Image processing device and image processing method

Abstract

To provide an image processing device and an image processing method with which quick and highly accurate analysis is possible.SOLUTION: An image processing device according to the present invention comprises: an image acquisition unit 11 for acquiring a captured image in which the crystal structure of a sintered compact having a parent phase and a second phase is captured; a parent phase determination unit 12 for determining the crystal grain and crystal grain boundary of the parent phase or the crystal grain boundary of the parent phase and generating a parent phase determination image by using a model having been trained regarding the parent phase using, as teaching data, a known captured image and a parent phase setting image in which the crystal grain and crystal grain boundary of the parent phase or the crystal grain boundary of the parent phase in the known captured image are set; and a second phase determination unit 13 for determining the crystal grain and crystal grain boundary of the second phase or the crystal grain boundary of the second phase and generating a second phase determination image by using a model having been trained regarding the second phase using, as teaching data, a known captured image and a second phase setting image in which the crystal grain and crystal grain boundary of the second phase or the crystal grain boundary of the second phase in the known captured image are set; and a measurement unit 14 for measuring the crystal grain sizes of the parent phase and the second phase.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to an image processing apparatus and an image processing method.

Background Art

[0002] When measuring the crystal grain size of a sintered body used for a sputtering target material, when manually observing an image of the crystal structure of the sintered body and determining each crystal phase and crystal grain boundary, it has imposed a great deal of labor on the operator.

[0003] On the other hand, the crystal analysis apparatus disclosed in Patent Document 1 constructs a three-dimensional crystal orientation map showing the crystal orientation in the normal direction of each surface for a plurality of surfaces of the polyhedron image of the sample based on the EBSP data of the cross section formed on the sample, whereby the attribution of the crystal phase and the crystal grain size can be measured.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] However, in the crystal analysis apparatus disclosed in Patent Document 1, it is necessary to perform cross-section processing in which the sample is sliced at a predetermined interval (10 mm interval in Patent Document 1) and a plurality of cross sections that are substantially parallel to each other are formed, which requires a laborious operation. Further, since the three-dimensional crystal orientation map is constructed by arranging the two-dimensional crystal orientation maps constructed from the EBSP data of each cross section based on the slice interval of the cross-section processing, it has taken a long time to measure the crystal grain size.

[0006] In view of the above problems, the present invention provides an image processing apparatus and an image processing method capable of rapid and highly accurate analysis.

Means for Solving the Problem

[0007] An image processing apparatus of the present invention made to solve the above problems is an image processing apparatus for a captured image of a crystal structure of a sintered body having a matrix phase and a second phase different from the matrix phase, including an image acquisition unit that acquires the captured image, a known captured image, crystal grains and crystal grain boundaries of the matrix phase in the known captured image, or a matrix phase setting image in which the crystal grain boundaries of the matrix phase are set, and a matrix phase learned model generated by performing machine learning using the matrix phase dataset as teacher data, and determines crystal grains and crystal grain boundaries of the matrix phase, or crystal grain boundaries of the matrix phase in the captured image, and generates a matrix phase determination image; a second phase determination unit that determines crystal grains and crystal grain boundaries of the second phase, or crystal grain boundaries of the second phase in the captured image, and generates a second phase determination image, using a second phase learned model generated by performing machine learning using a second phase dataset including the known captured image, crystal grains and crystal grain boundaries of the second phase in the known captured image, or a second phase setting image in which the crystal grain boundaries of the second phase are set as teacher data; and a measurement unit that measures crystal grain diameters of the matrix phase and the second phase using the matrix phase determination image and the second phase determination image.

[0008] Using the image processing apparatus of the present invention, the sintered body to be image-processed may be a sintered body for a sputtering target, for example, an ITO sintered body. Here, in the ITO sintered body, the matrix phase is the In2O3 phase, and the second phase different from the matrix phase is the In4Sn3O 12 phase. Note that the second phase is a general term indicating a crystal structure different from the matrix phase. In a sintered body having a plurality of crystal structures different from the matrix phase, all crystal structures different from the matrix phase are regarded as the second phase.

[0009] The image acquisition unit acquires a captured image of the crystal structure of a sintered body having a matrix phase and a second phase. Specifically, the image acquisition unit acquires a captured image captured by a scanning electron microscope or the like communicably connected to the image processing apparatus of the present invention, whether by wire or wirelessly. Further, examples of the captured image include a captured image captured by an optical microscope, a stereomicroscope, or a laser microscope. Furthermore, the captured image may be a captured image such as a secondary electron image, a reflection electron image, or a three-dimensional image captured by a scanning electron microscope, which is an electron microscope, or a secondary electron image, a reflection electron image, or a three-dimensional image captured by a transmission electron microscope.

[0010] Here, the captured image according to the present invention is an image of the crystal structure of a sintered body having a matrix phase and a second phase. Further, when the captured image is an SEM image captured using a scanning electron microscope, since the particle sizes of the matrix phase and the second phase are several micrometers and fine, it is preferable in that their grain boundaries can be clearly captured. The captured image is captured as follows. First, the cut surface obtained by cutting the sintered body is polished step by step using abrasive papers #180, #400, #800, #1000, and #2000, and finally buffed to finish it into a mirror surface. Next, the cut surface finished into a mirror surface is captured using a scanning electron microscope (SU3500, manufactured by Hitachi High-Technologies Corporation).

[0011] Furthermore, when the captured image is an image of a cross section of a sintered body subjected to an etching treatment with aqua regia, it is preferable in that unnecessary substances attached to the cross section of the sintered body can be removed. Also, when the captured image is an image in which the contrast of the image is adjusted, the cross section of the sintered body becomes clear, and the measurement accuracy of the crystal grain size can be improved.

[0012] The parent phase determination unit uses a machine learning model trained with a parent phase dataset including a known captured image, crystal grains and crystal grain boundaries of the parent phase in the known captured image, or a parent phase setting image with crystal grain boundaries of the parent phase set as teacher data to determine crystal grains and crystal grain boundaries of the parent phase, or crystal grain boundaries of the parent phase in the captured image, and generates a parent phase determination image. Note that it is preferable from the viewpoint of accuracy improvement that the known captured image and the parent phase setting image have the same image data format, for example, a secondary electron image, etc., and the imaging conditions such as the acceleration voltage are also the same conditions.

