WO2023188711A1 - Procédé d'analyse de tendance d'occurrence de défaut - Google Patents

Procédé d'analyse de tendance d'occurrence de défaut Download PDF

Info

Publication number
WO2023188711A1
WO2023188711A1 PCT/JP2023/001281 JP2023001281W WO2023188711A1 WO 2023188711 A1 WO2023188711 A1 WO 2023188711A1 JP 2023001281 W JP2023001281 W JP 2023001281W WO 2023188711 A1 WO2023188711 A1 WO 2023188711A1
Authority
WO
WIPO (PCT)
Prior art keywords
original fabric
sheet
defects
product
position information
Prior art date
Application number
PCT/JP2023/001281
Other languages
English (en)
Japanese (ja)
Inventor
洋介 村上
裕司 山下
清貴 堤
剛 神丸
空斗 深見
Original Assignee
日東電工株式会社
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 日東電工株式会社 filed Critical 日東電工株式会社
Priority to CN202380030286.5A priority Critical patent/CN118871775A/zh
Publication of WO2023188711A1 publication Critical patent/WO2023188711A1/fr

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
    • G01B11/00—Measuring arrangements characterised by the use of optical techniques
    • G01B11/30—Measuring arrangements characterised by the use of optical techniques for measuring roughness or irregularity of surfaces
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/84—Systems specially adapted for particular applications
    • G01N21/88—Investigating the presence of flaws or contamination
    • G01N21/89—Investigating the presence of flaws or contamination in moving material, e.g. running paper or textiles
    • G01N21/892—Investigating the presence of flaws or contamination in moving material, e.g. running paper or textiles characterised by the flaw, defect or object feature examined
    • G—PHYSICS
    • G02—OPTICS
    • G02B—OPTICAL ELEMENTS, SYSTEMS OR APPARATUS
    • G02B5/00—Optical elements other than lenses
    • G02B5/30—Polarising elements

Definitions

  • the present invention is manufactured by unwinding, transporting, and cutting a long raw material wound into a roll, such as a polarizing film or a retardation film, or a long film wound into a roll.
  • a long raw material wound into a roll such as a polarizing film or a retardation film, or a long film wound into a roll.
  • Defects that analyze trends related to the occurrence of defects based on the defects that have occurred in multiple sheet-like products produced by unwinding and transporting the raw material to cut out large-sized sheet-like intermediates, and cutting the intermediates.
  • Concerning methods for analyzing trends in the occurrence of In particular, the present invention relates to a method for analyzing the tendency of occurrence of defects, which can fully investigate the cause of defects occurring in sheet-like products and can sufficiently contribute to improving the yield of sheet-like products.
  • sheet-like products such as polarizing films and retardation films have been produced by unwinding a long raw material wound into a roll, conveying it, and cutting it, or by cutting a long raw material wound into a roll.
  • a large sheet-like intermediate body is cut out by unrolling and conveying the fabric, and by cutting this intermediate body, a plurality of sheet-like products are manufactured.
  • each manufactured sheet-like product is bonded to an optical display unit such as a liquid crystal panel, thereby manufacturing an optical display device such as a liquid crystal display device.
  • Patent Document 1 proposes a method of marking a mark representing positional information on the original fabric on the original fabric before cutting. Specifically, Patent Document 1 discloses that the original fabric of the sheet-shaped product is cut into the original fabric (rolled fabric in Patent Document 1) before being cut into the sheet-shaped product (optical film in Patent Document 1).
  • a method of marking a mark (roll information holding means in Patent Document 1) representing longitudinal position information (roll information in Patent Document 1) has been proposed (for example, paragraphs 0131 to 0133 of Patent Document 1). reference). According to the method described in Patent Document 1, it is considered that it is possible to some extent to contribute to investigating the cause of abnormalities in sheet products and improving the yield of sheet products.
  • the positional information represented by the mark marked on the sheet-like product is only the positional information in the longitudinal direction of the original fabric (Paragraph 0131 of Patent Document 1), so it is difficult to investigate the cause of defects. There is a problem in that the process cannot be carried out sufficiently, and as a result, it cannot sufficiently contribute to improving the yield of sheet-like products. Furthermore, even if the positional information represented by marks marked on sheet-like products includes not only the positional information in the longitudinal direction of the raw fabric but also the positional information in the widthwise direction of the raw fabric, how can they be used? Until now, there has been no proposal as to whether this method can be used to investigate the cause of defects.
  • the present invention was made in order to solve the problems of the conventional technology as described above, and when a defect occurs in a sheet-shaped product, the cause can be fully investigated, and the yield of sheet-shaped products can be improved. It is an object of the present invention to provide a method for analyzing the occurrence tendency of defects that can fully contribute to the development of problems.
  • the present invention is manufactured by unwinding a long raw material wound into a roll, conveying it and cutting it, or manufacturing a long raw material wound into a roll.
  • a generation tendency analysis method comprising: position information in the longitudinal direction and width direction in the original fabric, for the original fabric, the intermediate, or a plurality of sheet-like products before cutting or cutting out the intermediate.
  • a plurality of position information marks are provided on each of a plurality of sheet-like products, which are marks representing position information in the longitudinal direction and width direction of the original fabric. marked on sheet-like products. Therefore, in the reading process, by reading the positional information mark marked on each sheet-like product, it is possible to acquire the positional information (positional information in the longitudinal direction and width direction) on the original fabric of each sheet-like product. Then, in the analysis process, based on the positional information on the original fabric of each sheet-like product obtained in the reading process, the positional information on the original fabric of the defects present in each sheet-like product detected in the sheet-like product inspection process is identified. be able to.
  • the center position of each sheet-like product is It can be specified as information.
