Classification apparatus, classification method, program, and trained model

Abstract

To provide a classification apparatus that allows classification of a suitable polishing time for dross in a glass plate manufactured by a float method.SOLUTION: A classification apparatus is provided with a determination section that inputs image data, in which a glass plate manufactured by a float method is imaged, into a trained model of machine learning, and acquires an output result from the trained model as a determination result concerning the time necessary for polishing dross of the glass plate.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to a classification device, a classification method, a program, and a trained model. [Background technology]

[0002] Glass sheets are manufactured by the float process, in which molten glass continuously supplied onto molten tin in a bath is caused to flow over the molten tin and formed into a strip shape. In such a manufacturing method, dross is generated on the surface of the glass plate (the surface in contact with the molten tin, that is, the so-called lower surface), and a step of polishing off the dross is provided.

[0003] Patent Document 1 describes a method for removing dross, particularly mold-retained dross, present on the surface of a glass substrate produced by the float process by polishing (see Patent Document 1). Patent Document 1 describes, for example, that dross is a convex adhesion defect found scattered in dots when the glass is visually observed under a fluorescent lamp, and that the dross generated by the float process is mostly tin oxide-based dross whose main component is tin oxide (SnO2), an oxide of metal tin, which is a metallic component of the molten metal bath (see paragraph 0003 of Patent Document 1). [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Laid-Open No. 2016-28988 Summary of the Invention [Problem to be solved by the invention]

[0005] However, in the conventional techniques, it has not been possible to sufficiently use the optimum polishing time for each state of dross. For example, the optimal polishing time is determined depending on the state of the dross (e.g., the state of tin), and there may be dross that can be polished for a short time and dross that requires a long polishing time.However, conventional technology has not been able to adequately classify these polishing times, which has resulted in poor takt time and reduced productivity of glass plates.

[0006] The present disclosure has been made in consideration of these circumstances, and aims to provide a classification device, classification method, program, and trained model that can classify appropriate polishing times for dross on glass sheets manufactured by the float method. [Means for solving the problem]

[0007] One aspect of the present disclosure is a classification device that includes a determination unit that inputs image data of a glass plate manufactured by the float method into a machine learning trained model and obtains an output result from the trained model as a determination result regarding the time required to polish the dross of the glass plate.

[0008] One aspect of the present disclosure is a classification method in which a classification device inputs image data of a glass plate manufactured by the float method into a trained machine learning model, and obtains an output result from the trained model as a judgment result regarding the time required to polish the dross of the glass plate.

[0009] One aspect of the present disclosure is a program for enabling a computer to input image data of a glass plate manufactured by the float method into a trained machine learning model, and to obtain the output result from the trained model as a determination result regarding the time required to polish the dross of the glass plate.

[0010] One aspect of the present disclosure is a machine learning trained model that inputs image data of a glass plate manufactured by the float method and outputs judgment result data regarding the time required to polish the dross of the glass plate. [Effects of the Invention]

[0011] The information processing device, classification device, classification method, program, and trained model according to the present disclosure can classify the appropriate polishing time for dross on glass sheets manufactured by the float process. [Brief explanation of the drawings]

[0012] [Figure 1] 1 is a diagram illustrating a schematic configuration example of a glass manufacturing apparatus according to an embodiment. [Figure 2] FIG. 1 is a cross-sectional view showing an example of the configuration of a float bath according to an embodiment. [Figure 3] FIG. 2 is a diagram illustrating an example of the arrangement of an imaging unit according to the embodiment. [Figure 4] 1 is a diagram showing an example of a procedure for manufacturing glass according to an embodiment. FIG. [Figure 5] FIG. 1 is a diagram illustrating an example of the configuration of a classification device according to an embodiment. [Figure 6A] FIG. 10 is a diagram showing an example of an image including type A dross according to the embodiment. [Figure 6B] FIG. 10 is a diagram showing an example of an image including type B dross according to the embodiment. [Figure 6C] FIG. 10 is a diagram showing an example of an image including type C dross according to the embodiment. [Figure 7] 5A, 5B, and 5C are diagrams showing an example of the change over time in the distribution of dross in the width direction of a glass sheet according to an embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0013] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings.

[0014] [Glass manufacturing equipment] FIG. 1 is a diagram showing a schematic configuration example of a glass manufacturing apparatus 1 according to an embodiment. For convenience of explanation, FIG. 1 shows an XYZ Cartesian coordinate system, which is a three-dimensional Cartesian coordinate system. The glass manufacturing apparatus 1 includes a melting furnace 11 , a float bath 12 , an annealing furnace 13 , an inspection and cutting section 14 , and a polishing section 15 .

