Inspection device, learning data generation method, learning model generation method, and method for manufacturing object

Abstract

To provide a technique capable of specifying a type of a defect by using a learning model.SOLUTION: An inspection device includes: a specification part for specifying a defect position in a learning object having a predetermined defect by using wide-area learning image data representing the learning object; a microscopic image generation part for generating microscopic learning image data representing the predetermined defect by micro-imaging the specified defect position; and an acquisition part for acquiring a combination of type information representing the type of the predetermined defect obtained by using the generated microscopic learning image data and defect image data representing the predetermined defect included in the wide-area learning image data. The specification part specifies the type of the defect in an inspection object through the use of wide-area inspection image data representing the inspection object by using a learning model generated by machine learning using the acquired combination of the type information and the defect image data as learning data.SELECTED DRAWING: Figure 6

Description

Technical Field

[0001] The technology disclosed in this specification relates to inspecting abnormalities of an object.

Background Art

[0002] Patent Document 1 discloses a technology for detecting defects in a product. In this technology, image data representing a good product or a defective product is used as teaching data, and a learning model is generated by machine learning such as a neural network. Using the learning model, good products and defective products are classified from the image data of the product.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] When a defect is included in an object to be inspected, it may be necessary to specify the type of defect indicating what kind of defect it is.

[0005] This specification provides a technology capable of specifying the type of defect by using a learning model.

Means for Solving the Problems

[0006] The first aspect disclosed in this specification relates to an inspection apparatus. The inspection apparatus includes a specifying unit that specifies a defect position in the learning object of a predetermined defect using wide-area learning image data representing the learning object, a microscopic image generation unit that generates microscopic learning image data representing the predetermined defect by microscopically imaging the specified defect position, an acquisition unit that acquires a combination of type information representing the type of the predetermined defect obtained using the generated microscopic learning image data and defect image data representing the predetermined defect included in the wide-area learning image data, and the specifying unit uses a learning model generated by machine learning using the acquired combination of the type information and the defect image data as learning data to specify the type of a defect in the inspection object using wide-area inspection image data representing the inspection object.

[0007] According to this configuration, it is possible to generate microscopic learning image data in which a predetermined defect with a specified defect position is magnified by microscopically imaging the defect. By using the microscopic learning image data, it is possible to easily specify the type of a defect even if the defect is small. By generating a learning model using, as learning data, a combination of the type of a defect obtained using the microscopic learning image data and the defect image data, it is possible to specify the type of a defect even when the size of the defect is small using the learned learning model.

[0008] In a second aspect, in the above first aspect, the learning object includes a glass plate, the inspection apparatus further includes an inspection image generation unit that generates the wide-area learning image data, and the inspection image generation unit includes a bright-field wide-area imaging unit that generates bright-field wide-area image data representing the glass plate by imaging, in a field of view including a bright field, first transmitted light that has passed through the glass plate, and a dark-field wide-area imaging unit that generates dark-field wide-area image data representing the glass plate by imaging, in a field of view including a dark field, second transmitted light that is incident on the glass plate at an angle different from that of the first transmitted light and has passed through the glass plate, and the wide-area learning image data may include the bright-field wide-area image data and the dark-field wide-area image data.

[0009] Defects of various types such as small surface irregularities, air bubbles, and foreign substances can occur on the glass plate. Depending on the type of these defects, the appearance of the defect images when imaging in a field of view including a bright field and a field of view including a dark field is different. According to the above configuration, by using the bright-field wide-area image data captured in a field of view including a bright field and the dark-field wide-area image data captured in a field of view including a dark field, it is possible to easily identify the type of defect.

[0010] In the third form, in the above first or second aspect, the object to be learned includes a glass plate, and the microscopic image generation unit includes a bright-field microscopic imaging unit that generates bright-field microscopic image data representing the glass plate by imaging the reflected light reflected by the glass plate in a field of view including a bright field, and a dark-field microscopic imaging unit that generates dark-field microscopic image data representing the glass plate by imaging the scattered light scattered by the defect in a field of view including a dark field. The microscopic learning image data may include the bright-field microscopic image data and the dark-field microscopic image data.

[0011] According to the above configuration, by using the bright-field microscopic image data captured in a field of view including a bright field and the dark-field microscopic image data captured in a field of view including a dark field, it is possible to easily identify the type of defect.

[0012] In the fourth form, in any one of the above first to third aspects, the microscopic image generation unit includes a camera unit that microscopically images the predetermined defect, and a moving unit that moves the camera unit to the defect position. When the object to be learned includes a plurality of defects including one or more of the predetermined defects, the microscopic learning image data of a number of the predetermined defects less than the number of the plurality of defects may be generated.

[0013] According to the above configuration, by moving the camera unit, a predetermined defect can be appropriately imaged. Furthermore, it is not necessary to use all of the plurality of defects included in the object as learning data. As a result, the time required to generate the microscopic learning image data can be shortened, and the burden of generating the learning data and the learning period in the learning model can be suppressed.

[0014] In the fifth aspect, in the above-described fourth aspect, the microscopic image generation unit may generate the microscopic learning image data of the predetermined defects in ascending order of size from among the predetermined defects.

[0015] Defects with relatively small sizes are more difficult to identify in terms of type compared to defects with relatively large sizes. According to the above configuration, by selectively adopting image data representing defects with relatively small sizes as learning data, it is possible to improve the accuracy of identifying the types of defects with relatively small sizes.

[0016] In the sixth aspect, in any one of the first to fifth aspects described above, the object to be learned and the object to be inspected may include a glass plate used for a display device.

[0017] The smaller the size of the pixels displayed on the display device, the smaller the size of the defects allowed in the glass plate used for the display device. The above-described inspection device can be suitably used to identify the types of small defects in the glass plate used for the display device.

[0018] In the seventh aspect, in the above-described sixth aspect, the wide-area inspection image data may include the entire effective surface of the glass plate.

[0019] In the glass plate used for the display device, in a subsequent process, it may be used as a product after cutting the peripheral portion. The effective surface means the surface used as the product. According to this configuration, it is possible to comprehensively identify the defects included in the effective surface of the glass plate.

[0020] This specification further discloses a method for generating training data for generating a learning model. The method for generating training data includes a wide-area learning image acquisition step of acquiring wide-area learning image data representing an object, a defect position acquisition step for learning of acquiring a defect position in the object of a predetermined defect of the object, a microscopic learning image generation step of generating microscopic learning image data representing the predetermined defect by microscopically imaging the acquired defect position, a storage step of storing a combination of type information representing the type of the predetermined defect obtained using the generated microscopic learning image data and defect image data representing the predetermined defect included in the wide-area learning image data.