  • the position information of defects existing in each sheet-like product is known based on each sheet-like product, the position information on the original fabric of each sheet-like product obtained in the reading process and the position information on each sheet-like product as a reference Based on the positional information of the defects present in each sheet-like product, it is possible to specify the positional information on the original fabric of the defects present in each sheet-like product.
  • position information in the longitudinal direction and width direction of the original fabric represented by the position information mark of the present invention means information that includes the position at least in the longitudinal direction (conveying direction) of the original fabric and the width direction perpendicular to this. do.
  • the position in the longitudinal direction of the original fabric may be expressed as the distance from the longitudinal direction (transfer direction) tip of the original fabric (downstream end in the conveyance direction), or the distance from the longitudinal direction (transfer direction) tip of the original fabric. It may also be represented by serial numbers assigned according to the distance.
  • the position information mark may represent information regarding the original fabric for mutually identifying the plurality of original fabrics. Further, the position information mark may also represent other incidental information.
  • the position information mark may be marked using ordinary colored ink or transparent ink, or may be marked by laser engraving.
  • Transparent ink is ink that cannot be seen by the human eye under normal lighting, but becomes visible by emitting fluorescence when irradiated with light of a specific wavelength.
  • the transparent ink include UV ink that emits fluorescence when irradiated with ultraviolet light. Marking the position information mark using transparent ink has the advantage that the position information mark becomes invisible under normal lighting and does not spoil the appearance of the sheet-like product.
  • a one-dimensional code (barcode) or a two-dimensional code can be exemplified.
  • the two-dimensional code examples include DataMatrix (registered trademark) and QR code (registered trademark). Furthermore, if the sheet-shaped product has a layer such as a protective film that is removed when the sheet-shaped product is used (i.e., a layer that is not affected when the sheet-shaped product is used), the position information mark must be placed on that layer. is preferred.
  • the term "tendency related to the occurrence of defects" in the present invention refers to the tendency of defect occurrence (occurrence pattern) itself, as well as the tendency for defects to be detected in sheet products but difficult to detect in the original fabric, and Trends in the manufacturing process history information of sheet-like products where defects are more likely to be detected, trends in the types of sheet-like products (original materials) where defects are more likely to be detected, and inspections where defects are more likely to be detected when inspecting the sheet-like products Generally refers to trends related to the occurrence of defects, such as trends in methods and inspection conditions.
  • the present invention is applicable both when the inspection results of the original fabric are not used (when only the inspection results of the sheet-like product are used) and when the inspection results of the original fabric are used (the inspection results of the sheet-shaped product and the inspection results of the original fabric). (if both are used).
  • the analysis step is based on the position information on the original fabric of the identified defects present in the sheet-shaped product, and A map image generation procedure that generates a map image in which the position of the defect on the original fabric is plotted, and pattern matching is performed between the map image and a template image prepared in advance, so that the position of the defect is plotted on the map image.
  • the method includes a matching image generation procedure for generating a matching image, which is an image in which only the positions of the defects that match the template image among the positions of the defects are plotted, as an image representing a tendency related to the occurrence of defects.
  • generating a map image is not limited to the case where the map image is actually displayed on a display device such as a monitor, but also the position of the defect on the original fabric (original fabric)
  • This concept includes cases in which the map image is made displayable by specifying the location (in the longitudinal direction and the width direction) of the map image.
  • a map image is generated, which is an image in which the positions of defects present in the sheet product on the original fabric are plotted, and in the matching image generation procedure, the map image is generated.
  • a matching image is generated in which only the positions of the defects that match the template image are plotted among the positions of the defects plotted in the template image.
  • the "template image” may reflect the pattern you want to extract from the map image, such as an image in which pixels are plotted at a constant cycle in the longitudinal direction according to the cycle (outer circumference length) of the nip roller that conveys the original fabric. All you need to do is prepare an image.
  • an original fabric inspection step of inspecting the original fabric before cutting or cutting out the intermediate and a pre-cutting inspection result detected in the original fabric inspection step. or the position information of the defect existing in the original fabric before cutting out the intermediate, the type of defect existing in the original fabric, and execution of the original fabric inspection step in which the defect existing in the original fabric was detected.
  • a linking step of storing the location information mark in association with the location information mark, and the analysis step includes the location information and the location information on the original fabric of the identified defect existing in the sheet-like product.
  • a map image generation procedure that generates a The method includes a matching image generation procedure for generating a matching image, which is an image in which only the position is plotted, as an image representing a tendency related to the occurrence of defects.
  • the original fabric inspection step of inspecting the original fabric is not limited to being performed at a single location in the fabric manufacturing process, but can also be performed at multiple locations.
  • the position information of the defect existing in the original fabric detected in the original fabric inspection process, the type of defect existing in the original fabric, and the defect existing in the original fabric are detected.
  • the execution location of the original fabric inspection process is stored in association with the position information mark.
  • the defect detected in the original fabric inspection process is stored in association with a positional information mark marked on the same sheet-like product as the sheet-like product where the defect is located.
  • the position information on the original fabric of the defect existing in the sheet-like product identified as described above i.e., the position information on the original fabric of the sheet-like product obtained in the reading process is used
  • position information B the position information on the original fabric of the identified defect in the sheet-like product
  • position information A and position information B match means that position information A includes a position in the longitudinal direction and width direction of the original fabric, and position information B includes a position in the longitudinal direction and width direction of the original fabric. Not only when the positions in the longitudinal direction and the width direction of the original fabric included in the position information B are in a predetermined vicinity of the position in the longitudinal direction and the width direction of the original fabric included in the position information A. This concept also includes cases where there is a position in the width direction. In the combination identification procedure, if position information A and position information B match (in other words, a defect on the original fabric detected in the original fabric inspection process is located at the same position as a defect detected on the sheet-like product).