[0015] In the melting furnace 11, the melting process is carried out. In the melting process, glass raw materials prepared by mixing multiple types of raw materials are melted to obtain molten glass. The glass raw materials are charged into the melting furnace 11 and then melted by heating means such as radiant heat from a flame sprayed from a burner or electric melting to become molten glass.

[0016] In the float bath 12, the molding process is carried out. In the forming process, the molten glass obtained in the melting process is continuously supplied onto molten tin in a bath, and the molten glass is caused to flow and form on the molten tin to obtain a glass sheet (so-called glass ribbon), which is a sheet of glass. The glass sheet is cooled while flowing in a predetermined direction and is pulled up from the molten tin.

[0017] In the annealing furnace 13, a annealing process is carried out. In the annealing step, the glass sheet obtained in the forming step is annealed inside the annealing furnace 13. Inside the annealing furnace 13, the glass sheet is annealed while being transported horizontally on rolls from the entrance to the exit of the annealing furnace 13.

[0018] In the inspection and cutting section 14, the inspection process and the cutting process are carried out. In the inspection step, the glass sheet that has been slowly cooled in the slow cooling step is inspected. In the cutting process, the glass plate that has been annealed in the annealing process is cut into predetermined dimensions using a cutting machine. In the cutting process, both widthwise edge portions (so-called edge portions) of the glass plate are cut off. Both widthwise edge portions of the glass plate are cut off because they become thick due to the influence of surface tension, etc.

[0019] In the polishing section 15, a polishing process is carried out. In the polishing step, the surface of the glass plate after cutting is polished. After the polishing process, the glass plates are shipped.

[0020] In this embodiment, the polishing step involves polishing and removing dross present on the surface of the glass sheet. Note that the dross removal may be performed to an extent that does not impair the glass sheet as a finished product. Here, the time suitable for polishing the dross (polishing time) may vary depending on the state of the dross. That is, the polishing time for the dross needs to be long enough to sufficiently remove the dross, but from the viewpoint of the production efficiency of the glass plate, it is better to have a shorter time.

[0021] In this embodiment, the polishing unit 15 can change the polishing time for each unit area (polishing unit area) to be polished on the glass plate. The polishing unit area is not particularly limited, and may be, for example, the entire area of ​​each cut glass plate, or a predetermined divided area set on the glass plate (an area in which the entire area is divided into multiple areas according to a predetermined rule), or an area set for each portion of dross present on the glass plate.

[0022] [Float Bus] FIG. 2 is a cross-sectional view showing an example of the configuration of the float bath 12 according to the embodiment. FIG. 2 shows an XYZ Cartesian coordinate system similar to that shown in FIG. The cross section shown in FIG. 2 is a side cross section of the float bath 12 as seen from the negative side to the positive side of the X axis. In this embodiment, for convenience of explanation, the upper surface (the surface on the positive side of the Z axis) of the molten glass G (glass sheet) in the float bath 12 will be described as the front surface, and the lower surface (the surface on the negative side of the Z axis) will be described as the back surface, but these surfaces may be called by other names.

[0023] The float bath 12 forms molten glass G, which is continuously supplied onto molten tin M in a bath 22, by causing it to flow on the molten tin M. After being supplied onto the molten tin M near the inlet 21 of the float bath 12, the molten glass G is cooled while flowing in a predetermined direction and is pulled up from the molten tin M near the outlet 23 of the float bath 12. The float bath 12 is composed of a bath 22 that contains molten tin M, a sidewall 24 that is installed along the upper outer periphery of the bath 22, and a ceiling 26 that is connected to the sidewall 24 and covers the upper part of the bath 22. A gas supply path 30 is provided in the ceiling 26 to supply a reducing gas to a space 28 formed between the bath 22 and the ceiling 26. A heater 32 serving as a heat source is inserted into the gas supply path 30.

[0024] The gas supply passage 30 supplies a reducing gas to the space 28 in the float bath 12 to prevent oxidation of the molten tin M. The reducing gas contains, for example, 1 to 15 volume % hydrogen gas and 85 to 99 volume % nitrogen gas. The space 28 in the float bath 12 is set to a pressure higher than atmospheric pressure to prevent air from entering through gaps between the bricks that make up the side wall 24. A plurality of heaters 32 are provided at intervals in the flow direction (Y direction) and width direction (X direction) of the molten glass G to adjust the temperature distribution in the float bath 12. The output of the heaters 32 is controlled so that the temperature of the molten glass G decreases from the inlet 21 toward the outlet 23 of the float bath 12. The output of the heaters 32 is controlled so that the thickness of the molten glass G becomes uniform in the width direction (X direction).