[0021] According to this configuration, it is possible to generate microscopic learning image data in which a predetermined defect with a specified defect position is magnified by microscopically imaging the defect. By using the microscopic learning image data, it is possible to easily identify the type of the defect even if it is a small defect. It is possible to generate, as training data, a combination of the type of defect obtained using the microscopic learning image data and the defect image data. As a result, it is possible to use, as training data, the type of defect more accurately identified using a microscopic image in which a predetermined defect is magnified. As a result, the accuracy of identifying the type of defect by the learning model can be improved.

[0022] This specification further discloses a method for generating a learning model. The method for generating a learning model may include a wide-area learning image acquisition step of acquiring wide-area learning image data representing an object, a defect position acquisition step for learning of acquiring a defect position in the object of a predetermined defect of the object, a microscopic learning image generation step of generating microscopic learning image data representing the predetermined defect by microscopically imaging the acquired defect position, a storage step of storing a combination of type information representing the type of the predetermined defect obtained using the generated microscopic learning image data and defect image data representing the predetermined defect included in the wide-area learning image data, and a learning model generation step of generating a learning model by machine learning using the stored combination as training data.

[0023] According to this configuration, a highly accurate learning model can be generated using learning data generated using microscopic learning image data in which a predetermined defect is enlarged.

[0024] This specification further discloses a method for manufacturing an object. The method for manufacturing an object includes a learning defect type identification step of identifying a type of a predetermined defect in the learning object using wide-area learning image data representing the learning object, a position acquisition step of acquiring a defect position in the learning object of the predetermined defect using the wide-area learning image data, a microscopic learning image generation step of generating microscopic learning image data representing the predetermined defect by microscopically imaging the acquired defect position, a storage step of storing a combination of type information representing the type of the predetermined defect obtained using the generated microscopic learning image data and a defect image representing the predetermined defect included in the wide-area learning image data, and an inspection target defect type identification step of identifying the type of a defect in an inspection target object using wide-area inspection image data representing the inspection target object by using, as learning data, the combination of the stored type information and the defect image data generated by machine learning.

[0025] According to this configuration, the same effect as that of the described inspection apparatus can be achieved.

[0026] The manufacturing method may further include a quality determination step of determining the quality of the object based on defect information including at least one of the number of the defects included in the object, the dimensions of the defects, and the type of the defects identified in the inspection target defect type identification step.

[0027] According to this configuration, the quality of the object can be determined based on defect information corresponding to the quality required for the object.

[0028] The object includes a glass plate used for a display device, and the manufacturing method may further include a carrying-out step of sorting the object according to the quality based on the quality determined in the quality determination step.

[0029] According to this configuration, by sorting glass plates by quality, it is possible to ship glass plates of a quality that meets the required quality to the destination when the glass plates are shipped.

[0030] The manufacturing method may further include a defective product sorting step of distinguishing the object determined to be of poor quality in the quality determination step from the object not determined to be of poor quality.

[0031] According to this configuration, it is possible to prevent defective products from being included in the objects to be shipped.

Brief Description of the Drawings

[0032]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Best Mode for Carrying Out the Invention

[0033] (Configuration of the Manufacturing Unit) As shown in FIG. 1, the manufacturing unit 10 manufactures a glass plate G (see FIG. 2). The glass plate G is used, for example, for the cover or substrate of a display device. Note that the use of the glass plate G is not particularly limited. The glass plate G includes a learning glass plate G1 and an inspection target glass plate G2. Note that the learning glass plate G1 is an example of a "learning object", and the inspection target glass plate G2 is an example of an "inspection object".

[0034] The manufacturing unit 10 includes a forming device 12, a cooling device 14, a cutting device 16, an inspection device 18, and a transfer device 22. The forming device 12 stretches the molten glass downward to form a strip-shaped glass ribbon. The forming device 12 forms a glass ribbon from the molten glass melted in a melting furnace, for example, using the overflow down-draw method. Note that the forming device 12 is not particularly limited, and for example, a glass ribbon may be produced using other down-draw methods such as the slot down-draw method or the redraw method, or the float method.

[0035] The cooling device 14 gradually cools the glass ribbon extending from the forming device 12. The cooling device 14 includes a slow-cooling furnace in which a predetermined temperature gradient is provided downward in the internal space. The glass ribbon is gradually cooled so that its temperature decreases as it moves downward through the internal space of the slow-cooling furnace while being guided by a transfer device such as an annealing roller. Thereby, the strain of the glass ribbon is reduced. The cooling device 14 further includes a cooling chamber that allows the slowly cooled glass ribbon to cool to near room temperature.

[0036] The cutting device 16 cuts the glass ribbon into a predetermined length. The cutting device 16 clamps both ends of the glass ribbon, forms a scribe line using a cutter, and applies a bending stress along the scribe line to cut (i.e., sever) the glass ribbon along the scribe line. Thereby, a rectangular glass plate G of a predetermined length is obtained from the glass ribbon. Note that the cutting method used by the cutting device 16 is not limited to severing by bending stress, and may be, for example, laser severing or laser melting. Further, the cutting device 16 may further cut both ends in the width direction of the glass plate G. Both ends in the width direction of the glass plate G may be relatively thicker than the central portion in the width direction, and these both ends are called ears.

[0037] The glass plate G cut by the cutting device 16 is held by the conveying device 22. The glass plate G is conveyed to the inspection device 18 while being held by the conveying device 22. As shown in FIG. 2, the conveying device 22 includes clamping mechanisms 24 and 25 that respectively clamp the upper edge and the lower edge of the glass plate G, and a pair of rails 26, 26 extending in the conveying direction. The clamping mechanism 24 includes a plurality of chucks 24a that clamp the upper edge of the glass plate G, and a moving body 24b that movably attaches the plurality of chucks 24a to the rail 26. The clamping mechanism 25 includes a plurality of chucks 25a that clamp the lower edge of the glass plate G, and a moving body 25b that movably attaches the plurality of chucks 24a to the rail 26. The clamping mechanisms 24 and 25 convey the glass plate G along the rails 26, 26 while clamping the glass plate G. By performing the inspection while clamping the upper edge and the lower edge of the glass plate G, the shaking of the glass plate G can be reduced. Thereby, the deviation of the focus when imaging the glass plate G by the inspection device 18 can be reduced, and the type of defect can be accurately specified.