  • the defect that exists on the original fabric is stored in association with the position information mark read in the reading process.
  • a combination of the type of defect and the execution location of the original fabric inspection process where the defect existing in the original fabric was detected is specified.
  • position information B that is, there is a plurality of position information of defects existing on the original fabric stored in association with the position information mark read in the reading process
  • extraction condition determination step of the analysis step extraction conditions are determined based on the combination of type and execution location specified in the combination identification step.
  • a map image is generated in which the position of the defect on the original fabric is plotted.
  • defects detected for the corresponding original fabric are selected.
  • a map image is generated in which only the positions of defects on the original fabric that meet the conditions (combination of defect type and execution location of the original fabric inspection process) are plotted. Therefore, for example, by visually checking the matching image generated from this map image in the matching image generation procedure, it is possible to identify the tendency (occurrence pattern) of defects that may be detected in sheet products. trends).
  • the combination identification procedure if there are a large number of sheet products in which a defect was detected in the sheet product inspection process, for example, in the combination identification procedure, if a defect was detected in the sheet product inspection process, For each of the plurality of sheet products, a plurality of combinations of the types and the execution locations are specified, and in the extraction condition determination step, the combinations of the plurality of types and the execution locations specified in the combination identification step are determined. Among them, it is preferable that a majority of the combinations be determined as the extraction conditions.
  • the unit period for determining whether the number of multiple sheet products is large (in other words, the unit period for determining the majority of combinations as the extraction condition) may be, for example, one day or one week. can be mentioned.
  • the combination identification procedure if the number of sheet products in which a defect is detected in the sheet product inspection process is small, for example, in the combination identification procedure, the defect is detected in the sheet product inspection process.
  • the combinations of the plurality of types and the execution locations are specified, and in the extraction condition determination step, the combinations of the plurality of types and the execution locations identified in the combination identification step are determined. It is also possible to determine all combinations as the extraction conditions. If the number of multiple sheet products in which defects were detected in the sheet product inspection process is small, the number of combinations of multiple types and execution locations identified in the combination identification procedure is also likely to be small.
  • the defects detected in the sheet product inspection process are often not detected in the original fabric inspection process (in other words, the defects detected in the sheet product inspection process are If the defect is rarely detected in the inspection process), it may be possible to output this situation as a trend related to the occurrence of defects without generating a map image or matching image. That is, in the above preferred method, for example, in the combination identification step, a plurality of combinations of the types and the execution locations are identified for each of the plurality of sheet products in which a defect was detected in the sheet product inspection step. In the analysis step, the number of combinations of the plurality of types and the execution locations identified in the combination identification procedure is the total number of defects detected in the sheet-like product inspection step for the plurality of sheet-like products.
  • the extraction condition determination procedure, the map image generation procedure, and the matching image generation procedure are not executed. It is preferable to have an index output procedure for outputting the fact that defects are rarely detected even in the non-inspection process as an index representing a tendency related to the occurrence of defects.
  • the number of combinations of multiple types and execution locations specified in the combination identification procedure in other words, For example, the number of defects detected in both the sheet product and the original fabric is less than the majority of the total number of defects detected in multiple sheet products, indicating a trend in the occurrence of defects.
  • the original fabric inspection step of inspecting the original fabric before cutting or before cutting out the intermediate body, and the detection result detected in the original fabric inspection step.
  • the analysis step further includes a linking step of linking and storing an execution location of the original fabric and manufacturing process history information of the original fabric to the position information mark, and the analysis step includes Determine whether or not the positional information of the existing defect on the original fabric matches the positional information of the defect existing on the original fabric that is stored in association with the positional information mark read in the reading step.
  • a combination specifying procedure for specifying a combination of the execution location of the original fabric inspection process and the manufacturing process history information of the original fabric, and extraction conditions are determined based on the combination of the type and the execution location specified in the combination identification procedure.
  • the linking step the position information of the defect existing in the original fabric detected in the original fabric inspection process, the type of defect existing in the original fabric, and the In addition to the execution location of the original fabric inspection process where a defect existing in the fabric was detected, the manufacturing process history information of the original fabric is stored in association with the position information mark.
  • the combination identification procedure of the analysis process the type of defect existing in the original fabric, the execution location of the original fabric inspection process where the defect existing in the original fabric was detected, and the manufacturing process of the original fabric. A combination with process history information is specified.
  • the extraction condition determination step the extraction conditions are determined based on the combination of the specified type and execution location, and in the index output step, the existence of defects that meet the extraction conditions among all defects present in the original fabric is determined.
  • the specified manufacturing process history information of the original fabric is output as an index representing a tendency related to the occurrence of defects. Therefore, according to the above-mentioned preferred method, it is expected that trends in the manufacturing process history information of the original fabric in which defects may be detected in the sheet-like product can be appropriately grasped.
  • the combination identification procedure if there are a large number of sheet products in which a defect was detected in the sheet product inspection process, for example, in the combination identification procedure, if a defect was detected in the sheet product inspection process, For each of the plurality of sheet products, a plurality of combinations of the types and the execution locations are specified, and in the extraction condition determination step, the combinations of the plurality of types and the execution locations specified in the combination identification step are determined. Among them, it is preferable that a majority of the combinations be determined as the extraction conditions.
  • the present invention utilizes inspection results of the original fabric
  • the method further includes a linking step of storing a location and manufacturing process history information of the original fabric in a linked manner to the position information mark
  • the analysis step includes a step of storing a location and manufacturing process history information of the original fabric in association with the position information mark
  • the analyzing step includes a step of storing a location and manufacturing process history information of the original fabric in a manner that is associated with the location information mark. Determine whether or not the positional information of the defect on the original fabric matches the positional information of the defect existing on the original fabric that is stored in association with the positional information mark read in the reading step.