[0013] Here, the trained parent phase model is generated by performing machine learning using a parent phase dataset in which a known captured image, crystal grains and crystal grain boundaries of the parent phase in the known captured image, or a parent phase setting image with crystal grain boundaries of the parent phase set are associated as teacher data.

[0014] Also, the setting of crystal grains and crystal grain boundaries of the parent phase, or crystal grain boundaries of the parent phase in the known captured image is set by an operator or the like and associated with the known captured image. For example, when the known captured image is a captured image of the crystal structure of an ITO sintered body, an operator or the like sets the parent phase (for example, In2O3 phase) based on contrast, color tone in the SEM image, or the shape of the crystal phase, and sets the crystal grains of the parent phase and the boundary with the second phase as the crystal grain boundaries of the parent phase. Then, the operator or the like associates the set crystal grains and crystal grain boundaries of the parent phase, or crystal grain boundaries of the parent phase in the known captured image with the known captured image.

[0015] Furthermore, the trained parent phase model may be generated separately for each raw material of the sintered body, that is, divided into an ITO sintered body and an IGZO sintered body, or may be combined into one regardless of the raw material of the sintered body.

[0016] Specifically, when the sintered body captured in the captured image is an ITO sintered body, the matrix phase determination unit performs machine learning using, as teacher data, a matrix phase dataset including a known captured image of the ITO sintered body, crystal grains of the matrix phase and grain boundaries of the matrix phase in the known captured image, or a matrix phase setting image in which grain boundaries of the matrix phase are set, to generate a learned matrix phase model. Using this model, the crystal grains of the matrix phase (e.g., In2O3 phase) and grain boundaries of the matrix phase, or grain boundaries of the matrix phase in the captured image are determined, and a matrix phase determination image is generated.

[0017] The second phase determination unit uses a learned second phase model generated by performing machine learning using, as teacher data, a second phase dataset including a known captured image, crystal grains of the second phase and grain boundaries of the second phase in the known captured image, or a second phase setting image in which grain boundaries of the second phase are set, to determine the crystal grains of the second phase and grain boundaries of the second phase, or grain boundaries of the second phase in the captured image, and generate a second phase determination image. Note that it is preferable from the viewpoint of improving accuracy that the known captured image and the second phase setting image have the same image data format, such as a secondary electron image, and the same imaging conditions, such as acceleration voltage.

[0018] Here, the learned second phase model is generated by performing machine learning using, as teacher data, a second phase dataset in which a known captured image is associated with crystal grains of the second phase and grain boundaries of the second phase, or a second phase setting image in which grain boundaries of the second phase are set, in the known captured image. Note that the setting of the crystal grains of the second phase and grain boundaries of the second phase, or grain boundaries of the second phase in the known captured image is set by an operator or the like and associated with the known captured image.

[0019] Also, the setting of the crystal grains of the second phase and grain boundaries of the second phase, or grain boundaries of the second phase in the known captured image is set by an operator or the like and associated with the known captured image. For example, when the known captured image is a captured image of the crystal structure of an ITO sintered body, an operator or the like uses contrast, color tone in an SEM image, or the shape of the crystal phase as a reference to determine the second phase (e.g., In4Sn3O 12Set the second phase, and set the boundary with respect to the crystal grains of the second phase and the matrix phase as the grain boundary of the second phase. Then, an operator or the like associates the crystal grains of the second phase and the grain boundary of the second phase, or the grain boundary of the second phase, in the set known captured image with the known captured image.

[0020] In addition, the second-phase learned model is a learned model based on teacher data divided for each second phase according to the composition system of the sintered body or the like.

[0021] Specifically, when the sintered body captured in the captured image is an ITO sintered body, the second-phase determination unit uses, as teacher data, a second-phase data set including a known captured image of the ITO sintered body and the crystal grains of the second phase and the grain boundary of the second phase, or the second-phase setting image in which the grain boundary of the second phase is set, in the known captured image, and performs machine learning to generate a second-phase learned model. Then, using the second-phase learned model, the crystal grains of the second phase (for example, In4Sn3O 12 phase) and the grain boundary of the second phase, or the grain boundary of the second phase are determined, and a second-phase determination image is generated.

[0022] For convenience of explanation, the matrix-phase determination unit and the second-phase determination unit are described as separate configurations, but a configuration may be adopted in which a determination unit capable of determining both the matrix phase and the second phase is used. Similarly, a configuration may be adopted in which the matrix-phase learned model and the second-phase learned model are integrated into one learned model.

[0023] The measurement unit measures the crystal grain sizes of the matrix phase and the second phase using the matrix-phase determination image and the second-phase determination image. Specifically, the measurement unit performs particle analysis on the matrix-phase determination image and the second-phase determination image to obtain the areas of the particles constituting the matrix phase and the second phase, respectively. Next, from the areas of the particles constituting the obtained matrix phase and second phase, the equivalent circle diameters of the respective areas are calculated. Then, the calculated equivalent circle diameters of the particles constituting the matrix phase and the second phase are taken as the crystal grain sizes of the matrix phase and the second phase.

[0024] In addition to the crystal grain sizes of the matrix phase and the second phase, the measurement unit may measure, for example, the area ratios of the matrix phase and the second phase, the horizontal Feret diameter, the vertical Feret diameter, the particle area, the equivalent circle diameter of the particle, the perimeter of the particle, the maximum diameter, the minimum diameter, etc.

[0025] Furthermore, the image processing apparatus of the present invention includes a control unit (CPU) that controls the above-described image acquisition unit, matrix phase determination unit, second phase determination unit, and measurement unit, a storage unit (RAM, ROM, etc.) that stores captured images, matrix phase setting images, matrix phase learned models, second phase setting images, second phase learned models, matrix phase determination images, second phase determination images, etc., and various programs, a communication unit that can communicate with a scanning electron microscope used for capturing a captured image, other information communication terminals, etc., an input unit that receives an input from an operator, etc., and a display unit that can display a captured image, a matrix phase setting image, a second phase setting image, a matrix phase determination image, a second phase determination image, etc.

