 11 under the control of the control unit 121 of the control processing unit 12 a. Then, the acquired image data of the cross section of the laminate is stored in the storage unit 13 and output to the output unit 15 (S11). The output unit 15 outputs (for example, displays) an image of a cross section of the laminate.

  More specifically, for example, in the case where the image acquisition unit 11 is configured to include the camera, the stack is set so that the cross section can be observed on a microscope, and the control unit 121 is attached to the microscope. The cross section of the laminate is imaged by a camera to generate image data of the cross section of the laminate, and the generated image data of the cross section of the laminate is acquired from the camera.

  For example, when the image acquisition unit 11 includes the communication IF unit, the control unit 121 uses the communication IF unit to transmit a communication signal for requesting image data of a cross section of the laminate to the server device. And the communication signal of the reply is received from a server apparatus, The image data of the cross section of the said laminated body accommodated in this received said communication signal is acquired from the said communication IF part. In the server device, a plurality of cross-sectional image data of the laminate is stored (recorded) and managed in advance.

  For example, when the image acquisition unit 11 is configured to include the IF unit, the control unit 121 exchanges data with a recording medium by the IF unit, and records the image data of the cross section of the stacked body. Obtained from the medium via the IF unit. It is assumed that a plurality of image data of the cross section of the laminated body is recorded (stored) in advance on the recording medium.

  Next, the laminate boundary detection device Ma performs brightness projection in the orthogonal direction orthogonal to the stacking direction on the image data acquired by the image acquisition unit 11 in step S11 by the brightness projection processing unit 122 of the control processing unit 12a. The luminance projection processing obtained along the stacking direction is executed, the luminance projection in the orthogonal direction obtained along the stacking direction is stored in the storage unit 13 and output to the output unit 15 (S12). The output unit 15 outputs (for example, displays) each luminance projection in the orthogonal direction obtained along the stacking direction. For example, the luminance projections in the orthogonal direction obtained along the stacking direction are displayed side by side in the image of the cross section of the stack displayed on the output unit 15 in step S11. More specifically, the luminance projection processing unit 122 applies the luminance value Y of each pixel G (x, y) arranged in a line along the orthogonal direction x orthogonal to the stacking direction y to the image data. The sum ΣY (x, y) of (x, y) is obtained, and the obtained sum ΣY (x, y) is calculated by the total number Sn of the pixels G (x, y) arranged along the orthogonal direction x. By dividing, the luminance average value Yave of the pixels G (x, y) arranged along the orthogonal direction x is obtained as the luminance projection Yp (y) (Yp = Yave = (ΣY (x, y)) / Sn). Here, in the above-described case, Σ is an operator for obtaining the sum of Y (x, Y) for x. Then, the brightness projection processing unit 122 obtains the brightness projection Yp (y) in the orthogonal direction x along the stacking direction y for the image data. That is, the luminance projection processing unit 122 obtains each luminance projection Yp (y) in the horizontal direction x for each row along the horizontal direction x along the vertical direction y (along the column) with respect to the image data. .

  Next, the laminated body boundary detection device Ma obtains a differential value for the luminance projection obtained along the lamination direction by the luminance projection processing unit 122 in Step S12 by the differential absolute value processing unit 123 of the control processing unit 12a, Differential absolute value processing is performed to obtain the absolute value of the obtained differential value as a differential absolute value, and the obtained differential absolute value is stored in the storage unit 13 (S13). More specifically, the differential absolute value processing unit 123 calculates the luminance projections Yp (y) adjacent to each other for each luminance projection Yp (y) arranged in a line along the stacking direction y obtained in step S12. Is obtained as the differential value Dv (y) (Dv (y) = Sv (y) = Yp (y) −Yp (y−1)). Then, the differential absolute value processing unit 123 calculates the absolute value | Dv (y) | of the calculated differential value Dv (y) as the differential absolute value DA (y).

  In the above description, the luminance projections Yp (y) arranged in a line along the stacking direction y obtained in the process S12 are discrete values with respect to the stacking direction y. Although the differential value Sv (y) is obtained as the differential value Dv (y), the differential absolute value processing unit 123 performs the luminance projections Yp (y) arranged in a line along the stacking direction y obtained in Step S12. The differential value Dv (y) may be obtained by obtaining a continuous function that is continuous with respect to the stacking direction y and differentiating the obtained continuous function.

  Next, the laminated body boundary detection device Ma has a predetermined absolute value for the maximum value of the differential absolute value from the differential absolute value obtained by the differential absolute value processing unit 123 in step S13 by the noise removal processing unit 124 of the control processing unit 12a. A noise removal process for extracting a differential absolute value that is greater than or equal to a ratio or a differential absolute value that is greater than a predetermined ratio with respect to a maximum value of the differential absolute value is executed, and the extracted differential absolute value is stored in the storage unit 13 and output 15 (S14). The extracted differential absolute value (differential absolute value after noise removal processing) is output (for example, displayed) to the output unit 15. For example, the image of the cross section of the stacked body displayed on the output unit 15 in the process S11 and the noise displayed side by side on each luminance projection in the orthogonal direction obtained along the stacking direction displayed on the output unit 15 in the process S12. The differential absolute value after the removal process is displayed. The predetermined ratio is appropriately set according to the magnitude of noise to be removed. For example, it is set to 20%, 20%, 30%, 30%, 40%, etc. from a plurality of samples. The In the present embodiment, as an example, the predetermined ratio is set to 30%. Therefore, the noise removal processing unit 124 extracts a differential absolute value DA (y) that is 30% or more of the maximum value DAmax of the differential absolute value DA (y) ((extracted differential absolute value DA (y)). ≧ (0.3 × DAmax)). As a result, the differential absolute value DA (y), which is less than 30% of the maximum value DAmax, is removed as noise. The differential absolute value DA (y), which is less than 30% of the maximum value DAmax, may be replaced by a preset constant m (including 0) of less than 30% ((30% of the maximum value DAmax The differential absolute value DA (y)) ← m) which is less than. Alternatively, the noise removal processing unit 124 extracts a differential absolute value DA (y) larger than 30% of the maximum value DAmax of the differential absolute value DA (y) ((differential absolute value DA (y) to be extracted)> (0.3 × DAmax)). As a result, the differential absolute value DA (y) that is 30% or less of the maximum value DAmax is removed as noise. In the same manner as described above, the differential absolute value DA (y) that is 30% or less of the maximum value DAmax may be replaced with a preset constant m (including 0) that is 30% or less (( Differential absolute value DA (y)) ← m) which is 30% or less of maximum value DAmax.