  • a combination identification procedure for identifying a combination of an inspection process execution location and manufacturing process history information of the original fabric, and extraction conditions are determined based on the combination of the type and execution location identified in the combination identification procedure. It has an extraction condition determination procedure, and a determination procedure for making a determination using a learning model as a determination of a tendency related to the occurrence of defects, and in the determination step, the Among all existing defects, the manufacturing process history information identified in the combination identification procedure of the original fabric in which defects that meet the extraction conditions determined in the extraction condition determination procedure exist is input to the learning model. The learning model then outputs a determination result as to whether or not a defect present in the original fabric that meets the extraction conditions is detected as a defect in the sheet-like product.
  • the manufacturing process history information identified for the original fabric having defects that meet the extraction conditions among all the defects present in the original fabric is input into the learning model,
  • the learning model outputs a determination result as to whether or not a defect present in the original fabric that meets the extraction conditions is detected as a defect in the sheet-like product. Therefore, according to the above-mentioned preferred method, it is possible to evaluate (determine) the possibility that a defect will be detected (occurrence) in a sheet-like product according to the manufacturing process history information of the original fabric.
  • the learning model may be generated by supervised learning using a combination of known inputs and outputs as training data, or may be generated by unsupervised learning. It's okay.
  • the cause when a defect occurs in a sheet-like product, the cause can be fully investigated, and it is possible to sufficiently contribute to improving the yield of sheet-like products.
  • FIG. 1 is a diagram schematically showing a schematic configuration of a system for executing an analysis method according to an embodiment of the present invention.
  • FIG. 1 is a flowchart schematically showing steps of a method for manufacturing a sheet-like product to which an analysis method according to an embodiment of the present invention is applied.
  • FIG. 3 is a diagram schematically illustrating the states of the original fabric S1 and the sheet-like product S2 in the manufacturing method shown in FIG. 2.
  • FIG. FIG. 2 is a flowchart schematically showing the steps of the analysis method according to the first embodiment. It is a figure which shows typically the example of a map image and a template image.
  • FIG. 7 is a flowchart schematically showing steps of an analysis method according to a second embodiment.
  • FIG. 1 is a diagram schematically showing a schematic configuration of a system for executing an analysis method according to an embodiment of the present invention.
  • FIG. 1 is a flowchart schematically showing steps of a method for manufacturing a sheet-like product to which an
  • FIG. 7 is a diagram schematically illustrating the contents of analysis step ST26 of the analysis method according to the second embodiment. It is a flowchart which shows the outline process of the analysis method based on 3rd Embodiment. It is a flowchart which shows the outline process of the analysis method based on 4th Embodiment. It is a flowchart which shows the outline process of the analysis method based on 5th Embodiment.
  • FIG. 3 is a diagram showing a schematic configuration of a manufacturing apparatus for executing an analysis method according to a modification of the present invention.
  • the method for analyzing the occurrence tendency of defects (hereinafter referred to simply as the "analysis method") according to the embodiments (first to fifth embodiments) of the present invention will be explained before cutting.
  • An example of marking a position information mark on the original fabric will be explained.
  • sheet products to which the analysis method according to the present embodiment is applied include optical products such as polarizing films, retardation films, visual compensation films, brightness enhancement films, and laminated films made by laminating two or more of these films.
  • An example is a film.
  • the sheet-like product is a laminated film in which a retardation film or the like is laminated on a polarizing film will be described as an example.
  • the structure of each film is publicly known, detailed description thereof will be omitted here.
  • FIG. 1 is a diagram schematically showing a schematic configuration of a system for executing the analysis method according to the present embodiment.
  • the arrow X indicates the transport direction (horizontal direction) of the original fabric S1 for manufacturing the sheet-like product S2
  • the arrow Y indicates the width direction (horizontal direction perpendicular to the transportation direction) of the original fabric S1.
  • the arrow Z indicates the normal direction (vertical direction) of the surface of the original fabric S1.
  • the system 100 of the present embodiment includes a sheet-like product S2 manufacturing apparatus 100a disposed in a sheet-like product manufacturing process for manufacturing a sheet-like product S2, and a sheet-like product S2 manufacturing apparatus 100a for inspecting the sheet-like product S2. and a reading device 6 disposed in the product inspection process.
  • the manufacturing device 100a of this embodiment includes an inspection device 1, a marking device 2, a cutting device 3, and a control analysis device 4.
  • the control analysis device 4 is electrically connected to the inspection device 1, the marking device 2, and the cutting device 3.
  • the manufacturing apparatus 100a of this embodiment includes a feed roller R1, a nip roller R2, and a conveyor R3.
  • the manufacturing apparatus 100a of this embodiment unwinds a long raw fabric S1 wound into a roll around a unwinding roller R1, conveys it in the X direction by a nip roller R2, a conveyor R3, etc., and cuts it by a cutting device 3.
  • the original fabric S1 is a long laminated film in which a retardation film or the like is laminated on a polarizing film, and is manufactured by known processes such as a stretching process, an adhesion process, and a bonding process.
  • the inspection device 1 is a device that detects defects present in the original fabric S1 by inspecting the original fabric S1 before cutting.
  • the inspection apparatus 1 shown in FIG. 1 includes a light source 11 that is disposed on one side of the original fabric S1 in the Z direction (in the example shown in FIG. 1, below the original fabric S1) and emits light toward the original fabric S1. , is placed on the other side in the Z direction with respect to the original fabric S1 (in the example shown in FIG. 1, above the original fabric S1), and by receiving the light that has passed through the original fabric S1 and forming an image (imaging), a transmitted image is created.