[0026] The control unit controls the entire image processing apparatus of the present invention. The control unit performs various processes according to an input operation received from an operator via the input unit. For example, the control unit receives a captured image captured by a scanning electron microscope via the communication unit, stores the received captured image in the storage unit, or causes the display unit to display the crystal grain sizes of the matrix phase and the second phase measured by the measurement unit.

[0027] The memory unit stores various images, various programs, etc. Among the images and information stored in the memory unit, the captured images taken by a scanning electron microscope, the matrix phase learned model, the second phase learned model, the matrix phase determination image generated by the matrix phase determination unit, the second phase determination image generated by the second phase determination unit, the matrix phase and the crystal grain size of the second phase measured by the measurement unit may be stored and preserved not in the memory unit within the image processing apparatus of the present invention but in a database on a server communicably connected to the image processing apparatus of the present invention via a network. Each image and information stored in the database is transmitted and received between the image processing apparatus of the present invention via the network. Thus, the database on the server communicably connected to the image processing apparatus of the present invention via the network is considered to be included in the configuration of the image processing apparatus of the present invention. Note that the network may be a LAN or a WAN.

[0028] The communication unit performs transmission and reception of various data between the image processing apparatus of the present invention, a scanning electron microscope, and a database on a server. For example, the communication unit receives a captured image taken by a scanning electron microscope, receives a matrix phase learned model or a second phase learned model stored in a database on a server, and transmits a matrix phase determination image generated by the matrix phase determination unit or a second phase determination image generated by the second phase determination unit to the database on the server.

[0029] The input unit receives various instructions from an operator and input operations of various information. The input unit may be a keyboard, a mouse, a touch pad, etc.

[0030] The display unit displays various images and various information. The display unit may be a liquid crystal display, a CRT display, an organic EL display, a touch panel, etc.

[0031] Further, the image processing apparatus of the present invention further includes a binarization processing unit that generates a binarized matrix phase image and a binarized second phase image by performing binarization processing on the matrix phase determination image and the second phase determination image, and the measurement unit measures the crystal grain sizes of the matrix phase and the second phase using the binarized matrix phase image and the binarized second phase image. The image processing apparatus of the present invention generates a binarized matrix phase image and a binarized second phase image by binarizing the matrix phase determination image and the second phase determination image generated by the matrix phase determination unit and the second phase determination unit, and measures the crystal grain sizes of the matrix phase and the second phase using the binarized matrix phase image and the binarized second phase image. This is preferable in that it becomes easier to distinguish the difference in color tone between the matrix phase and the second phase, and it becomes easier to measure the crystal grain sizes of the matrix phase and the second phase.

[0032] Specifically, the binarization processing unit generates a binarized matrix phase image and a binarized second phase image by performing binarization processing on the matrix phase determination image and the second phase determination image. Next, the measurement unit performs particle analysis on the binarized matrix phase image and the binarized second phase image to obtain the areas of the particles constituting the matrix phase and the second phase, respectively. Next, the equivalent circle diameter of each area is calculated from the areas of the particles constituting the obtained matrix phase and second phase. Then, the equivalent circle diameters of the areas of the particles constituting the calculated matrix phase and second phase are taken as the crystal grain sizes of the matrix phase and the second phase.

[0033] The binarization processing unit is controlled by the above-described control unit in the same manner as the image acquisition unit, the matrix phase determination unit, the second phase determination unit, and the measurement unit. Further, the binarized matrix phase image and the binarized second phase image generated by the binarization processing unit are stored in the above-described storage unit or a database on the server.

[0034] Instead of the binarization processing unit that generates a binarized matrix phase image and a binarized second phase image by binarizing the matrix phase determination image and the second phase determination image, a color separation processing unit that color-separates the matrix phase and the second phase into different color tones may be used. Even in a sintered body having a plurality of second phases, it is possible to measure the crystal grain size for each second phase.

[0035] Furthermore, the image processing apparatus of the present invention is an image processing apparatus for a captured image of a crystal structure of a sintered body having a matrix phase and a second phase different from the matrix phase, the image acquisition unit for acquiring the captured image, a known captured image, and a matrix phase data set including the crystal grains and matrix grain boundaries of the matrix phase in the known captured image, or a matrix phase setting image in which the matrix grain boundaries of the matrix phase are set, and uses the learned model generated by performing machine learning using the data set as teacher data to determine the crystal grains and matrix grain boundaries of the matrix phase, or the matrix grain boundaries in the captured image, and a matrix phase determination unit for generating a matrix phase determination image, a second phase data set including the known captured image, the crystal grains and second phase grain boundaries of the second phase in the known captured image, or a second phase setting image in which the second phase grain boundaries of the second phase are set, and uses the learned model generated by performing machine learning using the data set as teacher data to determine the crystal grains and second phase grain boundaries of the second phase, or the second phase grain boundaries in the captured image, and a second phase determination unit for generating a second phase determination image. As another embodiment of the image processing apparatus of the present invention described above, a configuration may be adopted in which a matrix phase determination image and a second phase determination image are generated using a captured image of a crystal structure of a sintered body having a matrix phase and a second phase different from the matrix phase. Since the image acquisition unit, the matrix phase determination unit, and the second phase determination unit according to another embodiment of the image processing apparatus of the present invention are as described above, detailed description thereof will be omitted.

[0036] Furthermore, the image processing apparatus of the present invention is an image processing apparatus for a captured image of a crystal structure of a sintered body having a matrix phase, the image acquisition unit for acquiring the captured image, a known captured image, and a matrix phase data set including the crystal grains and matrix grain boundaries of the matrix phase in the known captured image, or a matrix phase setting image in which the matrix grain boundaries of the matrix phase are set, and uses the learned model generated by performing machine learning using the data set as teacher data to determine the crystal grains and matrix grain boundaries of the matrix phase, or the matrix grain boundaries in the captured image, and a matrix phase determination unit for generating a matrix phase determination image. As yet another embodiment of the image processing apparatus of the present invention described above, a configuration may be adopted in which, for a captured image of the crystal structure of a sintered body composed only of a matrix phase, crystal grains and grain boundaries of the matrix phase, or grain boundaries of the matrix phase in the captured image are determined to generate a matrix phase determination image. Note that since the image acquisition unit and the matrix phase determination unit according to yet another embodiment of the image processing apparatus of the present invention are as described above, detailed description thereof is omitted.