  Next, the stacked body boundary detection device Ma detects the position of the boundary surface in the cross section (based on the differential absolute value extracted by the noise removal processing unit 124 in step S14 by the boundary extraction processing unit 125a of the control processing unit 12a). A boundary extraction process for obtaining the position of the boundary line is executed (S15a). More specifically, in the first embodiment, the boundary extraction processing unit 125a performs a predetermined first as to the differential absolute value DA (y) extracted by the noise removal processing unit 124 in step S14 as the boundary extraction processing. The peaks located within the range are grouped as a peak group, and the barycentric position of the peak in the grouped peak group is obtained as the position of the boundary surface.

  For example, when the differential absolute value DA (y) after the process S14 is displayed on the output unit 15, the user refers to the displayed differential absolute value DA (y) after the process S14 and inputs The predetermined first range is input from the unit 14. For example, when the input unit 14 is a mouse, two diagonal vertices in a rectangle (an example of the predetermined first range) surrounding the peak group are designated and input by the mouse. When the predetermined first range is input from the input unit 14, the boundary extraction processing unit 125a extracts a peak from the predetermined first range input and received from the input unit 14, and the extracted peak Is obtained as the position of the boundary surface. That is, when there are a plurality of extracted peaks, the boundary extraction processing unit 125a obtains the weighted average of the positions of the peaks with the magnitudes of the peaks as the weights as the centroid positions of the peaks.

  Further, for example, the predetermined first range is stored in the storage unit 13 in advance, and the boundary extraction processing unit 125a extracts a peak from the predetermined first range stored in the storage unit 13 in advance. The center of gravity of the peak is obtained as the position of the boundary surface. In the case where the laminate boundary detection device Ma is used for the sampling inspection of the product, since the product to be subjected to the sampling inspection is determined in advance, the position of the boundary surface (position of the boundary line) in the cross-sectional image of the stack Is roughly defined, the predetermined first range can be preset and stored in the storage unit 13 in advance. The predetermined first range is set to, for example, about 1/100 of the length in the stacking direction in the cross-sectional image of the stack from a plurality of samples.

  Next, the stacked body boundary detection device Ma stores in the storage unit 13 the position of the boundary surface (the position of the boundary line) obtained by the boundary extraction processing unit 125a in step S15a by the control unit 121 of the control processing unit 12a. Then, the data is output to the output unit 15 (S16), and this process ends. The output unit 15 outputs (for example, displays) the position of the boundary surface (the position of the boundary line). For example, the position of the boundary surface (the position of the boundary line) is displayed in the vicinity (periphery) of the differential absolute value DA (y) after the processing of processing S14 displayed on the output unit 15 in processing S14.

  By such processes S11 to S16, the position of the boundary surface in the cross section of the stacked body, that is, the position of the boundary line that appears in the cross section of the stacked body is detected based on the boundary surface.

  In one example, for example, when the cross-sectional image PLB1 illustrated in FIG. 3A in the stacked body LB1 (not illustrated) is acquired in the process S11 and the luminance projection process is performed in the process S12, the stacking direction (vertical) illustrated in FIG. The luminance projection α1 in the orthogonal direction (horizontal direction) x obtained along (direction) y is obtained. The laminate LB1 is a member obtained by laminating the B1 member on the C1 member and further laminating the A1 member on the B1 member. In the example shown in FIG. 3A, in the image PLB1 of the cross section, each brightness of each layer of each member. Are substantially uniform in this example. For this reason, the luminance projection α1 in the orthogonal direction x shown in FIG. 3B changes relatively smoothly along the stacking direction (vertical direction) y, and has a profile that is recessed in the layer of the B1 member. For this reason, when the differential absolute value process is executed in process S13 and the noise removal process is executed in process S14, the differential absolute value β1 shown in FIG. 3C is obtained, and when the boundary extraction process is executed in process S15a, Within each of the predetermined first ranges, as shown in FIG. 3C, a single peak (P1) and a single peak (P12) P12 are extracted, and the barycentric positions of the peaks P11 and P12, That is, the positions of the peaks P11 and P12 are obtained as the positions of the boundary surface (positions of the boundary line) BP1 (CB) and BP1 (BA).