  • control analysis device 4 By applying known image processing such as binarization to extract a pixel region having a different luminance value from other pixel regions to the transmitted image input from the imaging means 12, An image processing means 13 for detecting defects is provided.
  • the control analysis device 4 also functions as the image processing means 13 of the inspection device 1, but it is also possible to provide the image processing means 13 separately from the control analysis device 4.
  • the inspection device 1 is not limited to the configuration that detects defects based on a transmitted image, but is generated by arranging a light source and an imaging means on one side in the Z direction with respect to the original fabric S1. It is also possible to adopt a configuration in which defects are detected based on reflected images.
  • the inspection device 1 is arranged with a light source and a polarizing filter for inspection on one side of the original fabric S1 in the Z direction, and Either an imaging means is placed on the other side in the Z direction with respect to the original web S1, or a light source is placed on one side of the original web S1 in the Z direction, and polarized light for inspection is placed on the other side of the original web S1 in the Z direction. It is also possible to adopt a configuration in which defects are detected based on a crossed Nicol image generated by arranging a filter and an imaging means.
  • FIG. 1 shows a case where one inspection device 1 is used to inspect the original fabric S1 at a single location, but the present invention is not limited to this, and the manufacturing device 100a is It is also possible to arrange the inspection device 1 in the previous process and inspect the original fabric S1 at a plurality of locations.
  • the control analysis device 4 (image processing means 13) that constitutes a part of the inspection device 1 is capable of recognizing the position (XY coordinates) of a defect within the transmitted image. Therefore, for example, the control analysis device 4 determines the position (XY coordinates) of the recognized defect in the transmitted image, the separation distance L1 in the X direction between the feeding roller R1 and the inspection device 1, and the encoder attached to the nip roller R2 (Fig. It is possible to recognize the position of a defect present in the original fabric S1 on the basis of the conveyance amount of the original fabric S1 measured by the method (not shown) or the like.
  • the marking device 2 is a device that marks the original fabric S1 before cutting with a position information mark that is a mark representing positional information on the original fabric S1.
  • the control analysis device 4 controls the marking device 2 to mark position information marks such that a plurality of position information marks are present on each of the plurality of cut sheet products S2.
  • the marking device 2 of this embodiment is an inkjet marking device using transparent ink (specifically, UV ink), and by discharging transparent ink from a large number of nozzles arranged along the Y direction. , and is configured to mark a position information mark.
  • transparent ink specifically, UV ink
  • the marking device 2 it is also possible to adopt a configuration in which marking is performed by an inkjet method using ordinary colored ink, or a configuration in which marking is performed by laser engraving.
  • ink it is also possible to adopt a configuration in which marking is performed not only by an inkjet method but also by a pen method. Since the specific configurations of these marking devices are well known, detailed explanation thereof will be o
  • the cutting device 3 is a device that manufactures a plurality of sheet-like products S2 by cutting the original fabric S1 using a known processing method such as punching or laser processing. Specifically, the control analysis device 4 controls the cutting device 3, so that the original fabric S1 is cut along a predetermined cutting line.
  • the control analysis device 4 is comprised of a computer in which a program for controlling the operations of the marking device 2 and the cutting device 3 and a program for executing the analysis method described below are installed.
  • the manufacturing apparatus 100a includes a marking device (defect mark) that marks, on the original fabric S1 before cutting, a mark (defect mark) at a position where a defect is detected by inspecting the original fabric S1 using the inspection device 1 or the like. (not shown) may also be provided.
  • the control analysis device 4 recognizes the position where a defect exists, which is detected by inspecting the original fabric S1 with the inspection device 1 etc., and the control analysis device 4 controls the above-mentioned marking device.
  • a defect mark may be marked at a position where a defect exists.
  • the above marking device similarly to the marking device 2, it is possible to adopt a configuration in which marking is performed using an inkjet method using normal colored ink, or a configuration in which marking is performed by laser engraving.
  • FIG. 2 is a flowchart schematically showing the steps of a method for manufacturing a sheet-like product to which the analysis method according to the present embodiment is applied.
  • the manufacturing method of this embodiment includes a position information marking step ST1, a cutting step ST2, and a collection step ST3.
  • the position information marking step ST1 described here is the same as the position information marking steps ST11, ST22, ST32, ST42, and ST52 of the first to fifth embodiments described later.
  • FIG. 1 is the same as the position information marking steps ST11, ST22, ST32, ST42, and ST52 of the first to fifth embodiments described later.
  • FIG. 3 is a diagram schematically explaining the state of the original fabric S1 and the sheet-like product S2 in the manufacturing method shown in FIG.
  • FIG. 3(a) is a diagram schematically showing the state of the original fabric S1 before performing the position information marking step ST1.
  • FIG. 3(b) is a diagram schematically showing the state of the original fabric S1 immediately after performing the position information marking step ST1.
  • FIG. 3(c) is a diagram schematically showing the state of the original fabric S1 (sheet-like product S2) immediately after performing the cutting step ST2.
  • FIG. 3(d) is a diagram schematically showing the state of the sheet-like product S2 in the process of performing the collection step ST3.
  • Each step ST1 to ST3 will be explained in order below.
  • FIG. 3(a) by inspecting the original fabric S1 before cutting with the inspection device 1, defects F (in FIG. 3(a), for convenience, all defects F are shown as black circles) is detected.
  • the marking device 2 marks the raw web S1 before cutting with a positional information mark M representing positional information on the raw web S1.