[0037] Further, the image processing apparatus of the present invention is an image processing apparatus for a captured image of the crystal structure of a sintered body having a matrix phase and a second phase different from the matrix phase, and includes an image acquisition unit that acquires the captured image, a known captured image, crystal grains of the second phase and grain boundaries of the second phase, or a second phase setting image in which grain boundaries of the second phase are set in the known captured image. A second phase determination unit that determines crystal grains of the second phase and grain boundaries of the second phase, or grain boundaries of the second phase in the captured image and generates a second phase determination image using a learned model generated by performing machine learning using the second phase dataset as teacher data. As yet another embodiment of the image processing apparatus of the present invention described above, a configuration may be adopted in which, for a captured image of the crystal structure of a sintered body having a matrix phase and a second phase different from the matrix phase, crystal grains of the second phase and grain boundaries of the second phase, or grain boundaries of the second phase in the captured image are determined to generate a second phase determination image. Note that since the image acquisition unit and the second phase determination unit according to yet another embodiment of the image processing apparatus of the present invention are as described above, detailed description thereof is omitted.

[0038] Further, the image processing apparatus of the present invention further includes a binarization processing unit that generates a binarized matrix phase image and / or a binarized second phase image by binarization processing using the matrix phase determination image and / or the second phase determination image, and a measurement unit that measures the crystal grain size of the matrix phase and / or the second phase using the binarized matrix phase image and / or the binarized second phase image. In another embodiment of the image processing apparatus of the present invention described above, a binarization processing unit that binarizes the generated matrix phase determination image and / or the second phase determination image to generate a matrix phase binarized image and / or a second phase binarized image, and a measurement unit that measures the crystal grain size of the matrix phase and / or the second phase using the matrix phase binarized image and / or the second phase binarized image may be further provided. Since the binarization processing unit and the measurement unit according to another embodiment of the image processing apparatus of the present invention are as described above, detailed description thereof will be omitted.

[0039] The image processing method of the present invention is an image processing method for a captured image of a crystal structure of a sintered body having a matrix phase and a second phase different from the matrix phase, the method including: a step of obtaining the captured image; a machine learning performed using, as teacher data, a matrix phase data set including a known captured image, crystal grains and crystal grain boundaries of the matrix phase in the known captured image, or a matrix phase setting image in which crystal grain boundaries of the matrix phase are set, to generate a learned model, and determining crystal grains and crystal grain boundaries of the matrix phase or crystal grain boundaries of the matrix phase in the captured image using the learned model to generate a matrix phase determination image; a machine learning performed using, as teacher data, a second phase data set including the known captured image, crystal grains and crystal grain boundaries of the second phase in the known captured image, or a second phase setting image in which crystal grain boundaries of the second phase are set, to generate a learned model, and determining crystal grains and crystal grain boundaries of the second phase or crystal grain boundaries of the second phase in the captured image using the learned model to generate a second phase determination image; a step of generating a matrix phase binarized image and a second phase binarized image obtained by binarizing the matrix phase determination image and the second phase determination image; and a step of measuring crystal grain sizes of the matrix phase and the second phase using the matrix phase binarized image and the second phase binarized image.

[0040] First, the image acquisition unit acquires a captured image of a crystal structure of a sintered body having a matrix phase and a second phase different from the matrix phase. Then, the image acquisition unit outputs the acquired captured image to the storage unit of the image processing apparatus of the present invention or a database on the server.

[0041] Here, when the image processing method of the present invention photographs a cross-section of a sintered body with a scanning electron microscope, it is preferable in that the grain boundaries become clearer if it further has a step of generating an SEM image of the cross-section of the sintered body that has been etched with aqua regia. Further, if it further has a step of adjusting the contrast of the SEM image to generate a photographed image, it is preferable in that the grain boundaries become even clearer and precise and accurate drawing of the grain boundaries becomes possible.

[0042] Next, the matrix phase determination unit uses a learned model generated by performing machine learning using, as teacher data, a matrix phase dataset including a known photographed image, the matrix phase crystal grains and matrix phase grain boundaries in the known photographed image, or a matrix phase setting image in which the matrix phase grain boundaries are set, to determine the matrix phase crystal grains and matrix phase grain boundaries, or the matrix phase grain boundaries in the photographed image, and generates a matrix phase determination image. Further, the matrix phase determination unit outputs the generated matrix phase determination image to the storage unit of the image processing apparatus of the invention or to a database on a server.

[0043] Specifically, when the sintered body photographed in the photographed image is an ITO sintered body, the matrix phase determination unit uses a matrix phase learned model generated by performing machine learning using, as teacher data, a matrix phase dataset including a known photographed image of the ITO sintered body, the matrix phase crystal grains and matrix phase grain boundaries in the known photographed image, or a matrix phase setting image in which the matrix phase grain boundaries are set, to determine the crystal grains of the matrix phase (for example, In2O3 phase) and the matrix phase grain boundaries, or the matrix phase grain boundaries in the photographed image, and generates a matrix phase determination image.

[0044] Further, the second-phase determination unit uses a learned model generated by performing machine learning using a second-phase dataset including a known captured image, crystal grains of the second phase and grain boundaries of the second phase in the known captured image, or a second-phase setting image in which the grain boundaries of the second phase are set, as teaching data, to determine crystal grains of the second phase and grain boundaries of the second phase, or grain boundaries of the second phase in the captured image, and generates a second-phase determination image. Further, the second-phase determination unit outputs the generated second-phase determination image to a storage unit of the image processing apparatus of the invention or a database on a server.