  In another example, for example, when the cross-sectional image PLB2 in the stacked body LB2 (not shown) shown in FIG. 4A is acquired in the process S11 and the luminance projection process is executed in the process S12, the stacking direction (vertical direction) y A luminance projection α2 (not shown) in the orthogonal direction (horizontal direction) x obtained along the line is obtained. The laminated body LB2 is a member obtained by laminating the B2 member on the C2 member and further laminating the A2 member on the B2 member. In the image PLB2 of the cross section, the luminance of the layer of the C2 member is the example shown in FIG. , Which varies due to the variation in particle size. For this reason, when differential absolute value processing is executed in step S13, a differential absolute value (differential absolute value before noise removal processing) γ shown in FIG. 4B is obtained. The differential absolute value γ shown in FIG. 4B has a profile having a plurality of small peaks in the portion corresponding to the layer of the C2 member. When the noise removal process is executed in the process S14, as shown in FIG. 4B, 30% of the maximum peak P23 is set as the threshold Th, and the differential absolute value γ greater than or equal to the threshold Th is obtained from the differential absolute value γ shown in FIG. 4B. A value DA (y) or a differential absolute value DA (y) larger than this threshold Th is extracted. When the boundary extraction process is executed in the process S15a, two peaks P21 and P22 are extracted in the predetermined first range AR1 as shown in FIG. 4B, and the barycentric positions of these peaks P21 and P22 are extracted. Is obtained as each position (each position of the boundary line) BP2 (CB) of the boundary surface, and a single peak (P1) P23 is extracted in the predetermined first range AR2 as shown in FIG. 4B. Then, the barycentric position of the peak P23, that is, the position of the peak P23 is obtained as each position (each position of the boundary line) BP2 (BA) of the boundary surface.

  As described above, since the luminance (reflectance) generally varies depending on the material, the differential value for the luminance projection in the orthogonal direction x obtained from the cross-sectional image along the stacking direction y is based on the boundary surface. It greatly changes at the boundary line appearing in the cross section. The laminated body boundary detection device Ma, the laminated body boundary detection method and the laminated body boundary detection program implemented in the laminated body boundary detection device Ma are a differential absolute value that is a predetermined ratio or more with respect to a maximum value of the differential absolute value, or a maximum value of the differential absolute value. Since the noise removal process is performed to extract a differential absolute value larger than a predetermined ratio with respect to the above, boundary extraction that is resistant to noise fluctuation and obtains the position of the boundary surface in the cross section based on the differential absolute value after the noise removal process Since the processing is performed, there is no need for the above-described correlation analysis.

  As an example, as shown in FIG. 4, depending on the laminate to be measured, the boundary line that appears in the cross section based on the boundary surface is unclear, so in the differential absolute value extracted by the noise removal process, the actual boundary surface In some cases, a plurality of peaks may occur in the vicinity (periphery) at the position. Since the laminate boundary detection device Ma, the laminate boundary detection method, and the laminate boundary detection program obtain the center of gravity of a peak in a peak group as the position of the boundary surface, even in such a case, the boundary in the cross section is obtained. The position of the surface can be obtained more appropriately.

  Note that, in the first embodiment described above, the boundary extraction processing unit 125a determines, as the boundary extraction processing, a peak located within a predetermined first range with respect to the differential absolute value extracted by the noise removal processing unit 124 as a peak group. And the center of gravity position of the peak in the group of collected peaks is obtained as the position of the boundary surface (position of the boundary line). Instead, the boundary extraction processing unit 125a performs noise extraction as the boundary extraction processing. With respect to the differential absolute value extracted by the removal processing unit 124, peaks located within a predetermined first range are grouped as a peak group, and the position of the maximum peak in the grouped peak group is defined as the position of the boundary surface (the boundary line). May be obtained as the position). For example, in the example shown in FIG. 4B, in the predetermined first range AR2, the peak P22 is the largest among the peak P21 and the peak P22, and therefore the position of the peak P22 is the position of the boundary surface (each of the boundary lines). Position) BP2 (CB). Such a laminated body boundary detection device Ma, a laminated body boundary detecting method, and a laminated body boundary detection program obtain the position of the maximum peak in the peak group as the position of the boundary surface. An apparent position can be obtained as the position of the boundary surface in the cross section.

Next, another embodiment will be described.
(Second Embodiment)
FIG. 5 is a diagram for explaining an example of boundary detection in the stacked body boundary detection device of the second embodiment. FIG. 5A schematically illustrates a cross-sectional image of the stack, FIG. 5B illustrates the luminance projection in the orthogonal direction x determined along the stack direction y of the stack, and FIG. The absolute value of the differential value (absolute value differential value) obtained with respect to the luminance projection in the orthogonal direction x obtained along the stacking direction y (FIG. 5B) is shown.

  The laminated body boundary detection device Ma according to the first embodiment collects, as the peak group, peaks located within a predetermined first range with respect to the differential absolute value extracted by the noise removal processing unit 124 as the boundary extraction process. The barycentric position of the peak in the group of collected peaks is obtained as the position of the boundary surface (the position of the boundary line). However, the stacked body boundary detection device Mb in the second embodiment performs noise removal processing as the boundary extraction processing. Peaks located within a predetermined second range are grouped as a peak group with respect to the differential absolute value extracted by the unit 124, and an average position between the position of the maximum peak and the position of the next maximum peak within the grouped peak group is It is obtained as the position of the boundary surface (position of the boundary line).

  As shown in FIG. 1, the stacked body boundary detection device Mb in the second embodiment has, for example, an image acquisition unit 11, a control processing unit 12 b, a storage unit 13, an input unit 14, and an output unit 15. With. The image acquisition unit 11, the storage unit 13, the input unit 14, and the output unit 15 in the laminate boundary detection device Mb of the second embodiment are respectively the image acquisition unit 11 in the laminate boundary detection device Ma of the first embodiment, Since it is the same as the memory | storage part 13, the input part 14, and the output part 15, the description is abbreviate | omitted.