  • the positional information represented by the positional information mark M means information including the position (that is, the XY coordinates) in at least the longitudinal direction (conveyance direction, X direction) and width direction (Y direction) of the original fabric S1.
  • the position in the longitudinal direction of the original fabric S1 may be expressed by the distance from the longitudinal direction (transfer direction) tip of the original fabric S1, or may be expressed according to the distance from the longitudinal direction (transfer direction) tip of the original fabric S1. It may also be represented by serial numbers.
  • the positional information represented by the position information mark M includes the position of the web S1 in the conveyance direction, and the position information mark M also includes information regarding the web S1 for mutually identifying the plurality of webs S1. may be displayed, or other incidental information may be displayed.
  • the position information mark M of this embodiment is DataMatrix (registered trademark), which is a type of two-dimensional code.
  • the present invention is not limited to this, and in addition to other two-dimensional codes such as QR code (registered trademark), one-dimensional codes (barcodes), as long as the positional information on the original fabric S1 can be expressed. , various forms of marks can be employed as the position information mark M.
  • the marking device 2 marks the original fabric S1 before cutting with a positional information mark M so that the positional information mark M is present on each of the plurality of sheet products S2 after cutting.
  • a planned cutting line CL shown by a broken line in FIG. 3(b) is determined in advance according to the dimensions and shape of the sheet-like product S2, and is stored in the control analysis device 4.
  • the planned cutting line CL is not actually drawn on the original fabric S1, but is stored, for example, in XY coordinates with the tip of the original fabric S1 in the longitudinal direction of transport as a reference.
  • the planned cutting lines CL are in the form of a grid, and within each rectangle (18 rectangles are shown in FIG.
  • the marking device 2 sets the position information so that the position information mark M exists within each rectangle defined by the planned cutting line CL, and so that it does not overlap with the planned cutting line CL. Mark mark M.
  • one positional information mark M is marked in each rectangle, and the position of the positional information mark M with respect to the cutting line CL is the same for all rectangles (each center of the rectangle).
  • the present invention is not limited to this.
  • a position information mark M may be marked.
  • the positions of the position information marks M to be marked may be different from each other.
  • the control analysis device 4 calculates the separation distance L2 in the X direction between the feed roller R1 and the marking device 2, and the original data measured by an encoder (not shown) attached to the nip roller R2.
  • the timing at which a predetermined portion of the original fabric S1 reaches the marking device 2 can be calculated based on the conveyance amount of the fabric S1.
  • the control analysis device 4 determines the timing at which a portion having a predetermined XY coordinate based on the cutting line CL (a portion to be marked with the position information mark M in each rectangle) reaches the marking device 2.
  • the marking device 2 is controlled to eject transparent ink from the nozzle of the marking device 2 corresponding to the Y coordinate of the predetermined XY coordinates. As a result, position information marks M are marked within each rectangle.
  • the cutting step ST2 As shown in FIG. 3(c), in the cutting step ST2, the original fabric S1 marked with the position information mark M is cut along the cutting line CL (see FIG. 3(b)) by the cutting device 3, and a plurality of A sheet-like product S2 is manufactured.
  • a punching device is used as the cutting device 3
  • the conveyance of the material S1 is temporarily stopped and the material S1 is cut at the timing when the planned cutting line CL of the material S1 reaches the cutting device 3. It turns out.
  • a laser processing device is used as the cutting device 3, it is also possible to cut the original fabric S1 while it is being conveyed without stopping the conveyance of the original fabric S1.
  • the control analysis device 4 determines the separation distance L3 in the X direction between the feed roller R1 and the cutting device 3, and the raw material S1 measured by an encoder (not shown) attached to the nip roller R2. It is possible to calculate the timing at which the scheduled cutting line CL of the original fabric S1 stored in the control analysis device 4 reaches the cutting device 3 based on the conveyance amount. Then, the control analysis device 4 controls and drives the cutting device 3 at the timing when the planned cutting line of the original fabric S1 reaches the cutting device 3. Thereby, the original fabric S1 is cut along the planned cutting line CL, and a plurality of sheet-like products S2 are manufactured.
  • the sheet-like product S2 manufactured as described above is carried out to a sheet-like product inspection process, as shown in FIG. 1, and is visually inspected by an inspector 5, for example.
  • the present invention is not limited to this, and it is also possible to perform the inspection using an optical automatic inspection device (not shown).
  • the sheet-like product manufacturing process is executed at a sheet-like product manufacturing factory, and the sheet-like product inspection process is an optical system for manufacturing an optical display device using the sheet-like product S2 shipped from the sheet-like product manufacturing factory. It is also conceivable that the process may be executed at a display device manufacturing factory.
  • the reading device 6 When inspecting the sheet-like product S2 at an optical display device manufacturing factory, for example, but not limited to this, a known test such as a lighting test may be performed with the sheet-like product S2 bonded to the optical display unit. is possible. If a defect is detected through the inspection, the reading device 6 reads the position information mark M marked on the sheet-like product S2.
  • the reading device 6 is a device that reads the position information mark M marked on the sheet-like product S2, and a known two-dimensional code reader or the like is used as the reading device 6.
  • position information (XY coordinates) of the sheet-like product S2 on the original fabric S1 is acquired.
  • the acquired position information is input to the control analysis device 4.
  • the acquired position information may be input manually by a person to the control analysis device 4, or the reading device 6 and the control analysis device 4 may be electrically connected via a wireless communication line, etc. 6 may be automatically input to the control analysis device.
  • FIG. 4 is a flow diagram schematically showing the steps of the analysis method according to the first embodiment.
  • the analysis method according to the first embodiment includes a position information marking step ST11, a sheet-like product inspection step ST12, a reading step ST13, and an analysis step ST14.