[0045] Specifically, when the sintered body captured in the captured image is an ITO sintered body, the second-phase determination unit uses a second-phase learned model generated by performing machine learning using a second-phase dataset including a known captured image of the ITO sintered body, crystal grains of the second phase and grain boundaries of the second phase in the known captured image, or a second-phase setting image in which the grain boundaries of the second phase are set, as teaching data, to determine crystal grains of the second phase (for example, In4Sn3O 12 phase) and grain boundaries of the second phase, or grain boundaries of the second phase in the captured image, and generates a second-phase determination image.

[0046] Further, the binarization processing unit generates a binarized matrix phase image and a binarized second-phase image by performing binarization processing on the matrix phase determination image and the second-phase determination image. Further, the binarization processing unit outputs the generated binarized matrix phase image and binarized second-phase image to a storage unit of the image processing apparatus of the present invention or a database on a server.

[0047] Then, the measurement unit measures the crystal grain sizes of the matrix phase and the second phase using the binarized matrix phase image and the binarized second-phase image. Further, the measurement unit outputs the measured crystal grain sizes of the matrix phase and the second phase to a storage unit of the image processing apparatus of the present invention or a database on a server.

[0048] Specifically, the measurement unit performs particle analysis on the matrix binary image and the second-phase binary image, and obtains the areas of the particles constituting the matrix and the second phase, respectively. Next, the equivalent circle diameters of the areas are calculated from the areas of the particles constituting the obtained matrix and the second phase. Then, the equivalent circle diameters of the areas of the particles constituting the calculated matrix and the second phase are used as the crystal grain sizes of the matrix and the second phase.

[0049] In addition, in this specification, when expressing "X to Y" (X and Y are arbitrary numbers), unless otherwise specified, it includes the meaning of "X or more and Y or less", and also includes the meaning of "preferably larger than X" or "preferably smaller than Y". Further, when expressing "X or more" (X is an arbitrary number) or "Y or less" (Y is an arbitrary number), it also includes the intention of "preferably larger than X" or "preferably less than Y".

Advantages of the Invention

[0050] The image processing apparatus of the present invention enables rapid and highly accurate analysis. Further, the image processing method of the present invention enables rapid and highly accurate analysis.

Brief Description of the Drawings

[0051]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

BEST MODE FOR CARRYING OUT THE INVENTION

[0052] Hereinafter, the image processing apparatus according to the embodiment of the present invention will be further described. However, the following embodiments do not limit the present invention.

[0053] (First Embodiment) FIG. 1 is a diagram showing a configuration example of the image processing apparatus according to the first embodiment. As shown in FIG. 1, the image processing apparatus 10 according to the first embodiment is communicably connected to the scanning electron microscope 20, regardless of whether it is wired or wireless.

[0054] In addition, the image processing apparatus 10 according to the first embodiment includes an image acquisition unit 11, a matrix phase determination unit 12, a second phase determination unit 13, and a measurement unit 14. Further, the image processing apparatus 10 according to the first embodiment includes a control unit (not shown) that controls the image acquisition unit 11, the matrix phase determination unit 12, the second phase determination unit 13, and the measurement unit 14, and a storage unit 31 that stores captured images, matrix phase setting images, matrix phase learned models, second phase setting images, second phase learned models, matrix phase determination images, second phase determination images, and various programs. The image processing apparatus 10 also includes a communication unit 32 that can communicate with a scanning electron microscope 20 and other information communication terminals, an input unit 33 that accepts input from an operator, and a display unit 34 that can display captured images, matrix phase setting images, second phase setting images, matrix phase determination images, second phase determination images, and the like.

[0055] The image processing method by the image processing apparatus according to the first embodiment will be described below.

[0056] First, the image acquisition unit 11 acquires, via the communication unit 32, a captured image of the crystal structure of a sintered body having a matrix phase and a second phase different from the matrix phase, which is captured by the scanning electron microscope 20. Then, the image acquisition unit 11 stores the acquired captured image in the storage unit 31. An example of the captured image is shown in FIG. 2.

[0057] Next, the matrix phase determination unit 12 calls a matrix phase learned model generated by performing machine learning using, as teacher data, a matrix phase data set including a known captured image stored in the storage unit 31 and a matrix phase setting image in which the grain boundaries of the matrix phase in the known captured image are set, and the captured image acquired by the image acquisition unit 11. Then, the matrix phase determination unit 12 determines the grain boundaries of the matrix phase in the captured image using the matrix phase learned model, and generates a matrix phase determination image as shown in FIG. 3(a). Then, the matrix phase determination unit 12 stores the generated matrix phase determination image in the storage unit 31. Here, FIG. 4(a) is a diagram schematically showing the matrix phase determination image shown in FIG. 3(a). The region indicated by the hatching in FIG. 4(a) indicates the region determined as the matrix phase.

[0058] Note that the parent phase determination unit 12 may determine the crystal grains and crystal grain boundaries of the parent phase in the captured image by using a machine learning model trained with a parent phase dataset including known captured images stored in the storage unit 31 and parent phase setting images in which the crystal grains and crystal grain boundaries of the parent phase in the known captured images are set as teacher data, and generate a parent phase determination image (see Fig. 5(a)).

[0059] Also, the secondary phase determination unit 13 calls a secondary phase trained model generated by performing machine learning using, as teacher data, a secondary phase dataset including known captured images stored in the storage unit 31 and secondary phase setting images in which the crystal grains and crystal grain boundaries of the secondary phase in the known captured images are set, and the captured image acquired by the image acquisition unit 11. Then, the secondary phase determination unit 13 determines the crystal grains and crystal grain boundaries of the secondary phase in the captured image by using the secondary phase trained model, and generates a secondary phase determination image as shown in Fig. 3(b). Then, the secondary phase determination unit 13 stores the generated secondary phase determination image in the storage unit 31. Here, Fig. 4(b) is a diagram schematically showing the secondary phase determination image shown in Fig. 3(b). The region indicated by the hatching in Fig. 4(b) indicates the region determined as the secondary phase.