[0038] (Configuration of the inspection device) The inspection device 18 includes a detection device 19 and a microscopic image generation device 20. The detection device 19 detects defects existing in the glass plate G and identifies the defect positions and defect types. The microscopic image generation device 20 selects a predetermined number (for example, 2 to 3) of defects from among the defects detected by the detection device 19 and captures an enlarged image of the selected defects (hereinafter referred to as "predetermined defects"). Defects include bubbles and foreign substances encapsulated in the glass plate G during the manufacture of the glass plate G, deposits such as dust adhering to the surface of the glass plate G, damage and unevenness formed on the surface of the glass plate G. Note that the detection device 19 is an example of a "specifying unit", and the microscopic image generation device 20 is an example of a "microscopic image generation unit".

[0039] (Configuration of the detection device) The detection device 19 includes a control device 30 and a wide-area imaging unit 40. The control device 30 includes a control unit 32 and a communication interface (hereinafter referred to as "communication I / F") 36. The control unit 32 controls the detection device 19. The control unit 32 includes a CPU and a memory constituted by a non-volatile memory or the like. The CPU executes the processes described below according to the computer program stored in the memory. Note that the wide-area imaging unit 40 is an example of an "inspection image generation unit", and the image data of the image captured by the wide-area imaging unit 40 is an example of "wide-area learning image data P1" and "wide-area inspection image data P3".

[0040] In addition to the computer program, a learning program and a learning model 34 are stored in the memory. The learning model 34 is a model (i.e., a mathematical formula) of a multi-layer neural network. The multi-layer neural network is a so-called deep learning model, and is, for example, a convolutional type, a fully connected type, or the like. The learning model 34 includes learning parameters that are values of each weight in the intermediate layer of the neural network. The multi-layer neural network is a known technique, and detailed description thereof is omitted here. The learning model 34 is machine-learned using learning data that combines image data representing defects of the learning glass plate G1 and the types of defects.

[0041] The communication I / F 36 is an interface for the detection device 19 to communicate with the microscopic image generation device 20 via a wired or wireless communication network such as a LAN or Wi-Fi (registered trademark).

[0042] The control device 30 controls the wide-area imaging unit 40. The wide-area imaging unit 40 includes a line camera unit 42 in which a plurality of cameras and light sources are arranged side by side in the vertical direction. As shown in FIG. 3, a plurality of combinations of a bright-field wide-area imaging system 44 and a dark-field wide-area imaging system 46 are arranged side by side in the vertical direction in the line camera unit 42. FIG. 3 is a view of a set of one bright-field wide-area imaging system 44 and one dark-field wide-area imaging system 46 as seen from above. Note that the bright-field wide-area imaging system 44 is an example of a "bright-field wide-area imaging unit", and the dark-field wide-area imaging system 46 is an example of a "dark-field wide-area imaging unit".

[0043] The glass plate G is sent by the conveying device 22 along the X direction that is perpendicular to the vertical direction and parallel to the surface of the glass plate G. The bright-field wide-area imaging system 44 includes a camera 44a, a shielding plate 44b, and a light source 44c. The camera 44a captures the transmitted light L1 that is irradiated from the light source 44c and transmitted through the glass plate G. The shielding plate 44b shields a part (for example, half) of the transmitted light L1 to form a bright part and a dark part within the field of view of the camera 44a. Note that the transmitted light L1 is an example of a "first transmitted light".

[0044] The light source 44c is arranged on the surface Ga side of the glass plate G, and the camera 44a is arranged on the surface Gb side of the glass plate G. The optical axis of the light source 44c is the direction in which light is incident perpendicularly to the surface Ga of the glass plate G. The optical axis of the camera 44a is arranged on the straight line of the optical axis of the light source 44c. Thereby, the camera 44a is in a state of capturing the transmitted light L1 in bright field without the shielding plate 44b. The camera 44a captures the transmitted light L1 in semi-bright field with a part of the transmitted light L1 blocked by the shielding plate 44b.

[0045] The dark-field wide-area imaging system 46 includes a camera 46a and light sources 46b and 46c. Note that the light source 46c is arranged so as to overlap with the light source 44c in the vertical direction. The camera 46a captures the transmitted light L2 transmitted through the glass plate G from the light source 46c in bright field. Further, the camera 46a captures the transmitted light L3 transmitted through the glass plate G from the light source 46b in dark field. Note that the transmitted light L3 is an example of the "second transmitted light".

[0046] The optical axis of the camera 46a is arranged on the straight line of the optical axis of the light source 46c separated by a beam splitter 48 described later so that the transmitted light L2 can be captured by the camera 46a. The camera 46a captures the transmitted light L2 in bright field.

[0047] The light source 46b is arranged on the surface Ga side of the glass plate G. The optical axis of the light source 46b is arranged in a direction in which light is incident while being inclined with respect to the surface Ga of the glass plate G. The light source 46b is arranged on each of both sides of the light sources 44c and 46c. The optical axis of the camera 46a is arranged at a position deviated from the straight line of the optical axis of the light source 46b so that the transmitted light L3 does not enter the camera 46a. The camera 46a captures the transmitted light L3 in dark field. The transmitted light L3 is received by the camera 46a when scattering occurs in the glass plate G due to a defect or the like. Note that the inclination angle of the transmitted light L3 is exaggerated larger than the actual one for easy understanding, but usually, the transmitted light L3 is incident on the beam splitter 48.

[0048] In the camera 46a, the light in which the transmitted light L2 and the transmitted light L3 are combined is captured. That is, according to the camera 46a, an image in which bright field and dark field are combined is generated. The light sources 44c, 46b, and 46c are arranged adjacent to each other, and the positions where the transmitted lights L1, L2, and L3 pass through the glass plate G are common. The lighting timings of the light sources 44c, 46b, and 46c may be simultaneous or may be in order according to the imaging timing. The lighting timings of the light sources 44c, 46b, and 46c are controlled by the control unit 32.