  • the analysis method according to the first embodiment is a method that does not use the inspection results of the original fabric S1. That is, the inspection device 1 is not necessary to execute the analysis method according to the first embodiment.
  • Each step ST12 to ST14 will be explained in order below.
  • Positional information marking step ST11 In the positional information marking step ST11, using the marking device 2, a plurality of positional information marks M, which are marks representing positional information in the longitudinal direction and the width direction on the raw web S1, are marked on the raw web S1 before cutting in the form of a plurality of sheets. A position information mark M is marked so that it exists on each product S2.
  • the reading device 6 is used to read the position information mark M marked on the sheet-like product S2 in which the defect was detected in the sheet-like product inspection step ST12. Get location information.
  • the position information acquired in the reading step ST13 is input to the control analysis device 4, and the control analysis device 4 analyzes the sheet based on the position information on the original fabric S1 of the sheet-like product S2 acquired in the reading step ST13.
  • the positional information on the original fabric S1 of the defect present in the sheet-like product S2 detected in the sheet-like product inspection step ST12 is specified.
  • the control analysis device 4 specifies the center position of each sheet-like product S2 as position information on the original fabric S1 of a defect present in each sheet-like product S2.
  • each sheet-like product S2 is It is possible to specify positional information on the original fabric S1 of existing defects. Then, in the analysis step ST14, the control analysis device 4 is used to analyze trends related to the occurrence of defects based on the position information in the original fabric S1 of the defects present in the identified sheet-like product S2.
  • the analysis step ST14 of the analysis method according to the first embodiment includes a map image generation procedure ST141 and a matching image generation procedure ST142.
  • FIG. 5 is a diagram schematically showing an example of a map image and a template image. 5(a) to 5(d) show examples of map images, and FIGS. 5(e) to 5(i) show examples of template images.
  • map image generation procedure ST141 In the map image generation procedure ST141, the control analysis device 4 plots the position of the defect existing in the sheet-like product S2 on the original web S1 based on the position information on the original web S1 of the defect existing on the identified sheet-like product S2.
  • a map image MP is generated as a map image MP.
  • the area divided by each rectangle corresponds to the original fabric S1
  • the position of the defect existing in the sheet-like product S2 on the original fabric S1 is indicated by a black circle. Indicated by a white circle.
  • pixels indicated by black circles and white circles have brightness values (pixel values) different from other pixels.
  • FIGS. 5(a) to 5(d) the area divided by each rectangle corresponds to the original fabric S1
  • white circle Indicated by a white circle.
  • pixels indicated by black circles and white circles have brightness values (pixel values) different from other pixels.
  • a map image MP is generated by combining a plurality of continuously manufactured original fabrics S1 (original fabrics S1a to original fabric S1d) in the X direction.
  • original fabrics S1a to original fabric S1d original fabrics S1a to original fabric S1d
  • the present invention is not limited to this, and a map image MP may be generated for each single web S1.
  • the control analysis device 4 performs pattern matching between the map image MP and the template image TE prepared in advance as shown in FIGS. 5(e) to 5(i) to generate the map image MP.
  • a matching image which is an image in which only the positions of defects that match the template image TE among the positions of defects plotted in , is plotted is generated as an image representing a tendency related to the occurrence of defects. That is, the control analysis device 4 scans the template image TE on the map image MP, and extracts a pixel region of the map image MP where the degree of coincidence between the map image MP and the template image TE is equal to or higher than a predetermined threshold. Then, a matching image is generated.
  • the template image TE shown in FIG. 5(e) is an image in which pixels indicated by white circles are plotted at a constant period P in the X direction, and this period P is the period (outer circumference length) of the feed roller R1, nip roller R2, etc. It will be set accordingly. It is preferable to prepare a plurality of template images TE with different periods P depending on the period of each roller used in the manufacturing process of the sheet-like product S2.
  • the pixel indicated by a white circle has a different brightness value (pixel value) from other pixels, and has the same brightness value as the pixel indicated by a black circle and a white circle in the map image MP. The same applies to the template images shown in FIGS. 5(f) to 5(i).
  • the map image shown in FIG. White circle pixels in MP are extracted, and a matching image is generated in which only the white circle pixels are plotted (black circle pixels are not plotted).
  • the three template images TE shown in FIG. 5(f) are all images in which pixels shown as white circles clustered in a predetermined shape are plotted.
  • the template image TE shown at the top of FIG. 5(f) is an example of an image in which pixels clustered in a parallelogram shape is plotted, and the template image TE shown in the center of FIG. 5(f) is an example of an image in which pixels clustered in a straight line.
  • the template image TE shown at the bottom of FIG. 5(f) shows an example of an image in which pixels clustered in a rhombus shape is plotted; An image in which pixels are plotted densely in an arbitrary shape, such as a square shape, a cross shape, etc., can be used as the template image TE.
  • the image shown in FIG. 5(f) For example, by scanning the template image TE shown at the top of FIG. 5(f) in the X direction and the Y direction on the map image MP shown in FIG. 5(b) and performing pattern matching, the image shown in FIG.
  • the pixels of the white circles in the map image MP shown are extracted, and a matching image is generated in which only the pixels of the white circles are plotted (the pixels of the black circles are not plotted).
  • the template image TE shown in FIG. 5(g) is an image in which pixels indicated by white circles are plotted at a specific Y coordinate (Y1) and a specific X coordinate (X1).
  • Y1 a specific Y coordinate
  • X1 a specific X coordinate
  • the map image MP shown in FIG. 5(c) is created.
  • a white circle pixel pixel whose Y coordinate is Y1 is extracted, and a matching image is generated in which only the white circle pixels are plotted (black circle pixels are not plotted).