[0025] The bath 22 is composed of a box-shaped metal case 34 that is open at the top, and bottom bricks 36 and side bricks 38 that are installed inside the metal case 34. The metal case 34 prevents air from entering the bath 22 from the sides or below. The bottom bricks 36 are arranged two-dimensionally at small intervals that prevent them from coming into contact with each other due to thermal expansion. The bottom bricks 36 are surrounded by side bricks 38 that are arranged in a ring shape. Also shown in FIG. 2 is the top surface 36a of the bottom brick 36.

[0026] <Example of imaging unit placement> FIG. 3 is a diagram illustrating an example of the arrangement of the imaging unit according to the embodiment. For convenience of explanation, FIG. 3 shows XY coordinates in an XYZ orthogonal coordinate system similar to that shown in FIG. Fig. 3 schematically shows the state of the molten glass G when viewed from above (the positive side of the Z axis) of the float bath 12. Specifically, Fig. 3 shows a glass width region E that represents the width region of the molten glass G in the float bath 12. The glass width region E in the flow direction represents the change in glass width in the flow direction.

[0027] Tensile forces F1 to F4 directed outward in the width direction (to both outer sides) are applied to the molten glass G. As a result, the width of the glass width region E gradually increases from the upstream side to the downstream side of the flow.

[0028] FIG. 3 shows a schematic diagram of a defect inspection section 311 (the inspection section of the inspection and cutting section 14 in FIG. 1). In the example of FIG. 3, the inspection unit 311 has three imaging units C1 to C3 arranged in the width direction. The imaging unit C1 captures an image of the inspection region R1. The imaging unit C2 captures an image of the inspection region R2. The imaging unit C3 captures an image of the inspection region R3. Here, the inspection region R1, the inspection region R2, and the inspection region R3 are aligned in the width direction corresponding to the imaging units C1 to C3, respectively. 3, a defect H12 exists in inspection area R1, and a defect H11 exists in inspection area R3. Here, the defect H11 corresponds to the defect H1 and the defect H1a, and the defect H12 corresponds to the defect H2 and the defect H2a. In the example of FIG. 3, there is no defect in the inspection area R2. In this way, each of the imaging units C1 to C3 captures an image of a partial area of ​​the glass plate. If the area contains a defect, the image contains a portion where the defect is captured (defective portion).

[0029] Here, in this embodiment, three imaging units C1 to C3 are shown, but the number of imaging units may be arbitrary. 3, the inspection regions R1 to R3 arranged in the width direction are shown spaced apart from each other to make the inspection regions R1 to R3 easier to see, but they may be adjacent to each other, that is, they may be capable of inspecting the entire width direction of the molten glass G. Alternatively, as shown in FIG. 3, a plurality of inspection regions spaced apart from each other may be used. Furthermore, although each of the imaging units C1 to C3 is shown simply in FIG. 3, each of the imaging units C1 to C3 may be a combination of devices such as a plurality of cameras.

[0030] [An example of glass manufacturing procedures] FIG. 4 is a diagram (flowchart) showing an example of a procedure for manufacturing glass according to the embodiment. In this embodiment, the method for manufacturing a glass sheet includes a melting step S10, a molten glass transporting step S20, a shaping step S30, an annealing step S50, a cutting step S60, and a polishing step S70.

[0031] In the melting step S10, molten glass G is produced by melting glass raw materials.

[0032] In the molten glass transport step S20, the molten glass G is transported from a melting apparatus (in the example of FIG. 1, melting furnace 11) to a forming apparatus (in the example of FIG. 1, float bath 12). The molten glass transport step S20 may also include a fining step of degassing bubbles contained in the molten glass G.

[0033] In the forming step S30, the molten glass G produced in the melting step S10 is formed into a band-shaped glass ribbon. For example, in the forming step S30, the molten glass G is continuously supplied onto the surface of molten metal (molten tin M), and the liquid surface of the molten metal is used to form the molten glass G into a band-shaped glass ribbon. The glass ribbon is gradually solidified while flowing in a predetermined direction (from the upstream side to the downstream side of the float bath 12 in the example of FIG. 1).

[0034] In the annealing step S50, the glass ribbon formed in the forming step S30 is annealed in an annealing section (annealing furnace 13 in the example of FIG. 1).

[0035] In the cutting step S60, the annealed glass ribbon is cut into a predetermined size by a cutting machine in a cutting section (inspection and cutting section 14 in the example of FIG. 1) to obtain glass plates.