[0049] The beam splitter 48 is disposed on the optical axes of the cameras 44a and 46a. The shielding plate 44b is disposed between the beam splitter 48 and the camera 44a. The beam splitter 48 transmits a specific wavelength and reflects other wavelengths. The beam splitter 48 separates the transmitted light that is irradiated from the light sources 44c, 46b, 46c and transmitted through the glass plate G into a component including the transmitted light L1 and a component including the transmitted lights L2 and L3. Specifically, by using the light sources 44c, 46b, 46c with different wavelengths, for example, LEDs of different colors, the beam splitter 48 separates the transmitted light L1 of the light source 44c from the transmitted light L2 of the light source 46b and the transmitted light L3 of the light source 46c. The component including the transmitted light L1 passes through the beam splitter 48 and is imaged by the camera 44a, and the component including the transmitted lights L2 and L3 is reflected by the beam splitter 48 and is imaged by the camera 46a. Note that the light sources 44c, 46b, 46c are not limited to LEDs, and may be, for example, metal halide lamps or laser light sources.

[0050] Although not shown, a plurality of the light sources 44c, 46b, 46c are arranged according to the plurality of cameras 44a, 46a that constitute the line camera unit 42. While the glass plate G is being conveyed by the conveying device 22, the entire effective surface of the glass plate G is imaged. The effective surface of the glass plate G is the surface that is used when the glass plate G is used as a product, i.e., a display device. For example, when the peripheral portion of the glass plate G is cut after inspection, the cut peripheral portion is not included in the effective surface, and the portion used for the product after cutting is included in the effective surface. Note that the surface to be imaged may include a surface other than the effective surface. Note that the image data of the image captured by the bright-field wide-area imaging system 44 is an example of the "bright-field wide-area image data P4", and the image data of the image captured by the dark-field wide-area imaging system 46 is an example of the "dark-field wide-area image data P5". Therefore, the "wide-area learning image data P1" includes the "bright-field wide-area image data P4" and the "dark-field wide-area image data P5".

[0051] (Configuration of Microscopic Image Generation Device) The microscopic image generation device 20 includes a control device 50 and a microscopic imaging unit 60. The control device 50 includes a control unit 52 and a communication I / F 56. The control unit 52 controls the microscopic image generation device 20. The control unit 52 includes a CPU and a memory constituted by a nonvolatile memory or the like. The CPU executes the processes described later according to a computer program stored in the memory.

[0052] The communication I / F 56 is an interface for the microscopic image generation device 20 to communicate with the detection device 19 via a wired or wireless communication network such as a LAN or Wi-Fi (registered trademark).

[0053] The control device 50 controls the microscopic imaging unit 60. The microscopic imaging unit 60 includes moving units 61 and 62 and a camera unit 63. As shown in FIG. 4, the camera unit 63 includes a microscope 66, a bright-field light source 67, and a dark-field light source 68. The microscope 66 has a higher magnification than cameras 44a and 46a. The bright-field light source 67 is used for the microscope 66 to image the glass plate G in bright field. The optical axis of the microscope 66 is set such that the microscope 66 images the reflected light L4 reflected by the defect D from the bright-field light source 67. That is, the microscope 66 and the bright-field light source 67 form a bright-field microscopic imaging system 64. The dark-field light source 68 is used for the microscope 66 to image the glass plate G in dark field. The optical axis of the microscope 66 is set such that the microscope 66 images the scattered light L5 irradiated on the defect D from the dark-field light source 68. That is, the microscope 66 and the dark-field light source 68 form a dark-field microscopic imaging system 65. In the present embodiment, when imaging a defect with the microscopic imaging unit 60, imaging is performed by the dark-field microscopic imaging system 65 after imaging by the bright-field microscopic imaging system 64, but imaging may also be performed by the bright-field microscopic imaging system 64 after imaging by the dark-field microscopic imaging system 65. Note that the bright-field microscopic imaging system 64 is an example of a "bright-field microscopic imaging unit", and the dark-field microscopic imaging system 65 is an example of a "dark-field microscopic imaging unit". Also, the image data of the image captured by the microscopic imaging unit 60 is an example of "microscopic learning image data P2", the image data of the image captured by the bright-field microscopic imaging system 64 is an example of "bright-field microscopic image data P6", and the image data of the image captured by the dark-field microscopic imaging system 65 is an example of "dark-field microscopic image data P7". Therefore, "microscopic learning image data P2" includes "bright-field microscopic image data P6" and "dark-field microscopic image data P7".

[0054] Specifically, for example, a digital microscope with a magnification of 5 times or more and 50 times or less can be used as the microscope 66. The magnification of the microscope 66 may be variable.

[0055] As shown in FIG. 2, the moving unit 61 includes a pair of guide rails and an actuator that move the camera unit 63 in a direction parallel to the X direction. The guide rails of the moving unit 61 extend parallel to the X direction outside the conveying device 22. The actuator moves the moving unit 62 along the guide rail. The moving unit 62 extends perpendicular to the pair of guide rails of the moving unit 61 and straddles the pair of guide rails of the moving unit 61. The moving unit 62 includes a guide rail and an actuator that is attached to the camera unit 63 and moves the camera unit 63 along the guide rail. The moving units 61 and 62 are controlled by the control device 50. The actuator of the moving unit 61 and the actuator of the moving unit 62 may include, for example, a servo motor and a ball screw, or may include a linear motor. Further, as an alternative to the ball screw, a timing belt, a chain, or the like may be used. The moving speed of the camera unit 63 by each of the moving units 61 and 62 is, for example, 500 mm / second or more and 2000 mm / second or less.

[0056] An acquisition terminal 70 is communicably connected to the inspection device 18. The acquisition terminal 70 is a desktop or laptop computer, a tablet terminal, a mobile terminal, or the like. The acquisition terminal 70 includes a control unit 72, an operation unit 74, a display unit 75, and a communication I / F 76. The control unit 72 includes a CPU and a memory constituted by a non-volatile memory or the like. The CPU executes the processes described below according to the computer programs stored in the memory. The control unit 72 is communicably connected to the operation unit 74, the display unit 75, and the communication I / F 76 by wiring (not shown). The operation unit 74 receives operations performed by the user. The operation unit 74 includes a keyboard and a mouse. The display unit 75 displays an image so as to be visible to the user. The display unit 75 includes a display device such as a liquid crystal display or an organic EL display. Note that the operation unit 74 and the display unit 75 may be integrally configured, for example, as a touch panel. The communication I / F 76 is configured in the same manner as the communication I / Fs 36 and 56 and is used for the acquisition terminal 70 to communicate with each of the detection device 19 and the microscopic image generation device 20. Note that the acquisition terminal 70 is an example of an “acquisition unit”.