[0060] Note that the secondary phase determination unit 13 may determine the crystal grain boundaries of the secondary phase by using a secondary phase trained model generated by performing machine learning using, as teacher data, a secondary phase dataset including known captured images stored in the storage unit 31 and secondary phase setting images in which the crystal grain boundaries of the secondary phase in the known captured images are set, and generate a secondary phase determination image (see Fig. 5(b)).

[0061] Then, the measurement unit 14 calls the matrix phase determination image and the second phase determination image stored in the storage unit 31, performs particle analysis on the matrix phase determination image and the second phase determination image, and obtains the areas of the particles constituting the matrix phase and the second phase respectively. Next, the measurement unit 14 calculates the equivalent circle diameters of the respective areas from the areas of the particles constituting the obtained matrix phase and the second phase. Then, the measurement unit 14 uses the calculated equivalent circle diameters of the particles constituting the matrix phase and the second phase as the crystal grain sizes of the matrix phase and the second phase. Note that the measurement unit 14 stores the measured crystal grain sizes of the matrix phase and the second phase in the storage unit 31.

[0062] (Second Embodiment) FIG. 6 is a diagram showing a configuration example of the image processing apparatus according to the second embodiment. The image processing apparatus 10A according to the second embodiment has a binarization processing unit 15 in addition to the configuration of the image processing apparatus 10 according to the first embodiment. Note that the same components as those in the configuration of the image processing apparatus 10 according to the first embodiment are denoted by the same reference numerals, and detailed descriptions thereof are omitted.

[0063] The binarization processing unit 15 uses the matrix phase determination image generated by the matrix phase determination unit 12 and the second phase determination image generated by the second phase determination unit 13 to generate a binarized matrix phase binarized image (see FIG. 7(a)) and a second phase binarized image (see FIG. 7(b)). For example, the region shown in white in FIG. 7(a) is the region determined to be the crystal grain boundary of the matrix phase by the matrix phase determination unit 12 and is the region binarized by the binarization processing unit 15. On the other hand, for example, the region shown in white in FIG. 7(b) is the region determined to be the crystal grains and the crystal grain boundaries of the second phase by the second phase determination unit 13 and is the region binarized by the binarization processing unit 15. Then, the binarization processing unit 15 stores the generated matrix phase binarized image and second phase binarized image in the storage unit 31.

[0064] Note that when the matrix phase determination unit 12 determines the crystal grains and crystal grain boundaries of the matrix phase and generates a matrix phase determination image, the binarization processing unit 15 may binarize the regions determined as the crystal grains and crystal grain boundaries of the matrix phase to generate a matrix phase binarized image. Further, when the second phase determination unit 13 determines the crystal grain boundaries of the second phase and generates a second phase determination image, the binarization processing unit 15 may binarize the regions determined as the crystal grain boundaries of the second phase to generate a second phase binarized image.

[0065] Next, the measurement unit 14A performs particle analysis on the matrix phase binarized image and the second phase binarized image generated by the binarization processing unit 15 to obtain the areas of the particles constituting the matrix phase and the second phase respectively. Next, the equivalent circle diameters of the areas are calculated from the obtained areas of the particles constituting the matrix phase and the second phase. Then, the equivalent circle diameters of the areas of the particles constituting the calculated matrix phase and the second phase are used as the crystal grain sizes of the matrix phase and the second phase. Note that the measurement unit 14A stores the measured crystal grain sizes of the matrix phase and the second phase in the storage unit 31.

[0066] As described above, the image processing apparatus 10A according to the second embodiment can measure the area ratios of the matrix phase and the second phase by performing the binarization processing by the binarization processing unit 15, and can qualitatively determine the quality of the sputtering target.

[0067] Here, a graph showing the crystal grain sizes of the matrix phase and the second phase calculated from the photographed image of the ITO sintered body, and a graph showing the crystal grain sizes of the matrix phase and the second phase measured from the photographed image of the ITO sintered body by a conventional operator himself / herself using the image processing apparatus 10A according to the second embodiment having the above-described configuration are shown in FIGS. 8 and 9.

[0068] Example 1 in Fig. 8(a) is the crystal grain size of the matrix phase calculated for 20 captured images (sample numbers 1 to 20 in Fig. 8(a)) of an ITO sintered body with a predetermined composition ratio using the image processing apparatus 10A according to the second embodiment of the present invention. On the other hand, Comparative Example 1 in Fig. 8(a) is the crystal grain size of the matrix phase measured for 20 captured images of the same ITO sintered body as in Example 1 by the operator himself / herself. Further, Example 1 in Fig. 8(b) is the crystal grain size of the second phase calculated for 20 captured images (sample numbers 1 to 20 in Fig. 8(b)) of an ITO sintered body with a predetermined composition ratio using the image processing apparatus 10A according to the second embodiment of the present invention. On the other hand, Comparative Example 1 in Fig. 8(b) is the crystal grain size of the second phase measured for 20 captured images of the same ITO sintered body as in Example 1 by the operator himself / herself. Note that the captured images of the ITO sintered body used in Example 1 and Comparative Example 1 are secondary electron images capturing different locations of the same ITO sintered body.

[0069] As shown in Fig. 8(a), the maximum value of the absolute value of the difference in the crystal grain size of the matrix phase between Example 1 and Comparative Example 1 was 0.12 μm, the minimum value of the absolute value of the difference was 0.0 μm, and the average value of the absolute value of the difference was 0.0535 μm. Also, as shown in Fig. 8(b), the maximum value of the absolute value of the difference in the crystal grain size of the second phase between Example 1 and Comparative Example 1 was 0.08 μm, the minimum value of the absolute value of the difference was 0.0 μm, and the average value of the absolute value of the difference was 0.027 μm.