[0036] In the polishing step S70, the main surface of the glass plate is polished using a polishing slurry in a polishing unit (polishing unit 15 in the example of FIG. 1). The polishing slurry contains cerium oxide, zirconium oxide, manganese oxide, lanthanum, or red iron oxide as abrasive grains. Either the first or second main surface of the glass plate is polished to produce a highly smooth glass plate. Both the first and second main surfaces of the glass plate may be polished. From the viewpoint of improving productivity, the polishing amount is preferably 3.5 μm or less, more preferably 2 μm or less, even more preferably 1.5 μm or less, and particularly preferably 1.0 μm or less, but is not limited thereto. In this embodiment, the first and second main surfaces of the glass plate are the lower and upper surfaces of the glass plate, respectively.

[0037] [Classification device] FIG. 5 is a diagram illustrating an example of the configuration of the classification device 111 according to the embodiment. The classification device 111 is, for example, configured by a computer. The classification device 111 includes an input unit 131 , an output unit 132 , a communication unit 133 , a storage unit 134 , and a control unit 135 . The output unit 132 includes a display unit 151 . The control unit 135 includes a learning control unit 191 , a determination unit 192 , and a dressing control unit 193 .

[0038] The input unit 131 has a function of inputting information in response to an operation by a user (for example, a person), and a function of inputting information from an external device. The output unit 132 has a function of outputting information to the outside. For example, the display unit 151 has a function of displaying information to be displayed on a screen. The communication unit 133 has a function of communicating with an external device.

[0039] The storage unit 134 stores various types of information. In this embodiment, the storage unit 134 stores a learning model 171, training data 172, and the like. The learning model 171 may be set in advance in the storage unit 134, or may be set externally at any timing. The teacher data 172 may be set in advance in the storage unit 134, or may be set externally at any timing.

[0040] Here, any model may be used as the learning model 171, for example, a neural network model may be used. The learning model 171 extracts, for example, feature quantities (image feature quantities) of input image data and outputs an estimation result (determination result) based on the feature quantities. In this embodiment, the feature quantities are information representing the state of dross on the surface of the glass plate.

[0041] In this embodiment, the function of the feature extraction unit that extracts the features of image data and the function of the estimation unit that outputs an estimation result (judgment result) regarding the dross polishing time based on the extracted features are integrated and included in the learning model 171. As another example, the function of the feature extraction unit and the function of the estimation unit may be configured using separate learning models. In this case, image data is input to a learning model having the function of the feature extraction unit, and the output result (feature extraction result) from the learning model is input to a learning model having the function of the estimation unit, and the output from the learning model is used as the estimation result (determination result).

[0042] The control unit 135 performs various processes and controls. The learning control unit 191 uses the teacher data 172 to learn the learning model 171. This learning is machine learning. Note that, in the learning stage, for example, instructions based on the judgment results of the user may be used for learning.

[0043] The determination unit 192 inputs image data to be determined to the trained learning model 171, and outputs the output from the learning model 171 as a determination result via the output unit 132. For example, the judgment unit 192 may directly adopt the output (estimation result) from the trained learning model 171 as the judgment result by the judgment unit 192, or may make a further judgment (for example, trend judgment) based on the output (estimation result) from the trained learning model 171, and use the result of that judgment as the judgment result by the judgment unit 192.

[0044] Here, the image data to be determined may be input from the input unit 131, or may be stored in advance in the storage unit . For example, the judgment unit 192 may always perform judgment processing in real time by inputting image data acquired by the inspection unit 311 or image data processed from the image data in real time from the input unit 131. The determination process may be performed automatically in the classifier 111, for example.

[0045] The polishing control unit 193 controls a polishing unit (in this embodiment, the polishing unit 15 shown in FIG. 1) that polishes the glass plate. In this embodiment, the polishing control unit 193 controls the time for polishing performed by the polishing unit 15 (polishing time).

[0046] In this embodiment, attention is focused on the color of the dross shown in the image of the glass plate. As a result of extensive research, it has been found that the abrasiveness differs depending on the classification of the dross color in the dross image, the area of ​​each color, and the ratio of the areas of the colors to each other.

[0047] Here, the tin oxide-based dross that forms on the surface of a glass sheet produced by the float process has a number of different forms. As a specific example, dross in the form of an aggregation of particles with dimensions on the order of several hundred nanometers may be generated. Such dross may be called, for example, ordinary dross. It is believed that ordinary dross is generated when tin oxide particles adhering to the conveying rollers of the glass ribbon are transferred to the glass ribbon. As another specific example, dross may be generated in the form of a tin oxide granule with a size of several μm at the center, with a thin film of tin oxide several tens of nanometers spreading around the tin oxide granule. Such dross may be called, for example, mold-retained dross. It is believed that the mold-retained dross is formed when metallic tin adhering to the lower surface of the glass ribbon becomes tin oxide in the process of being crushed by the conveying rollers.