[0057] (Method for manufacturing a glass plate) With reference to FIG. 5, a method for manufacturing a glass plate G will be described. First, in S12, a forming process is executed in the forming device 12. As a result, a glass ribbon is formed from molten glass. Next, in S14, a cooling process is executed in the cooling device 14. As a result, the glass ribbon is gradually cooled while moving in the lehr, and further cooled while moving in the cooling chamber. In S16, a cutting process is executed in the cutting device 16. As a result, a glass plate G is produced from the glass ribbon.

[0058] Next, in S18, an inspection process is executed in the inspection device 18. When the inspection process is completed, in S20, an unloading process is executed in the transfer device 22. In the unloading process, the glass plate G is unloaded from the manufacturing unit 10 by the transfer device 22, placed on a pallet or the like, and the manufacturing method is completed.

[0059] Note that the glass plate G manufactured in this way may have a plurality of display devices cut out from one glass plate G for the purpose of improving productivity. Therefore, the dimensions of the glass plate G in this embodiment are preferably 1000 mm × 1000 mm or more and 4000 mm × 4000 mm or less. The lower limit value of the dimensions of the glass plate G is more preferably 1500 mm × 1500 mm or more, and even more preferably 2000 mm × 2000 mm or more. The upper limit value of the dimensions of the glass plate G is more preferably 3500 mm × 3500 mm or less, and even more preferably 3000 mm × 3000 mm or less.

[0060] (Inspection process) Referring to FIG. 6, the inspection process of S18 will be described. In the inspection process, at S32, the glass plate G is conveyed from the cutting device 16 to the detection device 19 by the conveying device 22. Hereinafter, the detection device 19 and the microscopic image generation device 20 will be mainly described. However, the processes executed by the detection device 19 and the microscopic image generation device 20 are the processes executed by the control unit 32 of the detection device 19 for each part of the detection device 19 and the processes executed by the control unit 52 of the microscopic image generation device 20 for each part of the microscopic image generation device 20.

[0061] At S34, the detection device 19 images the glass plate G being conveyed by the conveying device 22 using the line camera unit 42 (wide-area learning image acquisition step). The line camera unit 42 images the glass plate G in bright field by the bright-field wide-area imaging system 44 and images the glass plate G in a state where the bright field and the dark field are combined by the dark-field wide-area imaging system 46. As shown in FIG. 7, for example, when the defect is a bubble, the defect images representing the defect are differently represented by the bright-field wide-area image data P4 and the dark-field wide-area image data P5. At S35, the detection device 19 determines whether the entire surface of the glass plate G has been imaged. Specifically, the glass plate G is continuously conveyed by the conveying device 22. The line camera unit 42 images the entire glass plate G by imaging a part of the glass plate G facing the line camera unit 42 a plurality of times. The detection device 19 determines that the entire surface of the glass plate G has been imaged when imaging of a predetermined number of times has been executed for one glass plate G (YES at S35) and proceeds to S36. On the other hand, the detection device 19 determines that the entire surface of the glass plate G has not been imaged when imaging of a predetermined number of times has not been executed for one glass plate G (NO at S35) and returns to S34. Thereby, the entire surface of the glass plate G is imaged. Note that at S35, it may be determined whether the entire surface of the glass plate G has been imaged by detecting the position of the glass plate G with a sensor (not shown).

[0062] The detection device 19 stores the bright-field wide-area image data P4 and the dark-field wide-area image data P5 in the memory of the control unit 32. The detection device 19 further stores, in combination with the bright-field wide-area image data P4 and the dark-field wide-area image data P5, position information representing the position of the image represented by the bright-field wide-area image data P4 and the dark-field wide-area image data P5 on the glass plate G. The position information is specified by the positions of the clamping mechanisms 24 and 25 in the X direction and the positions of the cameras 44a and 46a in the vertical direction.

[0063] In S36, the detection device 19 uses the bright-field wide-area image data P4 and the dark-field wide-area image data P5 to determine whether a defect is detected in the images represented by the bright-field wide-area image data P4 and the dark-field wide-area image data P5. As shown in FIG. 7, if a defect exists, in the images represented by the bright-field wide-area image data P4 and the dark-field wide-area image data P5, the portion representing the defect is shown darker compared to other portions. Thus, when the image contains a portion that is shown dark, i.e., has a low luminance, it is determined that a defect has been detected. If it is determined that no defect is detected (NO in S36), the inspection process ends. On the other hand, if it is determined that a defect is detected (YES in S36), the process proceeds to S38. Note that among the bright-field wide-area image data P4 and the dark-field wide-area image data P5, the one containing the defect is an example of "defect image data".

[0064] In S38, the detection device 19 specifies a defect position representing the position of the detected defect (learning defect position acquisition step). Specifically, using the information stored in S34, the position of the region represented by the image on the glass plate G is specified. Next, the position of the defect within the image is specified. Thereby, the position of the defect on the glass plate G is specified. Next, in S40, the detection device 19 specifies the type of the defect (learning defect type specification step and inspection target defect type specification step). Specifically, according to the learning model 34, using the bright-field wide-area image data P4 and the dark-field wide-area image data P5, the type of the defect detected in S36 is specified.

[0065] In addition, in S40, the detection device 19 may perform image processing using the bright-field wide-area image data P4 and the dark-field wide-area image data P5 to generate one or more processed image data so that defects can be easily discriminated. For example, filter processing such as edge enhancement and noise removal may be performed.

[0066] Next, in S42, as shown in FIG. 8, the detection device 19 causes the control unit 32 to store a combination of the defect position specified in S38, the type of the defect specified in S40, and the bright-field wide-area image data P4 and the dark-field wide-area image data P5 acquired in S34. When one or more processed image data are generated by image processing in S40, a combination further including the processed image data is caused to be stored in the control unit 32.

[0067] Next, in S44, the detection device 19 transmits the combination stored in S42 to the microscopic image generation device 20 and the acquisition terminal 70. When a plurality of defects are detected on one glass plate G, the detection device 19 transmits the stored combination to the microscopic image generation device 20 and the acquisition terminal 70 for each of the plurality of defects. When the control unit 72 of the acquisition terminal 70 receives the combination, it can cause the display unit 75 to display an image represented by the bright-field wide-area image data and the dark-field wide-area image data, the defect position, and the type of the defect. Thereby, the user can grasp the position and type of the defects existing on the glass plate G.