  The control processing unit 12b controls each unit 11, 13 to 15 of the laminate boundary detection device Mb according to the function of each unit, and obtains the position of the boundary surface (position of the boundary line) in the cross section of the laminate. Circuit. The control processing unit 12b functionally includes a control unit 121, a luminance projection processing unit 122, a differential absolute value processing unit 123, a noise removal processing unit 124, and a boundary extraction processing unit 125b by executing the control processing program. The control unit 121, the luminance projection processing unit 122, the differential absolute value processing unit 123, and the noise removal processing unit 124 in the control processing unit 12b are respectively the control unit 121 and the luminance projection processing unit in the control processing unit 12a of the first embodiment. Since it is the same as 122, the differential absolute value processing part 123, and the noise removal processing part 124, the description is abbreviate | omitted.

  The boundary extraction processing unit 125b performs the boundary extraction processing. More specifically, the boundary extraction processing unit 125b in the second embodiment peaks the peak located within a predetermined second range with respect to the differential absolute value extracted by the noise removal processing unit 124 as the boundary extraction processing. As a group, the average position of the position of the maximum peak (first largest peak) and the position of the next maximum peak (second largest peak) in the group of collected peaks is the position of the boundary surface (of the boundary line). As position).

  In the laminated body boundary detection device Mb having such a configuration, when obtaining the position of the boundary surface (position of the boundary line) in the cross section of the laminated body, as shown in FIG. The brightness projection process is executed in step S12, the differential absolute value process is executed in process S13, and the noise removal process is executed in process S14. Since the processes S11 to S14 in the multilayer boundary detection device Mb of the second embodiment are the same as the processes S11 to S14 of the multilayer boundary detection device Ma of the first embodiment, respectively. The description is omitted.

  Subsequent to step S14, the multilayer boundary detection device Mb performs the boundary surface in the cross section based on the differential absolute value extracted by the noise removal processing unit 124 in step S14 by the boundary extraction processing unit 125b of the control processing unit 12b. Boundary extraction processing for obtaining the position (position of the boundary line) is executed (S15b). More specifically, in the second embodiment, the boundary extraction processing unit 125b performs a predetermined second value on the differential absolute value DA (y) extracted by the noise removal processing unit 124 in step S14 as the boundary extraction processing. Peaks located within the range are grouped as a peak group, and an average position between the position of the maximum peak and the position of the next maximum peak in the grouped peak group is obtained as the position of the boundary surface.

  For example, when the differential absolute value DA (y) after the process S14 is displayed on the output unit 15, the user refers to the displayed differential absolute value DA (y) after the process S14 and inputs The predetermined second range is input from the unit 14. When the predetermined second range is input from the input unit 14, the boundary extraction processing unit 125b extracts a peak from the predetermined second range input and received from the input unit 14, and the extracted peak The maximum peak and the next maximum peak are further extracted from the above, and a simple average position between the extracted maximum peak position and the next maximum peak position is obtained as the position of the boundary surface.

  Further, for example, the predetermined second range is stored in the storage unit 13 in advance, and the boundary extraction processing unit 125b extracts a peak from the predetermined second range stored in the storage unit 13 in advance. The center of gravity of the peak is obtained as the position of the boundary surface. Depending on the laminate to be measured, the boundary line appearing in the cross section may be wavy based on the boundary surface. In such a case, in the differential absolute value extracted by the noise removal processing unit 124, a plurality of peaks occur within a predetermined second range corresponding to the degree of the swell. The laminated body boundary detection device Mb in the second embodiment is suitable for such a case. As described above, in the case where the stacked body boundary detection device Mb is used for the sampling inspection of the product, since the product to be subjected to the sampling inspection is determined in advance, the degree of the swell can be roughly estimated. The second range can be set in advance and stored in the storage unit 13 in advance. The predetermined second range is set to, for example, about 1/10 of the length in the stacking direction in the cross-sectional image of the stack from a plurality of samples.

  Here, as in the example shown in FIG. 4, for example, a plurality of peaks that appear indistinct in the boundary line are concentrated in the vicinity of the boundary line, while, for example, as shown in FIG. Since the plurality of peaks appearing in the undulation depend on the degree of undulation, the predetermined second range is usually wider than the predetermined first range (the predetermined first range is larger than the predetermined second range). Narrow, (predetermined second range)> (predetermined first range)).

  Next, the stacked body boundary detection device Mb stores in the storage unit 13 the position of the boundary surface (position of the boundary line) obtained by the boundary extraction processing unit 125b in step S15b by the control unit 121 of the control processing unit 12b. Then, the data is output to the output unit 15 (S16), and this process ends. The output unit 15 outputs (for example, displays) the position of the boundary surface (the position of the boundary line).

  By such processes S11 to S16, the position of the boundary surface in the cross section of the stacked body, that is, the position of the boundary line that appears in the cross section of the stacked body is detected based on the boundary surface.