[0048] Of these tin oxide dross, normal dross does not pose a risk of breaking the wiring formed on the substrate surface, and can be removed relatively easily by polishing the substrate surface. On the other hand, in the case of mold-retained dross, the granular portions can be removed relatively easily by grinding, but the thin film-like portions surrounding them are less susceptible to grinding pressure and are also highly hard and resistant to wear, so it may be difficult to completely remove them even with a large amount of grinding. The polishing method is not particularly limited, and an example may be a method using a polishing pad made of foamed polyurethane, cerium oxide as an abrasive, and a 4B single-sided polisher.

[0049] <Example of an image containing dross> The white and black areas of the dross in the image of the glass plate are determined by the composition of the dross, such as, but not limited to, whether it is metallic tin or tin oxide. The time required for polishing such white and black portions tends to differ. In this embodiment, it is assumed that the more white parts of the dross there are, the longer it takes to polish, and the more black parts of the dross there are, the shorter the time it takes to polish. This assumption is merely an example, and there may be cases where, for example, the more black the dross, the longer it takes to polish, and the more white the dross, the shorter the time required to polish. The relationship between color and the time required for polishing is determined in advance, for example, by experiment or theory, and the processing is set to be carried out in accordance with this relationship.

[0050] In this embodiment, for convenience of explanation, the color state of the dross will be classified into types A, B, and C. In this embodiment, the color characteristics of types A, B, and C are relative, and the boundaries between the types may be arbitrarily defined. Type A dross is completely black. Type B dross is mostly black with a small amount of white mixed in. Type C dross is a mixture of black and white rather than just a little white. In this embodiment, it is assumed that the time required for polishing Type A and Type B is short, and the time required for polishing Type C is long.

[0051] FIG. 6A is a diagram showing an example of an image 511 including type A dross 531 according to the embodiment. FIG. 6B is a diagram showing an example of an image 512 including type B dross 532 according to the embodiment. FIG. 6C is a diagram showing an example of an image 513 including type C dross 533 according to the embodiment. In the examples of FIGS. 6A, 6B, and 6C, for convenience of illustration, the reference numerals of the dross 531, 532, and 533 are attached to the black portions of the dross. It should be noted that the examples of FIGS. 6A, 6B, and 6C are merely illustrative examples for the purpose of explanation, and the present invention is not necessarily limited to these examples.

[0052] <Example of time-dependent change in dross distribution across the width of a glass plate> 7(A), 7(B), and 7(C) are diagrams showing an example of the change over time in the distribution of dross in the width direction of the glass sheet according to the embodiment. FIG. 7(A) is an example where Type A is predominant, FIG. 7(B) is an example where Type B is predominant, and FIG. 7(C) is an example where Type C is predominant. 7(A), 7(B), and 7(C), the horizontal axis represents the passage of time, and the vertical axis represents the position (location) in the width direction of the glass sheet. The dross at each position in the width direction of the glass sheet is represented by a circle (◯), and the dross is colored white or black. It should be noted that the examples of FIGS. 7(A), 7(B), and 7(C) are merely illustrative examples for the purpose of explanation, and the present invention is not necessarily limited to these examples.

[0053] For example, the optimum polishing time is determined depending on a combination of characteristics such as the location (position) of the dross, the number of dross, and the amount of white or black in the dross. For example, even if there is a lot of Type A or Type B dross, the time required for polishing tends to be short, whereas even if there is a little Type C dross, the time required for polishing tends to be long.

[0054] <First example of learning and judgment> As input data (teacher data 172, data to be determined) of the learning model 171, data of an image containing dross is used. Then, the output data of the learning model 171 is used as data of the determination result regarding the time required to polish the dross.

[0055] Here, the judgment result regarding the time required to polish the dross may be, for example, one or more of information indicating the time required for polishing (itself), information indicating the range of time required for polishing, or information indicating whether the time required for polishing is long or short. Generally, it is thought that the shorter the time required for polishing, the easier the polishing will be, and the longer the time required for polishing, the more difficult the polishing will be.

[0056] As a specific example, the determination unit 192 may determine an index of the time required for polishing from an image captured (scanned) of information related to dross using a trained model (trained learning model 171). The determination unit 192 may, for example, determine an index of the time required for polishing for a large number of images that change over time in the same inspection areas R1 to R3, and perform trend aggregation of the determination results, thereby more accurately identifying (determining) the index.