[0070] Example 2 in Fig. 9(a) is the crystal grain size of the matrix phase calculated from 20 captured images (sample numbers 1 to 20 in Fig. 9(a)) of an ITO sintered body with a composition ratio different from that of Example 1 using the image processing apparatus 10A according to the second embodiment of the present invention. On the other hand, Comparative Example 2 in Fig. 9(a) is the crystal grain size of the matrix phase measured from 20 captured images of the same ITO sintered body as Example 2 by the operator himself / herself. Further, Example 2 in Fig. 9(b) is the crystal grain size of the second phase calculated from 20 captured images (sample numbers 1 to 20 in Fig. 9(b)) of an ITO sintered body with a composition ratio different from that of Example 1 using the image processing apparatus 10A according to the second embodiment of the present invention. On the other hand, Comparative Example 2 in Fig. 9(b) is the crystal grain size of the second phase measured from 20 captured images of the same ITO sintered body as Example 2 by the operator himself / herself. Note that the captured images of the ITO sintered body used in Example 2 and Comparative Example 2 are secondary electron images of different locations of the same ITO sintered body.

[0071] As shown in Fig. 9(a), the maximum value of the absolute value of the difference in the crystal grain size of the matrix phase between Example 2 and Comparative Example 2 was 0.08 μm, the minimum value of the absolute value of the difference was 0.0 μm, and the average value of the absolute value of the difference was 0.04 μm. Further, as shown in Fig. 9(b), the maximum value of the absolute value of the difference in the crystal grain size of the eutectic phase different from the matrix phase between Example 2 and Comparative Example 2 was 0.09 μm, the minimum value of the absolute value of the difference was 0.01 μm, and the average value of the absolute value of the difference was 0.039 μm.

[0072] As shown in Figs. 8 and 9, there was almost no difference between the crystal grain size of the matrix phase calculated from the captured image of the ITO sintered body using the image processing apparatus 10A according to the second embodiment and the crystal grain size of the matrix phase measured from the captured image of the ITO sintered body by the conventional operator himself / herself. Further, there was almost no difference in the crystal grain size of the second phase either. Furthermore, even in ITO sintered bodies with different composition ratios, there was almost no difference in the crystal grain size of the matrix phase and the crystal grain size of the second phase.

[0073] In this way, it was confirmed that both had the same level of accuracy. On the other hand, while the series of processing times by the image processing apparatus 10A according to the second embodiment was about 1 minute, the series of processing times by the operator himself was 140 minutes, so a significant reduction in time could be achieved.

[0074] (Third Embodiment) FIG. 10 is a diagram showing a configuration example of an image processing apparatus according to the third embodiment. The image processing apparatus 10B according to the third embodiment includes an image acquisition unit 11, a matrix phase determination unit 12, and a second phase determination unit 13. Also, for the same configurations as those of the image processing apparatus 10 according to the first embodiment, the same reference numerals are given, and detailed descriptions thereof are omitted.

[0075] The image processing apparatus 10B according to the third embodiment is preferable in that it can be processed quickly by limiting it to the identification of the matrix phase or the second phase.

[0076] Also, the image processing apparatus 10B according to the third embodiment may have configurations corresponding to the binarization processing unit 15 and the measurement unit 14A of the image processing apparatus 10A according to the second embodiment.

[0077] (Fourth Embodiment) FIG. 11 is a diagram showing a configuration example of an image processing apparatus according to the fourth embodiment. The image processing apparatus 10C according to the fourth embodiment includes an image acquisition unit 11 and a matrix phase determination unit 12. For the same configurations as those of the image processing apparatus 10 according to the first embodiment, the same reference numerals are given, and detailed descriptions thereof are omitted.

[0078] The image processing apparatus 10C according to the fourth embodiment is preferable in that it can be processed quickly by limiting it to the identification of the matrix phase, and also in that the teaching data can be limited to only the matrix phase. Also, for a sintered body that is the object of image processing and is a single-phase matrix sintered body or a sintered body in which three or more second phases coexist, it is possible to perform image processing only on the matrix phase.

[0079] Further, the image processing apparatus 10C according to the fourth embodiment may have a configuration corresponding to the binarization processing unit 15 and the measurement unit 14A of the image processing apparatus 10A according to the second embodiment.

[0080] (Fifth Embodiment) FIG. 12 is a diagram showing a configuration example of an image processing apparatus according to the fifth embodiment. The image processing apparatus 10D according to the fifth embodiment includes an image acquisition unit 11 and a second-phase determination unit 13. For the same configurations as those of the image processing apparatus 10 according to the first embodiment, the same reference numerals are given, and detailed descriptions thereof are omitted.

[0081] The image processing apparatus 10D according to the fifth embodiment is preferable in that it can be processed quickly by limiting it to the identification of the second phase, and the training data can be limited to only the second phase.

[0082] Further, the image processing apparatus 10D according to the fifth embodiment may have a configuration corresponding to the binarization processing unit 15 and the measurement unit 14A of the image processing apparatus 10A according to the second embodiment.

[0083] The invention disclosed in this specification, in addition to the configurations of each invention and embodiment, within the applicable range, includes those in which these partial configurations are changed to other configurations disclosed in this specification and specified, or those in which other configurations disclosed in this specification are added to these configurations and specified, or those in which these partial configurations are deleted to the extent that partial operational effects can be obtained and a higher-level concept is specified.

Industrial Applicability

[0084] The image processing apparatus according to the present invention enables image processing of a captured image in which the crystal structure of a sintered body is simply captured without requiring a complicated preprocessing step or complicated apparatus analysis. For example, by measuring the crystal grain size of the matrix phase and / or the matrix phase and the second phase, defective sintered bodies can be found in advance. Specifically, in the case of a sintered body for sputtering target use, a sputtering target that is highly likely to cause defects during sputtering can be removed at the sintered body stage in advance, the generation of a sputtering film forming body that is a defective product can be prevented, and this leads to a reduction in waste. These aspects lead to the sustainable management of natural resources and efficient advantages, as well as the achievement of decarbonization (carbon neutrality).