  In one example, for example, when the cross-sectional image PLB3 illustrated in FIG. 5A in the stacked body LB3 (not illustrated) is acquired in the process S11 and the luminance projection process is performed in the process S12, the stacking direction (vertical) illustrated in FIG. The luminance projection α3 in the orthogonal direction (horizontal direction) x obtained along (direction) y is obtained. The laminated body LB3 is a member obtained by laminating the B3 member on the C3 member, and further laminating the A3 member on the B3 member. In the example shown in FIG. 5A, in the image PLB3 of the cross section, for example, a malfunction in the manufacturing process is performed. Due to the above, the boundary lines BL3 (CB) and BL3 (BA) appearing in the cross section of the stacked body LB3 based on the boundary surface are not straight lines but curves, and have inflection points. Therefore, the luminance projection α3 in the orthogonal direction x shown in FIG. 5B changes in a step shape along the stacking direction (vertical direction) y, and has a profile that is recessed in multiple steps around the layer of the B3 member. For this reason, when the differential absolute value process is executed in process S13 and the noise removal process is executed in process S14, the differential absolute value β3 shown in FIG. 5C is obtained, and when the boundary extraction process is executed in process S15b, Within the predetermined second range ER1 (CB), as shown in FIG. 5C, two peaks P31 and P32 are extracted, and from these peaks P31 and P32, the maximum peak and the next maximum peak, that is, in this example, Peaks P31 and P32 are further extracted, and a simple average position between the positions of the extracted peaks P31 and P32 is obtained as the position of the boundary surface (position of the boundary line) BP3 (CB). Within the predetermined second range ER1 (BA), two peaks P33 and P34 are extracted as shown in FIG. 5C, and the maximum is selected from these peaks P33 and P34. And the next maximum peak, that is, in this example, the peak P33 and the peak P34 are further extracted, and the simple average position between the position of the extracted peak P33 and the position of the peak P34 is the position of the boundary surface (the boundary line). Position) BP3 (BA).

  As described above, the layer boundary detection device Mb in the second embodiment, the layer boundary detection method and the layer boundary detection program mounted thereon are the same as those in the layer boundary detection device Ma in the first embodiment. Since the noise removal process is performed in the same manner as in the implemented laminated body boundary detection method and laminated body boundary detection program, it is resistant to noise fluctuations and the boundary extraction process is performed, so that the above-described correlation analysis is not necessary.

  And the laminated body boundary detection apparatus Mb in 2nd Embodiment, the laminated body boundary detection method mounted in this, and the laminated body boundary detection program are the average position of the position of the largest peak in the peak group, and the position of the next largest peak. Is obtained as the position of the boundary surface, so that the most probable position can be obtained as the position of the boundary surface in the cross section even when the boundary line appearing in the cross section is wavy based on the boundary surface.

Next, another embodiment will be described.
(Third embodiment)
FIG. 6 is a flowchart showing the operation of the stacked body boundary detection apparatus according to the third embodiment. FIG. 7 is a diagram for explaining an example of boundary detection in the multilayer body boundary detection device according to the third embodiment. 7A schematically shows a cross-sectional image of the laminate, and FIG. 7B shows a luminance projection in the orthogonal direction x determined along the laminate direction y of the laminate before rotation. Indicates the absolute value of the differential value (absolute value differential value) obtained with respect to the luminance projection (FIG. 5B) in the orthogonal direction x obtained along the stacking direction y before the rotation, and FIG. FIG. 5E shows the luminance projection in the orthogonal direction x determined along the stacking direction y of the stack after appropriate rotation, and FIG. 5E is determined along the stacking direction y after the appropriate rotation. The absolute value of the differential value (absolute value differential value) obtained for the luminance projection in the orthogonal direction x (FIG. 5D) is shown.

  The laminated body boundary detection device Ma according to the first embodiment collects, as the peak group, peaks located within a predetermined first range with respect to the differential absolute value extracted by the noise removal processing unit 124 as the boundary extraction process. The center-of-gravity position of the peak in the group of collected peaks was obtained as the position of the boundary surface (the position of the boundary line). However, the stacked body boundary detection device Mc in the third embodiment Until the number of boundary positions, that is, the number of boundary surfaces and the number of boundary lines, reaches a given predetermined number, the image of the cross section with a predetermined position in the image of the cross section is set to a predetermined value. The brightness projection process, the differential absolute value process, the noise removal process, and the boundary extraction process are executed while rotating each angle.

  As shown in FIG. 1, the stacked body boundary detection device Mc in the third embodiment has, for example, an image acquisition unit 11, a control processing unit 12 c, a storage unit 13, an input unit 14, and an output unit 15. With. The image acquisition unit 11, the storage unit 13, the input unit 14, and the output unit 15 in the stacked body boundary detection device Mc according to the third embodiment are respectively the same except that the input unit 14 further receives the number of the boundary surfaces. Since it is the same as that of the image acquisition part 11, the memory | storage part 13, the input part 14, and the output part 15 in the laminated body boundary detection apparatus Ma of 1 embodiment, the description is abbreviate | omitted.

  The control processing unit 12c controls each unit 11, 13 to 15 of the laminate boundary detection device Mc according to the function of each unit, and obtains the position of the boundary surface (position of the boundary line) in the cross section of the laminate. Circuit. The control processing unit 12c functionally includes a control unit 121, a luminance projection processing unit 122, a differential absolute value processing unit 123, a noise removal processing unit 124, and a boundary extraction processing unit 125c by executing the control processing program. The rotation processing unit 126c shown is functionally provided. The control unit 121, the luminance projection processing unit 122, the differential absolute value processing unit 123, and the noise removal processing unit 124 in the control processing unit 12c are respectively the control unit 121 and the luminance projection processing unit in the control processing unit 12a of the first embodiment. Since it is the same as 122, the differential absolute value processing part 123, and the noise removal processing part 124, the description is abbreviate | omitted.

  The boundary extraction processing unit 125c performs the boundary extraction processing. More specifically, the boundary extraction processing unit 125c in the third embodiment may perform the boundary extraction processing (including variations thereof) described above in the first embodiment, or described above in the second embodiment. The boundary extraction process may be performed.