[0057] In the trained model (trained learning model 171), for example, one or more of the following may be used as features: at least one of the area of ​​the white parts or the area of ​​the black parts of the dross; at least one of the shape of the white parts or the shape of the black parts of the dross; or at least one of the proportion of the white parts or the proportion of the black parts of the dross.

[0058] <Second example of learning and judgment> As input data (teacher data 172, data to be determined) of the learning model 171, data of an image containing dross and auxiliary data (auxiliary data) are used. Then, the output data of the learning model 171 is used as data of the determination result regarding the time required to polish the dross.

[0059] As a specific example, the judgment unit 192 may determine an indicator of the time required for polishing from an image captured (scanned) of information related to dross and auxiliary data using a trained model (trained learning model 171). The determination unit 192 may, for example, determine an index of the time required for polishing for a large number of images that change over time in the same inspection areas R1 to R3, and perform trend aggregation of the determination results, thereby more accurately identifying (determining) the index.

[0060] Here, the auxiliary data may include, for example, one or more of the location (position) of the dross and the number of the dross, regarding the dross. For example, compared to when auxiliary data is not used, the use of auxiliary data improves the accuracy of estimation (the accuracy of the determination result by the determining unit 192).

[0061] <Example of judgment results regarding polishing time> In this embodiment, the judgment result regarding the polishing time is an indicator of the time required for polishing, and may be, for example, information on the time required for polishing (polishing time) itself, or information used to derive the time required for polishing (polishing time). As a specific example, such an index may be used, which is defined for each of a plurality of types, such as Type A, Type B, and Type C. In this case, a polishing time may be set for each type, and a polishing time corresponding to each type may be set for dross in a state belonging to each type. As another example, an index suitable for each state of dross may be individually determined according to the state of each dross using a predetermined calculation formula or the like.

[0062] Here, the polishing time may be, for example, the polishing time for each individual dross, or the polishing time for a group of multiple dross. For example, when a unit area to be polished (polishing unit area) includes a plurality of drosses, a determination result regarding the polishing time for a set of these drosses may be acquired. As a specific example, image data obtained by capturing an image of the unit area to be polished (polishing unit area) may be used as image data (image data input to the learning model 171) used during machine learning learning and determination.

[0063] <Polishing time control> In this embodiment, in the classification device 111, the polishing control unit 193 controls the time (polishing time) of polishing performed by the polishing unit 15 shown in Fig. 1 based on the determination result by the determination unit 192. The polishing control unit 193 can, for example, change the polishing time for each unit area to be polished (polishing unit area). In this embodiment, the grinding control unit 193 is provided in the classification device 111, but as another example, the grinding control unit 193 may be provided in another device, and the other device may control the grinding time based on the judgment result obtained by the classification device 111.

[0064] <Example of imaging device> In this embodiment, a preferred example is a configuration in which a reflective optical camera is used as the imaging device (the imaging device constituting the imaging units C1 to C3 shown in FIG. 3). However, the imaging device is not limited to this, and other devices may also be used. Here, in an image captured by a reflective optical system, the metal parts that make up the dross (the shiny parts of the metal) appear white, and the parts other than the metal appear black. Furthermore, the positions at which the imaging devices are installed (the positions of the imaging units C1 to C3 shown in FIG. 3) are not necessarily limited to the example in FIG. 3, and they may be installed at other positions (other locations) in the manufacturing process of the glass plate.

[0065] As described above, the classification device 111 according to this embodiment can use machine learning to classify the appropriate polishing time for dross on glass sheets manufactured by the float process. As a result, the classification device 111 according to this embodiment can set the polishing time required for the polishing step to an appropriate time, and can shorten the polishing time while ensuring that it is sufficient, thereby improving the manufacturing efficiency of glass sheets. Here, for example, in the past, the amount of white area generated by the metallic luster of dross was not taken into consideration, but in this embodiment, the polishing time is estimated taking the amount of white area into consideration, thereby improving the accuracy of estimating the polishing time.

[0066] Regarding the trend in the length of polishing time required to polish dross, the trained learning model 171 may, for example, learn in accordance with the trend, or may learn in a way that deviates slightly from the trend.

[0067] In this embodiment, the classification device 111 acquires images of glass sheets manufactured by the float process, and performs image classification (learning of a learning model) related to dross in advance by machine learning. Then, the classification device 111 predicts the operational sensitivity of dross using the trained learning model 171.

[0068] It should be noted that the classification device 111 may be, for example, a non-online device or an online device (for example, a cloud server or an on-premise server).

[0069] As one configuration example, in the classification device 111, the judgment unit 192 inputs image data of a glass plate manufactured by the float method into a machine learning trained model, and obtains the output result from the trained model as a judgment result regarding the time required to polish the dross of the glass plate. Therefore, the sorting device 111 can sort the dross on the glass plate produced by the float method into appropriate polishing times.