[0068] In S46, the microscopic image generation device 20 specifies defects to be imaged by the microscopic image generation device 20 from among the combinations received in S44. Specifically, when the number of received combinations is equal to or less than a predetermined number (for example, 3), the microscopic image generation device 20 specifies that all defects are defects to be imaged. On the other hand, when the number of received combinations is larger than the predetermined number, the microscopic image generation device 20 specifies that a predetermined number of defects having smaller dimensions among the received combinations are defects to be imaged.

[0069] When the entire glass plate G is imaged by the detection device 19, the glass plate G is conveyed from the detection device 19 to the microscopic image generation device 20 by the conveying device 22. In S48, the microscopic image generation device 20 drives the moving parts 61 and 62 to move the camera part 63 to the position of the defect identified in S46. Next, in S50, the microscopic image generation device 20 causes the camera part 63 to image the defect (microscopic learning image generation step). In S50, the microscopic image generation device 20 images the defects of the number identified in S46.

[0070] In the microscopic image generation device 20, an enlarged defect is imaged using the microscope 66. As shown in FIG. 9, in the microscopic image generation device 20, an image of the bright-field defect D is imaged in the bright-field microscopic imaging system, and an image of the dark-field defect D is imaged in the dark-field microscopic imaging system. Hereinafter, the image captured by the microscopic image generation device 20 is referred to as a microscopic image, and the image data representing the microscopic image is referred to as microscopic image data. When the defect D is a foreign object, it appears as a black image in the bright-field wide-area image data P4 and as a white image in the dark-field wide-area image data P5, but there may be cases where the image itself cannot be obtained. When the defect D is a bubble, it has a shape close to an ellipse in both the bright-field wide-area image data P4 and the dark-field wide-area image data P5. When the defect D is an attachment such as dust, it has, for example, a wavy black shape in the bright-field wide-area image data P4 and a wavy white shape in the dark-field wide-area image data P5.

[0071] Next, in S52, as shown in FIG. 10, the microscopic image generation device 20 adds the image data representing the image captured in S50 to the combination identified in S46 and transmits it to the acquisition terminal 70 to end the process. That is, in S52, the wide-area learning image data P1 (bright-field wide-area image data P4 and dark-field wide-area image data P5), the microscopic learning image data P2 (bright-field microscopic image data P6 and dark-field microscopic image data P7), the defect position, and the type of defect are transmitted to the acquisition terminal 70.

[0072] The user specifies the type of defect using the combination received at the acquisition terminal 70, namely, the wide-area learning image data P1, the microscopic learning image data P2, the defect position, and the type of defect, among which the wide-area learning image data P1 and the microscopic learning image data P2 are used. According to this configuration, the user can specify the type of defect using the enlarged microscopic learning image data P2 for small defects. Thereby, the type of defect can be accurately specified for small defects.

[0073] As shown in FIG. 11, the user stores the specified type of defect and the wide-area learning image data P1 in the acquisition terminal 70 as learning data. As a result, as shown in FIG. 12, a learning data library including a plurality of learning data in which a plurality of image data and the type of defect are combined for a plurality of defects is stored (storage step).

[0074] When the learning data library is generated, the user transmits the generated learning data library to the detection device 19. When the detection device 19 receives the learning data library, it updates the learning model using the learning data included in the learning data library. Thereby, the accuracy of specifying the type of defect by the learning model can be improved. Further, since the microscopic learning image data P2 of the defect is generated by the manufacturing unit 10, it is not necessary to separately execute the process for generating the microscopic learning image data P2.

[0075] According to the inspection method executed in the inspection process of this embodiment, the microscopic learning image data P2 of the defect of the learning object detected by the detection device 19 can be captured by the microscopic image generation device 20. Thereby, even for a defect with a small size, the user can easily identify the type of the defect by checking the microscopic learning image data P2. By generating a learning model using the combination of the type of defect and the image data as learning data, even when the size of the defect of the inspection object is small, the type of the defect can be identified using the learned learning model 34. Note that the image data of the image captured by the microscopic image generation device 20 for the learning glass plate G1 (learning object) is "wide-area learning image data P1", and the image data of the image captured by the microscopic image generation device 20 for the inspection target glass plate G2 (inspection object) is "wide-area inspection image data P3". Further, the process of identifying the type of defect included in the learning glass plate G1 (learning object) is the "learning defect type identification process", and the process of identifying the type of defect included in the inspection target glass plate G2 (inspection object) is the "inspection target defect type identification process".

[0076] Multiple types of defects such as small irregularities, air bubbles, and foreign substances on the surface occur in the glass plate G. Depending on the type and state of the defect, there are defects that are more clearly imaged in bright field and defects that are more clearly imaged in dark field. In the detection device 19, by using the image data captured in bright field and the image data in which dark field and bright field are combined, the type of defect can be easily identified in the learning model. In the microscopic image generation device 20, by using the bright-field microscopic image data P6 captured in bright field and the dark-field microscopic image data P7 captured in dark field, the user can easily identify the type of defect.

[0077] The microscopic image generation device 20 selects defects to be imaged from among a plurality of defects included in the glass plate G. As a result, it is not necessary to use all the defects included in the glass plate G as learning data. Thereby, the time required to generate microscopic learning image data can be shortened, and the burden of generating learning data and the learning period in the learning model 34 can be suppressed. In addition, since defects with relatively small dimensions are selected, image data representing defects with relatively small dimensions can be adopted as learning data. For defects with relatively large dimensions, the type of defect can be identified relatively easily from the images captured by the detection device 19. On the other hand, for defects with relatively small dimensions, it is difficult to identify the type of defect from the images captured by the detection device 19. In the present embodiment, by collecting learning data for defects with small dimensions, the accuracy of identifying the type of defect using the learning model 34 can be improved.

[0078] As the dimensions of the pixels displayed on the display device become smaller, the allowable dimensions of the defects in the glass plate G used for the display device become smaller. By improving the accuracy of identifying the type of defect in the learning model 34, the type of defect can be appropriately identified for the glass plate G used for the display device using the inspection device 18.

[0079] In addition, in the present embodiment, the dimensions of the glass plate G are, for example, 1000 mm × 1000 mm or more. In such a large glass plate G, if all the defects included in the glass plate G are imaged by the microscopic image generation device 20, the time required for inspection will increase significantly, deteriorating the production efficiency of the glass plate G. By selecting defects with relatively small dimensions and imaging them with the microscopic image generation device 20, the accuracy of identifying the type of defect using the learning model 34 can be improved without deteriorating the production efficiency of the glass plate G.