  The rotation processing unit 126c moves the image of the cross section at a predetermined angle around a predetermined position (for example, a central position) in the image of the cross section until the number of positions of the boundary surface reaches a predetermined number. While rotating in a predetermined direction each time, the luminance projection processing unit 122 performs the luminance projection processing, the differential absolute value processing unit 123 performs the differential absolute value processing, and the noise removal processing unit 124 performs the noise removal processing. The boundary extraction processing unit 125c performs the boundary extraction process. In the present embodiment, for example, the predetermined number is input from the input unit 14 and received before the processing of the rotation processing unit 126c. The predetermined angle is appropriately set, for example, 0.1 degree, 0.2 degree, 0.5 degree, and 1 degree, and is set to 0.1 degree in the present embodiment. The predetermined direction may be maintained in a certain direction until the number of positions of the boundary surface reaches a predetermined number set in advance, for example, may be counterclockwise or clockwise. good.

  In the laminated body boundary detection device Mc having such a configuration, when the measurement start is instructed from the input unit 14 when obtaining the position of the boundary surface (position of the boundary line) in the cross section of the laminated body, in FIG. The laminated body boundary detection device Mc acquires the image data of the cross section of the laminated body by the image obtaining unit 11 under the control of the control unit 121 of the control processing unit 12c, and the obtained laminated body. Is stored in the storage unit 13 and output to the output unit 15 (S21). The output unit 15 outputs (for example, displays) an image of a cross section of the laminate.

  Next, the stacked body boundary detection device Mc receives the number of positions of the boundary surface (the number of positions of the boundary line) as the predetermined number from the input unit 14 under the control of the control unit 121 of the control processing unit 12c. And stored in the storage unit 13 (S22). For example, when the image of the cross section of the stacked body acquired in step S11 is displayed on the output unit 15, the user refers to the displayed cross section image of the stacked body, and inputs the boundary surface from the input unit 14 Enter the number of positions.

  Next, the laminated body boundary detection device Mc performs the luminance projection processing by the luminance projection processing unit 122 of the control processing unit 12c in the same manner as the processing S12, and calculates the orthogonal direction obtained along the stacking direction. The luminance projection is stored in the storage unit 13 and output to the output unit 15 (S23).

  Next, the laminated body boundary detection device Mc performs the differential absolute value processing by the differential absolute value processing unit 123 of the control processing unit 12c in the same manner as the process S13, and stores the obtained differential absolute value in the storage unit 13. (S24).

  Next, the laminated body boundary detection device Mc performs the noise removal processing by the noise removal processing unit 124 of the control processing unit 12c, similarly to the processing S14, and stores the extracted differential absolute value in the storage unit 13. And output to the output unit 15 (S25). The extracted differential absolute value (differential absolute value after noise removal processing) is output (for example, displayed) to the output unit 15.

  Next, the stacked body boundary detection device Mc uses the boundary extraction processing unit 125c of the control processing unit 12c to perform the boundary extraction processing in the same manner as the processing S15a (including variations thereof) or in the same manner as the processing S15b. Is executed (S26).

  Next, in the multilayer body boundary detection device Mc, the rotation processing unit 126c of the control processing unit 12c has a predetermined number of the boundary surface positions (the number of the boundary line positions) obtained in step S26. The number, in this embodiment, it is determined whether or not it matches the number stored in the reception storage unit 13 in step S22 (S27). As a result of this determination, when these two match (YES), the stacked body boundary detection device Mc next executes the process S28. On the other hand, when they do not match (NO), the stacked body Next, the boundary detection device Mc executes the process S31.

  In the process S31, the rotation processing unit 126c rotates the cross-sectional image in a predetermined direction by a predetermined angle around a predetermined position in the cross-sectional image that is the target of each of the processes S23 to S26. Then, the process returns to process S23. For example, in the present embodiment, the rotation processing unit 126c rotates the cross-sectional image counterclockwise by 0.1 degrees around the center position in the cross-sectional image that is the target of each of the processes S23 to S26. Rotate. Accordingly, in each of the following processes S23 to S26, the luminance projection process, the differential absolute value process, the noise removal process, and the boundary extraction process are performed on the image of the cross section after the rotation. The Accordingly, the image of the cross section is centered on a predetermined position in the image of the cross section until the number of positions of the boundary surface reaches the predetermined number by each of the processes S23 to S27 and S31. While being rotated by a predetermined angle, the luminance projection processing unit 122 executes the luminance projection processing, the differential absolute value processing unit 123 executes the differential absolute value processing, and the noise removal processing unit 124 executes the noise removal processing. Then, the boundary extraction processing unit 125c executes the boundary extraction processing.

  On the other hand, in the process S28, the laminate boundary detection device Mc is obtained by the boundary extraction processing unit 125c in the process S26 when the control unit 121 of the control processing unit 12c determines that the two match in the process S27. The position of the boundary surface (the position of the boundary line) is stored in the storage unit 13 and output to the output unit 15, and this process is terminated. The output unit 15 outputs (for example, displays) the position of the boundary surface (the position of the boundary line).

  By such processes S21 to S28 and S31, the position of the boundary surface in the cross section of the stacked body, that is, the position of the boundary line that appears in the cross section of the stacked body is detected based on the boundary surface.