[0070] As an example configuration, in the classification device 111, the trained model uses one or more of the following features: at least one of the area of ​​the white parts or the area of ​​the black parts of the dross; at least one of the shape of the white parts or the shape of the black parts of the dross; or at least one of the proportion of the white parts or the proportion of the black parts of the dross. Therefore, the classification device 111 can estimate an appropriate polishing time for the dross by utilizing the fact that the state of the white portion or the black portion or both of the dross is related to the polishing time.

[0071] As an example configuration, in the classification device 111, the determination unit 192 inputs image data and auxiliary data into a trained model. The auxiliary data includes one or more of the location of the dross or the number of dross. Therefore, in the classification device 111, the accuracy of classification can be further improved by using auxiliary data.

[0072] As an example of configuration, the classification device 111 includes an imaging device (imaging units C1 to C3 in the example of FIG. 3) that captures images of the image data. The imaging device is a reflective optical camera. Therefore, the classification device 111 can emphasize the difference between the white and black parts of the dross in the image data, thereby further improving the accuracy of classification.

[0073] As one configuration example, in the classification device 111, the image data is data of an image of a part of a glass plate, and includes a part in which dross is captured. Therefore, the sorting device 111 can sort the appropriate polishing time for each portion of the glass plate where dross is present.

[0074] As an example configuration, in the classification device 111, the learning control unit 191 generates a trained model using training data. Therefore, the classifier 111 can generate a trained model and use the trained model to classify the appropriate polishing time for the dross.

[0075] In this embodiment, the classification device 111 is shown to have both the learning control unit 191 and the judgment unit 192. However, as another example, the learning control unit 191 may not be provided in the classification device 111 but in another device, and the classification device 111 may use a trained model obtained by the other device.

[0076] A program for implementing the functions of any of the components of any of the above-described devices may be recorded on a computer-readable recording medium and loaded into a computer system for execution. The term "computer system" as used herein includes hardware such as an operating system or peripheral devices. The term "computer-readable recording medium" refers to portable media such as flexible disks, optical magnetic disks, ROMs, and compact discs (CDs) or read-only memories (ROMs), as well as storage devices such as hard disks built into computer systems. The term "computer-readable recording medium" also includes devices that retain a program for a certain period of time, such as volatile memory within a computer system that acts as a server or client when a program is transmitted over a network such as the Internet or a communication line such as a telephone line. Such volatile memory may be, for example, random access memory (RAM). The recording medium may also be, for example, a non-transitory recording medium.

[0077] The above program may be transmitted from a computer system storing the program in a storage device or the like to another computer system via a transmission medium or by transmission waves in the transmission medium. Here, the "transmission medium" that transmits the program refers to a medium that has the function of transmitting information, such as a network such as the Internet or a communication line such as a telephone line. The above program may also be one that realizes part of the above-mentioned functions. Furthermore, the above program may be a so-called differential file that can realize the above-mentioned functions in combination with a program already recorded in a computer system. A differential file may also be called a differential program.

[0078] Furthermore, the functions of any of the components in any of the above-described devices may be implemented by a processor. For example, each process in the embodiments may be implemented by a processor operating based on information such as a program and a computer-readable recording medium storing information such as the program. Here, the functions of each unit of the processor may be implemented by, for example, individual hardware, or may be implemented by integrated hardware. For example, the processor may include hardware, and the hardware may include at least one of a circuit for processing digital signals and a circuit for processing analog signals. For example, the processor may be configured using one or more circuit devices mounted on a circuit board, or one or both of one or more circuit elements. An integrated circuit (IC) or the like may be used as the circuit device, and a resistor or a capacitor may be used as the circuit element.

[0079] Here, the processor may be, for example, a CPU. However, the processor is not limited to a CPU, and various types of processors such as a GPU (Graphics Processing Unit) or a DSP (Digital Signal Processor) may be used. The processor may also be, for example, a hardware circuit such as an ASIC (Application Specific Integrated Circuit). The processor may also be, for example, composed of multiple CPUs, or may be, for example, composed of a hardware circuit such as a multiple ASIC. The processor may also be, for example, composed of a combination of multiple CPUs and a hardware circuit such as a multiple ASIC. The processor may also include, for example, one or more of an amplifier circuit or a filter circuit that processes analog signals.

[0080] The embodiments of this disclosure have been described in detail above with reference to the drawings, but the specific configuration is not limited to this embodiment, and includes designs within the scope that do not deviate from the gist of this disclosure.

[0081] [Note] (Configuration example 1) to (Configuration example 9) are shown.