[0080] In another embodiment, the method for manufacturing a glass plate is different from the above-described embodiment. Specifically, as shown in FIG. 13, the steps from S12 to S18 are executed in the same manner as in the above-described embodiment. In the inspection step of S18, in addition to the types of defects included in the glass plate G, the number and dimensions of the defects included in the glass plate G are specified.

[0081] Specifically, in S40 of the inspection step, the detection device 19 further specifies the number of defects detected in S36 and the dimensions of each defect. For example, the detection device 19 may count the number of defects included in the bright-field wide-area image data P4 and the dark-field wide-area image data P5. Also, for example, the detection device 19 may count the number of times the type of defect is specified as the number of defects. The detection device 19 may specify the dimensions of the defects by measuring the dimensions of the defects included in the bright-field wide-area image data P4 and the dark-field wide-area image data P5. Further, the detection device 19 may classify the defects into a plurality of classes (for example, 10 μm or less, 10 μm to 50 μm, 50 μm or more) according to the dimensions of the defects included in the bright-field wide-area image data P4 and the dark-field wide-area image data P5. The detection device 19 stores the specified types, numbers, and dimensions of the defects.

[0082] Next, when the inspection step of S18 is completed, in S19, a quality determination step is executed. In the quality determination step, the quality of the glass plate G is determined. More specifically, in the quality determination step, the quality of the glass plate G is determined based on the defect information including the types, number, and dimensions of the defects specified in the inspection step of S18. The glass plate G is classified into a plurality of classes according to quality. The plurality of classes are three or more classes including one class of poor quality and two or more classes of good quality.

[0083] For example, in the control unit that controls the transfer device 22, defective product thresholds are stored in advance for each of the number of defects and the size of the defects. When at least one of the number of defects and the size of the defects identified in the inspection step of S18 exceeds the stored defective product threshold, the control unit of the transfer device 22 determines that the glass plate G is of poor quality. Further, when the glass plate G contains a specific type of defect, the control unit of the transfer device 22 determines that the glass plate G is of poor quality. When both the number of defects and the size of the defects identified in the inspection step of S18 do not exceed the stored defective product threshold and the glass plate G does not contain a specific type of defect, the control unit of the transfer device 22 determines that the glass plate G is of good quality.

[0084] In the control unit that controls the transfer device 22, class thresholds are further stored in advance for each of the number of defects and the size of the defects. The control unit of the transfer device 22 distinguishes classes according to whether both the number of defects and the size of the defects identified in the inspection step of S18 do not exceed the stored class threshold and whether at least one of the number of defects and the size of the defects identified exceeds the stored class threshold. Note that the class may be distinguished according to the type of defect. Note that the quality determination step may be executed by other configurations of the manufacturing unit 10, for example. Note that the quality determination step may determine the quality using at least one of the number of defects, the size of the defects, and the type of defect.

[0085] When the quality determination step is completed, the unloading step of S20 is executed. In the unloading step, the glass plate G is unloaded from the manufacturing unit 10 by the transfer device 22 and sorted into pallets for each class according to the quality class determined in the quality determination step of S19. Further, in S21, a defective product sorting step is executed in parallel with the unloading step of S20. In the defective product sorting step, the glass plate G determined to be of poor quality in the quality determination step of S19 is transported to a pallet for disposal or the like by the transfer device 22. As a result, the glass plate G determined to be of poor quality in the quality determination step of S19 is discarded. The manufacturing unit 10 may further execute a disposal step of disposing of the glass plate G transported to a pallet for disposal or the like.

[0086] In this embodiment, by sorting the glass plates G according to quality, it is not necessary to check the quality when the glass plates G are shipped. It is possible to ship the glass plates G of a quality that meets the quality required for the display device to the destination.

[0087] By executing the defective product sorting process, it is possible to prevent the plurality of glass plates G carried out from including defective products.

[0088] As described above, specific examples of the technology disclosed in this specification have been described, but these are merely examples and do not limit the scope of the claims. The technology described in the claims includes various modifications and changes of the specific examples illustrated above. For example, the following modification examples may be adopted.

[0089] (Modification Example 1) The bright-field wide-area imaging system 44 generates image data by imaging in the bright field. However, the bright-field wide-area imaging system 44 may generate image data by synthesizing the bright field and the dark field. The dark-field wide-area imaging system 46 may generate image data by imaging in the dark field. The same applies to the microscopic image generation device 20.

[0090] (Modification Example 2) The acquisition terminal 70 may be included in the inspection device 18. For example, the acquisition terminal 70 may be integrally configured with the microscopic image generation device 20.

[0091] (Modification Example 3) The inspection device 18 may not include an imaging system. For example, there may be another imaging device independent of the inspection device 18, and the inspection device 18 may acquire imaging image data from the other imaging device via wireless communication such as a wireless LAN.

[0092] (Modification Example 4) Another detection device for identifying the type of defect may be provided using the microscopic learning image data P2. That is, without the user identifying the type of defect using the microscopic learning image data P2, another detection device having a predetermined learning model may identify the type of defect according to the predetermined learning model.

[0093] (Modification Example 5) The transport device 22 may not be provided with a clamping mechanism 25 that clamps the lower edge of the glass plate G. That is, the glass plate G may be transported in a state where only the upper edge is clamped by the clamping mechanism 24.

[0094] (Modification Example 6) The microscopic imaging unit 60 includes a microscope 66, a bright-field light source 67, and a dark-field light source 68 on the surface Gb side of the glass plate G, and images the reflected light L4 and the scattered light L5 of the defect D. However, a microscope 66 may be provided on the surface Gb side of the glass plate G, and a bright-field light source 67 and a dark-field light source 68 may be provided on the surface Ga side, and the scattered light and the transmitted light of the defect D may be imaged. In this case, in addition to the moving parts 61 and 62 for moving the microscope 66, a moving part for moving the bright-field light source 67 and the dark-field light source 68 is further provided, and the microscope 66, the bright-field light source 67, and the dark-field light source 68 are moved synchronously.

[0095] (Modification Example 7) In the quality determination step S19, based on the defect information including the position of the defect in addition to the type of defect, the number of defects, and the dimensions of the defects identified in the inspection step of S19, the quality of the glass plate G may be determined. For example, as a subsequent step of the quality determination step S19, in the case of including a step of cutting out a second glass plate with a smaller dimension from the glass plate G according to a predetermined cutting pattern, and at least one of the number of defects and the dimensions of the defects included in the glass plate G is larger than the defective product threshold value, but there exists a cutting pattern such that the number of defects and the dimensions of the defects included in the second glass plate are below the defective product threshold value, the glass plate G may be determined to have good quality.