  In one example, for example, an image PLB4 of the cross section shown in FIG. 7A in the stacked body LB4 (not shown) is acquired in the process S21, and the number of positions of the boundary surface (the number of positions of the boundary line) is input from the input unit 14. When the brightness projection process is executed in the process S23 via the process S22 that has received “2”, the brightness in the orthogonal direction (horizontal direction) x obtained along the stacking direction (vertical direction) y shown in FIG. 7B. Projection α4 is determined. The laminated body LB4 is a member obtained by laminating a B4 member on a C4 member, and further laminating an A4 member on a B4 member. In the example shown in FIG. 7A, in the cross-sectional image PLB4, for example, the laminated body LB4 is a microscope. The boundary lines BL4 (CB) and BL4 (BA) appearing in the cross section of the laminate LB4 based on the boundary surface due to the lack of adjustment when setting to the microscope, the lack of adjustment when attaching the camera to the microscope, etc. Each is diagonal to Heisei direction x. For this reason, the luminance projection α4 in the orthogonal direction (horizontal direction) x shown in FIG. 7B changes in a step shape along the stacking direction (vertical direction) y, and becomes a profile that is recessed in multiple steps around the layer of the B4 member. Yes. Therefore, when the differential absolute value process is executed in the process S24 and the noise removal process is executed in the process S25, the differential absolute value β4 shown in FIG. 7C is obtained. In the process S26, for example, boundary extraction similar to the process S15a is performed. When the process is executed, four peaks P41 to P44 are extracted as shown in FIG. 7C, and the positions of these peaks P41 to P44 are the four positions of the boundary surface (positions of the boundary line). Desired. Therefore, in process S27, it is determined that the two do not match, and in process S31, the cross-sectional image PLB4 is rotated by 0.1 degree counterclockwise, and the process returns to process S23.

  When the processes S23 to S27 and S31 are repeated, the cross-sectional image PLB4 shown in FIG. 7A eventually appears on the boundary line BL4 ( CB) and BL4 (BA) are substantially horizontal images. When the luminance projection processing is executed in step S23 on the horizontal image, the luminance projection α5 in the orthogonal direction (horizontal direction) x obtained along the stacking direction (vertical direction) y shown in FIG. 7D. Is required. Since the image is horizontal, the luminance projection α5 in the orthogonal direction x shown in FIG. 7D changes relatively smoothly along the stacking direction y, and has a profile that is recessed in the layer of the B4 member. For this reason, when the differential absolute value process is executed in process S24 and the noise removal process is executed in process S25, the differential absolute value β5 shown in FIG. 7E is obtained, and when the boundary extraction process is executed in process S26, As shown in FIG. 7E, two peaks P51 and P52 are extracted, and the positions of these peaks P51 and P52 are obtained as the positions of the two boundary surfaces (positions of the boundary lines). Therefore, in process S27, it is determined that the two match, and in process S28, the positions of the peaks P51 and P52 are the positions of the boundary surfaces (positions of the boundary lines) BP4 (CB) and BP4 (BA ) Is output.

  As described above, the laminate boundary detection device Mc according to the third embodiment, the laminate boundary detection method and the laminate boundary detection program mounted thereon are the same as the laminate boundary detection device Ma according to the first embodiment. Since the noise removal process is performed in the same manner as in the implemented laminated body boundary detection method and laminated body boundary detection program, it is resistant to noise fluctuations and the boundary extraction process is performed, so that the above-described correlation analysis is not necessary.

  As an example, as shown in FIG. 7, depending on the image of the cross section, the boundary line that appears in the cross section based on the boundary surface does not coincide with the horizontal direction of the image but is inclined with respect to the horizontal direction of the image. There is a case. In the laminated body boundary detection device Mc according to the third embodiment, the laminated body boundary detecting method and the laminated body boundary detecting program mounted thereon, the number of positions of the boundary surface in the cross section is a predetermined number. Up to this point, the position of the boundary surface in the cross section is obtained while rotating the cross section image by a predetermined angle around a predetermined position (for example, the center position in the present embodiment) in the cross section image. Even in this case, the position of the boundary surface in the cross section can be obtained.

  In the third embodiment described above, the number of the positions of the boundary plane (the number of positions of the boundary line) is input as the predetermined number from the input unit 14 by the user. May be stored in advance in the storage unit 13 as the predetermined number, and the rotation processing unit 126c may use the number of positions of the boundary surface stored in advance in the storage unit 13. In the case where the laminate boundary detection device Mc is used for the sampling inspection of the product, since the products to be subjected to the sampling inspection are determined in advance, the number of the positions of the boundary surfaces in the cross-sectional image of the stack is substantially defined. Therefore, the number of positions on the boundary surface can be set in advance and stored in the storage unit 13 in advance.

  In the first to third embodiments described above, when the control processing program is not installed in the storage unit 13, an IF unit is provided from a computer-readable recording medium on which the control processing program is recorded. The control processing program may be read and installed via the image processing unit 11 configured as described above or the IF unit 17. As described above, the control processing program includes the luminance projection processing program, the differential absolute value processing program, the noise removal processing program, and the boundary extraction processing program. That is, the control processing program includes a laminate boundary detection program including these programs. According to this, it is possible to provide a computer-readable recording medium that records the above-described laminate boundary detection program that is resistant to noise fluctuation and does not require correlation analysis.

  In order to express the present invention, the present invention has been properly and fully described through the embodiments with reference to the drawings. However, those skilled in the art can easily change and / or improve the above-described embodiments. It should be recognized that this is possible. Therefore, unless the modifications or improvements implemented by those skilled in the art are at a level that departs from the scope of the claims recited in the claims, the modifications or improvements are not covered by the claims. To be construed as inclusive.