[0082] (Configuration example 1) A determination unit is provided that inputs image data of a glass plate manufactured by the float method into a machine learning trained model and obtains an output result from the trained model as a determination result regarding the time required to polish the dross of the glass plate. Classification device.

[0083] (Configuration example 2) In the trained model, at least one of the area of ​​the white portion or the area of ​​the black portion of the dross, the shape of the white portion or the shape of the black portion of the dross, or the proportion of the white portion or the proportion of the black portion of the dross is used as a feature. The classification device according to (Configuration Example 1).

[0084] (Configuration example 3) The determination unit inputs the image data and auxiliary data into the trained model, The auxiliary data includes one or more of the location of the dross or the number of the dross. The classification device according to (Configuration Example 1) or (Configuration Example 2).

[0085] (Configuration Example 4) an imaging device that captures an image of the image data; The imaging device is a reflective optical camera. The classification device according to any one of (Configuration Example 1) to (Configuration Example 3).

[0086] (Configuration Example 5) The image data is data of an image of a portion of the glass plate, and includes a portion in which the dross is captured. The classification device according to any one of (Configuration Example 1) to (Configuration Example 4).

[0087] (Configuration Example 6) A learning control unit that generates the trained model using training data. The classification device according to any one of (Configuration Example 1) to (Configuration Example 5).

[0088] It is also possible to provide a method of processing (classification method) performed by the classification device. (Configuration Example 7) A classification device inputs image data of a glass plate manufactured by a float method into a machine learning trained model, and obtains an output result from the trained model as a determination result regarding the time required to polish the dross of the glass plate. Classification method.

[0089] It is also possible to provide a program that is executed by a computer (computer program). (Configuration Example 8) On the computer, A function to input image data of glass sheets manufactured using the float method into a machine learning model. A function of acquiring an output result from the trained model as a determination result regarding the time required to polish the dross of the glass plate; A program to achieve this.

[0090] It is also possible to provide pre-trained machine learning models. (Configuration Example 9) inputting image data of a glass plate produced by the float method, and outputting judgment result data regarding the time required to polish the dross of the glass plate; A trained machine learning model. [Explanation of symbols]

[0091] 1...glass manufacturing apparatus, 11...melting furnace, 12...float bath, 13...lehr, 14...inspection and cutting section, 15...polishing section, 21...inlet, 22...bathtub, 23...outlet, 24...side wall, 26...ceiling, 28...space, 30...gas supply path, 32...heater, 34...metal case, 36...bottom brick, 36a...top surface, 38...side brick, 111...sorting device, 131...input section, 132...output section, 133...communication section, 13 4...Memory unit, 135...Control unit, 151...Display unit, 171...Learning model, 172...Teacher data, 191...Learning control unit, 192...Determination unit, 193...Polishing control unit, 311...Inspection unit, 511 to 513...Image, 531 to 533...Dross, C1 to C3...Imaging unit, E...Glass width area, G...Molten glass, H1, H1a, H2, H2a, H11, H12...Defect area, M...Molten tin, R1 to R3...Inspection area

Claims

1. A determination unit is provided that inputs image data of a glass plate manufactured by the float method into a machine learning trained model and obtains an output result from the trained model as a determination result regarding the time required to polish the dross of the glass plate. Classification device.

2. In the trained model, at least one of the area of ​​the white portion or the area of ​​the black portion of the dross, the shape of the white portion or the shape of the black portion of the dross, or the proportion of the white portion or the proportion of the black portion of the dross is used as a feature. The classification device of claim 1 .

3. The determination unit inputs the image data and auxiliary data into the trained model, The auxiliary data includes one or more of the location of the dross or the number of the dross. The classification device according to claim 1 or claim 2.

4. an imaging device that captures an image of the image data; The imaging device is a reflective optical camera. The classification device according to claim 1 or claim 2.

5. The image data is data of an image of a portion of the glass plate, and includes a portion in which the dross is captured. The classification device according to claim 1 or claim 2.

6. A learning control unit that generates the trained model using training data. The classification device according to claim 1 or claim 2.

7. A classification device inputs image data of a glass plate manufactured by a float method into a machine learning trained model, and obtains an output result from the trained model as a determination result regarding the time required to polish the dross of the glass plate. Classification method.

8. On the computer, A function to input image data of glass sheets manufactured using the float method into a machine learning model. A function of acquiring an output result from the trained model as a determination result regarding the time required to polish the dross of the glass plate; A program to achieve this.

9. inputting image data of a glass plate produced by the float method, and outputting judgment result data regarding the time required to polish the dross of the glass plate; A trained machine learning model.

Images