Explanation of Signs

[0085] 10, 10A, 10B, 10C, 10D… Image processing apparatus 11… Image acquisition unit 12… Matrix phase determination unit 13… Second phase determination unit 14, 14A… Measurement unit 15… Binarization processing unit 20… Scanning electron microscope 31… Storage unit 32… Communication unit 33… Input unit 34… Display unit

Claims

1. An image processing apparatus for a captured image of a crystal structure of a sintered body having a matrix phase and a second phase different from the matrix phase, comprising: an image acquisition unit that acquires the captured image; using a machine learning model trained with a dataset of known captured images, crystal grains and grain boundaries of the matrix phase in the known captured images, or a matrix phase setting image with grain boundaries of the matrix phase set as teacher data, determining crystal grains and grain boundaries of the matrix phase or grain boundaries of the matrix phase in the captured image, and generating a matrix phase determination image; using a second-phase trained model generated by performing machine learning using a second-phase dataset including the known captured image, crystal grains and grain boundaries of the second phase in the known captured image, or a second-phase setting image with grain boundaries of the second phase set as teacher data, determining crystal grains and grain boundaries of the second phase or grain boundaries of the second phase in the captured image, and generating a second-phase determination image; a measurement unit that measures crystal grain sizes of the matrix phase and the second phase using the matrix phase determination image and the second-phase determination image; An image processing apparatus, characterized by comprising the above components.

2. further comprising a binarization processing unit that generates a binarized matrix phase image and a binarized second-phase image by binarization processing using the matrix phase determination image and the second-phase determination image; The image processing apparatus according to claim 1, wherein the measurement unit measures crystal grain sizes of the matrix phase and the second phase using the binarized matrix phase image and the binarized second-phase image.

3. An image processing apparatus for a captured image of a crystal structure of a sintered body having a matrix phase and a second phase different from the matrix phase, comprising: an image acquisition unit that acquires the captured image; using a trained model generated by performing machine learning using a dataset of known captured images, crystal grains and grain boundaries of the matrix phase in the known captured images, or a matrix phase setting image with grain boundaries of the matrix phase set as teacher data, determining crystal grains and grain boundaries of the matrix phase or grain boundaries of the matrix phase in the captured image, and generating a matrix phase determination image; Using a learned model generated by performing machine learning using, as teacher data, a second-phase data set including the known captured image, crystal grains of the second phase and grain boundaries of the second phase in the known captured image, or a second-phase setting image in which the grain boundaries of the second phase are set, a second-phase determination unit that determines crystal grains of the second phase and grain boundaries of the second phase, or grain boundaries of the second phase in the captured image and generates a second-phase determination image; An image processing apparatus, characterized by comprising the same. **Claim 4** An image processing apparatus for a captured image of a crystal structure of a sintered body having a matrix phase, an image acquisition unit that acquires the captured image; Using a learned model generated by performing machine learning using, as teacher data, a matrix-phase data set including a known captured image, crystal grains of the matrix phase and grain boundaries of the matrix phase in the known captured image, or a matrix-phase setting image in which the grain boundaries of the matrix phase are set, a matrix-phase determination unit that determines crystal grains of the matrix phase and grain boundaries of the matrix phase, or grain boundaries of the matrix phase in the captured image and generates a matrix-phase determination image; An image processing apparatus, characterized by comprising the same. **Claim 5** An image processing apparatus for a captured image of a crystal structure of a sintered body having a matrix phase and a second phase different from the matrix phase, an image acquisition unit that acquires the captured image; Using a learned model generated by performing machine learning using, as teacher data, a second-phase data set including the known captured image, crystal grains of the second phase and grain boundaries of the second phase in the known captured image, or a second-phase setting image in which the grain boundaries of the second phase are set, a second-phase determination unit that determines crystal grains of the second phase and grain boundaries of the second phase, or grain boundaries of the second phase in the captured image and generates a second-phase determination image; An image processing apparatus, characterized by comprising the same. **Claim 6** a binarization processing unit that generates a binarized matrix-phase image and / or a binarized second-phase image by using the matrix-phase determination image and / or the second-phase determination image; a measurement unit that measures the crystal grain size of the matrix phase and / or the second phase by using the binarized matrix-phase image and / or the binarized second-phase image; The image processing apparatus according to any one of claims 3 to 5, further comprising the same. **Claim 7** The image processing apparatus according to any one of claims 1 to 5, characterized in that the sintered body is a sintered body for a sputtering target. **Claim 8** wherein the sintered body has a matrix phase of In ２ O ３ phase and a second phase of In ４ Sn ３ O １２ phase, and the image processing apparatus according to any one of claims 1 to 5, characterized in that it is an ITO sintered body.

9. The image processing apparatus according to any one of claims 1 to 5, wherein the captured image is any one of a secondary electron image, a backscattered electron image, and a stereoscopic image captured by a scanning electron microscope.

10. The image processing apparatus according to any one of claims 1 to 5, wherein the captured image is an image of a cross-section of the sintered body etched with aqua regia.

11. The image processing apparatus according to any one of claims 1 to 5, wherein the captured image is an image with adjusted contrast.

12. An image processing method for a captured image of a crystal structure of a sintered body having a matrix phase and a second phase different from the matrix phase, obtaining the captured image; using a learned model generated by performing machine learning using, as teacher data, a known captured image and a matrix phase dataset including crystal grains and crystal grain boundaries of the matrix phase in the known captured image, or a matrix phase setting image in which the crystal grain boundaries of the matrix phase are set, to determine crystal grains and crystal grain boundaries of the matrix phase in the captured image, or crystal grain boundaries of the matrix phase, and generating a matrix phase determination image; using a learned model generated by performing machine learning using, as teacher data, a known captured image and a second phase dataset including crystal grains and crystal grain boundaries of the second phase in the known captured image, or a second phase setting image in which the crystal grain boundaries of the second phase are set, to determine crystal grains and crystal grain boundaries of the second phase in the captured image, or crystal grain boundaries of the second phase, and generating a second phase determination image; generating a binarized matrix phase image and a binarized second phase image by performing binarization processing on the matrix phase determination image and the second phase determination image; measuring crystal grain sizes of the matrix phase and the second phase using the binarized matrix phase image and the binarized second phase image; An image processing method characterized by comprising the steps of.

13. generating an SEM image of a cross-section of a sintered body etched with aqua regia; adjusting the contrast of the SEM image to generate the captured image; The image processing method according to claim 12, characterized by comprising the steps of.

Images