[0096] The technical elements described in this specification or the drawings exhibit technical utility either individually or in various combinations, and are not limited to the combinations described in the claims at the time of filing. Further, the technology exemplified in this specification or the drawings achieves multiple objectives simultaneously, and achieving one of those objectives by itself has technical utility.

Description of Reference Numerals

[0097] 10: Manufacturing unit, 12: Molding device, 14: Cooling device, 16: Cutting device, 18: Inspection device, 19: Detection device, 20: Microscopic image generation device, 22: Conveying device, 30: Control device, 40: Wide-area imaging unit, 44: Bright-field wide-area imaging system, 46: Dark-field wide-area imaging system, 50: Control device, 60: Microscopic imaging unit, 61: Moving part, 62: Moving part, 63: Camera part, 64: Bright-field microscopic imaging system, 65: Dark-field microscopic imaging system, 70: Acquisition terminal, G: Glass plate, G1: Glass plate for learning, G2: Glass plate to be inspected, L1: Transmitted light, L2: Transmitted light, L3: Transmitted light, L4: Reflected light, L5: Scattered light, P1: Wide-area learning image data, P2: Microscopic learning image data, P3: Wide-area inspection image data, P4: Bright-field wide-area image data, P5: Dark-field wide-area image data, P6: Bright-field microscopic image data, P7: Dark-field microscopic image data

Claims

1. A specifying unit that specifies a defect position in the learning object of a predetermined defect using wide-area learning image data representing the learning object; A microscopic image generation unit that generates microscopic learning image data representing the predetermined defect by microscopically imaging the specified defect position; An acquisition unit that acquires a combination of type information representing the type of the predetermined defect obtained using the generated microscopic learning image data and defect image data representing the predetermined defect included in the wide-area learning image data; and The inspection apparatus, wherein the specifying unit specifies the type of a defect in an inspection object using wide-area inspection image data representing the inspection object by using a learning model generated by machine learning using the acquired combination of the type information and the defect image data as learning data.

2. The learning object includes a glass plate, The inspection apparatus further includes an inspection image generation unit that generates the wide-area learning image data, The inspection image generation unit includes: A bright-field wide-area imaging unit that generates bright-field wide-area image data representing the glass plate by imaging the first transmitted light transmitted through the glass plate in a field of view including a bright field; A dark-field wide-area imaging unit that generates dark-field wide-area image data representing the glass plate by imaging the second transmitted light transmitted through the glass plate at an angle different from the first transmitted light in a field of view including a dark field; and The wide-area learning image data includes the bright-field wide-area image data and the dark-field wide-area image data. The inspection apparatus according to claim 1.

3. The learning object includes a glass plate, The microscopic image generation unit includes: A bright-field microscopic imaging unit that generates bright-field microscopic image data representing the glass plate by imaging the reflected light reflected by the glass plate in a field of view including a bright field; A dark-field microscopic imaging unit that generates dark-field microscopic image data representing the glass plate by imaging the scattered light scattered by the defect in a field of view including a dark field; and The microscopic learning image data includes the bright-field microscopic image data and the dark-field microscopic image data. The inspection apparatus according to claim 1 or 2.

4. The microscopic image generation unit includes: A camera unit that microscopically images the predetermined defect; A moving unit that moves the camera unit to the defect position. When the object to be learned includes a plurality of defects including one or more of the predetermined defects, the inspection apparatus according to claim 1 or 2, which generates the microscopic learning image data of the predetermined defects in a number less than the number of the plurality of defects.

5. The microscopic image generation unit generates the microscopic learning image data of the predetermined defects in ascending order of size among the predetermined defects, the inspection apparatus according to claim 4.

6. The object to be learned includes a glass plate used for a display device, the inspection apparatus according to claim 1 or 2.

7. The wide-area inspection image data includes the entire effective surface of the glass plate, the inspection apparatus according to claim 6.

8. A wide-area learning image acquisition step of acquiring wide-area learning image data representing an object; A learning defect position acquisition step of acquiring a defect position in the object of a predetermined defect of the object; A microscopic learning image generation step of generating microscopic learning image data representing the predetermined defect by microscopically imaging the acquired defect position; A storage step of storing a combination of type information representing the type of the predetermined defect obtained using the generated microscopic learning image data and defect image data representing the predetermined defect included in the wide-area learning image data, a learning data generation method.

9. A wide-area learning image acquisition step of acquiring wide-area learning image data representing an object; A learning defect position acquisition step of acquiring a defect position in the object of a predetermined defect of the object; A microscopic learning image generation step of generating microscopic learning image data representing the predetermined defect by microscopically imaging the acquired defect position; A storage step of storing a combination of type information representing the type of the predetermined defect obtained using the generated microscopic learning image data and defect image data representing the predetermined defect included in the wide-area learning image data; A learning model generation step of generating a learning model by machine learning using the stored combination as learning data, a learning model generation method.

10. A learning defect type identification step of identifying the type of a predetermined defect in the object to be learned using wide-area learning image data representing the object to be learned; A learning defect position acquisition step of acquiring a defect position in the object to be learned of the predetermined defect using the wide-area learning image data; A microscopic learning image generation step of generating microscopic learning image data representing the predetermined defect by microscopically imaging the acquired defect position; A storage step of storing a combination of type information representing the type of the predetermined defect obtained using the generated microscopic learning image data and defect image data representing the predetermined defect included in the wide-area learning image data; An inspection target defect type specifying step of specifying the type of a defect in the inspection target object using wide-area inspection image data representing the inspection target object, by using a learning model generated by machine learning using the combination of the stored type information and the defect image data as learning data;

11. The method for manufacturing an object according to claim 10, further comprising a quality determination step of determining the quality of the object based on defect information including at least one of the number of the defects included in the object, the dimensions of the defects, and the type of the defects specified in the inspection target defect type specifying step.

12. The object includes a glass plate used for a display device, The method for manufacturing an object according to claim 11, further comprising a carrying-out step of sorting the object according to the quality based on the quality determined in the quality determination step.

13. The method for manufacturing an object according to claim 11 or 12, further comprising a defective product sorting step of distinguishing the object determined to be of poor quality in the quality determination step from the object not determined to be of poor quality.

Images