Ma, Mb, Mc Laminate boundary detection device 11 Image income unit 122 Luminance projection processing unit 123 Differential absolute value processing unit 124 Noise removal processing unit 125a, 125b, 125c Boundary extraction processing unit 126c Rotation processing unit

Claims (8)

An image acquisition unit that acquires image data of a cross section including a stacking direction intersecting a boundary surface in a stacked body that is a member in which a plurality of different materials are stacked;
A brightness projection processing unit that performs brightness projection processing for obtaining a brightness projection in the orthogonal direction orthogonal to the stacking direction along the stacking direction with respect to the image data acquired by the image acquisition unit;
A differential absolute value processing unit that performs differential absolute value processing for obtaining a differential value for the luminance projection obtained along the stacking direction in the luminance projection processing unit, and obtaining an absolute value of the obtained differential value as a differential absolute value;
From the differential absolute value obtained by the differential absolute value processing unit, a differential absolute value that is equal to or greater than a predetermined ratio with respect to the maximum value of the differential absolute value or a differential absolute value greater than a predetermined ratio with respect to the maximum value of the differential absolute value is extracted. A noise removal processing unit for performing noise removal processing,
A laminate boundary detection apparatus comprising: a boundary extraction processing unit that performs a boundary extraction process for obtaining a position of the boundary surface in the cross section based on the differential absolute value extracted by the noise removal processing unit. The boundary extraction processing unit, as the boundary extraction processing, summarizes the peaks located within a predetermined range as a peak group with respect to the differential absolute value extracted by the noise removal processing unit, and calculates the peak in the collected peak group. The center of gravity position is obtained as the position of the boundary surface. The boundary extraction processing unit, as the boundary extraction processing, summarizes the peaks located within a predetermined range as a peak group with respect to the differential absolute value extracted by the noise removal processing unit, and the maximum peak in the combined peak group The position of the laminated body is detected as the position of the boundary surface. The boundary extraction processing unit summarizes the peaks located within a predetermined second range as a peak group with respect to the differential absolute value extracted by the noise removal processing unit as the boundary extraction processing, and within the combined peak group The laminated body boundary detection device according to claim 1, wherein an average position of a maximum peak position and a next maximum peak position is obtained as the position of the boundary surface. The luminance projection processing unit is configured to rotate the image of the cross section around a predetermined position in the image of the cross section by a predetermined angle until the number of positions of the boundary surface reaches a predetermined number set in advance. Perform luminance projection processing, cause the differential absolute value processing unit to perform differential absolute value processing, cause the noise removal processing unit to perform the noise removal processing, and perform the boundary extraction processing to the boundary extraction processing unit The laminate boundary detection device according to claim 1, further comprising: a rotation processing unit that performs the rotation processing. An image acquisition step of acquiring image data of a cross section including a stacking direction intersecting a boundary surface in a stacked body that is a member in which a plurality of different materials are stacked,
A luminance projection processing step for obtaining a luminance projection along the stacking direction in an orthogonal direction orthogonal to the stacking direction for the image data acquired in the image acquisition step;
A differential absolute value processing step for obtaining a differential value for the luminance projection obtained along the stacking direction in the luminance projection processing step, and obtaining an absolute value of the obtained differential value as a differential absolute value;
From the differential absolute value obtained in the differential absolute value processing step, a differential absolute value that is equal to or greater than a predetermined ratio with respect to the maximum value of the differential absolute value or a differential absolute value greater than a predetermined ratio with respect to the maximum value of the differential absolute value is extracted. Noise removal processing step,
And a boundary extraction processing step for obtaining a position of the boundary surface in the cross section based on the differential absolute value extracted in the noise removal processing step. On the computer,
An image acquisition step of acquiring image data of a cross section including a stacking direction intersecting a boundary surface in a stacked body that is a member in which a plurality of different materials are stacked,
A luminance projection processing step for obtaining a luminance projection along the stacking direction in an orthogonal direction orthogonal to the stacking direction for the image data acquired in the image acquisition step;
A differential absolute value processing step for obtaining a differential value for the luminance projection obtained along the stacking direction in the luminance projection processing step, and obtaining an absolute value of the obtained differential value as a differential absolute value;
From the differential absolute value obtained in the differential absolute value processing step, a differential absolute value that is equal to or greater than a predetermined ratio with respect to the maximum value of the differential absolute value or a differential absolute value greater than a predetermined ratio with respect to the maximum value of the differential absolute value is extracted. Noise removal processing step,
Based on the differential absolute value extracted in the noise removal processing step, a boundary extraction processing step for obtaining the position of the boundary surface in the cross section;
Laminate boundary detection program to execute. On the computer,
An image acquisition step of acquiring image data of a cross section including a stacking direction intersecting a boundary surface in a stacked body that is a member in which a plurality of different materials are stacked,
A luminance projection processing step for obtaining a luminance projection along the stacking direction in an orthogonal direction orthogonal to the stacking direction for the image data acquired in the image acquisition step;
A differential absolute value processing step for obtaining a differential value for the luminance projection obtained along the stacking direction in the luminance projection processing step, and obtaining an absolute value of the obtained differential value as a differential absolute value;
From the differential absolute value obtained in the differential absolute value processing step, a differential absolute value that is equal to or greater than a predetermined ratio with respect to the maximum value of the differential absolute value or a differential absolute value greater than a predetermined ratio with respect to the maximum value of the differential absolute value is extracted. Noise removal processing step,
Based on the differential absolute value extracted in the noise removal processing step, a boundary extraction processing step for obtaining the position of the boundary surface in the cross section;
The computer-readable recording medium which recorded the laminated body boundary detection program for performing this.

Images