  • the template image TE shown in FIG. 5(h) is located at the leading edge in the X direction of the map image MP (corresponding to the leading edge in the longitudinal direction of the raw fabric S1), and at the rear end in the X direction of the map image MP (corresponding to the trailing edge in the longitudinal direction of the raw fabric S1). (equivalent).
  • the template image TE shown in FIG. 5(h) is an image for extracting defects present at the leading edge and trailing edge of the map image MP.
  • pattern matching is performed by scanning the template image TE shown in FIG. 5(h) in the Y direction on the tip of the map image MP shown in FIG. 5(d) in the X direction (the upper end in FIG. 5(d)). By doing so, the white circle pixels in the map image MP shown in FIG. 5(d) are extracted, and a matching image is generated in which only the white circle pixels are plotted (the black circle pixels are not plotted). .
  • the template image TE shown in FIG. 5(i) is an image in which pixels indicated by white circles are plotted so as to swing in the Y direction. Although a specific example will be omitted, by using this template image TE for pattern matching, it is possible to create a matching image that corresponds to a defect that occurs at a position where the web S1 swings in the width direction due to meandering or the like of the web S1. It is possible to generate
  • the map image generation step ST141 of the analysis step ST14 an image in which the positions of defects present in the sheet-like product S2 on the original fabric S1 are plotted is generated.
  • a certain map image MP is generated, and in matching image generation step ST142, a matching image is generated in which only the positions of defects that match the template image TE are plotted among the positions of defects plotted on the map image MP. Ru. Therefore, for example, by visually confirming the matching image, it is possible to grasp the tendency related to the occurrence of defects detected in the sheet-like product S2 (the tendency of the occurrence (occurrence pattern) of defects).
  • FIG. 6 is a flow diagram schematically showing the steps of the analysis method according to the second embodiment.
  • the analysis method according to the second embodiment includes an original fabric inspection process ST21, a position information marking process ST22, a linking process ST23, a sheet-like product inspection process ST24, a reading process ST25, analysis step ST26.
  • the analysis method according to the second embodiment is different from the analysis method according to the first embodiment in that it uses the inspection results of the original fabric S1. Therefore, the analysis method according to the second embodiment includes an original fabric inspection step ST21 and a linking step ST23, which the analysis method according to the first embodiment does not have. Further, the contents of the analysis step ST26 are also different from the analysis step ST14 of the analysis method according to the first embodiment.
  • Each step ST21 to ST26 will be explained in order below.
  • the inspection device 1 is used to inspect the original fabric S1 before cutting, and detect defects present in the original fabric S1.
  • the inspection device 1 can be placed at a plurality of locations, and the inspection device 1 can be configured to detect defects based on a transmitted image, a reflected image, or a crossed Nicol image.
  • the control analysis device 4 determines the defect based on which image the defect was detected, in which location the inspection device 1 placed it, and the location information of the defect on the original fabric S1. The type can be determined.
  • the marking device 2 is used to mark the length and width of the original web S1 on the original web S1 before cutting.
  • the position information mark M is marked so that the position information mark M, which is a mark representing the position information in the direction, is present on each of the plurality of sheet-like products S2.
  • the control analysis device 4 uses the position information of the defects present in the original fabric S1 before cutting, which were detected in the original fabric inspection process ST21, the types of defects present in the original fabric S1, and the original fabric.
  • the execution location of the original fabric inspection step ST21 location of the inspection device 1 where the defect existing in S1 was detected is stored in association with the position information mark M.
  • the control analysis device 4 stores it in association with the position information mark M marked on the same sheet-like product S2 as the sheet-like product S2 where the defect detected in the original fabric inspection step ST21 is located.
  • the reading device 6 is used to read the position information marked on the sheet-like product S2 in which the defect was detected in the sheet-like product inspection step ST24.
  • the mark M positional information of the sheet-like product S2 on the original fabric S1 is acquired.
  • the position information acquired in the reading step ST25 is input to the control analysis device 4, and the control analysis device 4 acquires it in the reading step ST25. Based on the positional information on the original fabric S1 of the sheet-like product S2, the positional information on the original fabric S1 of the defect existing in the sheet-like product S2 detected in the sheet-like product inspection step ST24 is specified. Then, in the analysis step ST26, similarly to the analysis step ST14 of the analysis method according to the first embodiment, the control analysis device 4 is used to determine the position information in the original fabric S1 of the defect existing in the identified sheet-like product S2.
  • the analysis step ST26 of the analysis method according to the second embodiment includes a combination identification procedure ST261, an extraction condition determination procedure ST262, a map image generation procedure ST263, and a matching image generation procedure.
  • FIG. 7 is a diagram schematically explaining the contents of the analysis step ST26 of the analysis method according to the second embodiment.
  • Combination identification procedure ST261 In the combination identification procedure ST261, the control analysis device 4 extracts the positional information on the original fabric S1 of the defects present in the identified sheet-like product S2 (i.e., the positional information on the original fabric S1 of the sheet-like product S2 acquired in the reading step ST25).
  • the position information (position information A) on the original fabric S1 of the defect existing in the sheet-like product S2 identified based on It is determined whether the position information of the existing defect (that is, the position information (position information B) of the defect existing in the original fabric S1 detected in the original fabric inspection step ST21) matches.
  • the control analysis device 4 detects the position information mark M read in the reading step ST25.
  • a combination of the type of defect existing in the original fabric S1, which is stored in association with the original fabric S1, and the execution location of the original fabric inspection step ST21 where the defect existing in the original fabric S1 was detected is specified. For example, No. 1 ⁇ No.
  • Defects are detected in each of the 11 sheet-like products S2 of 11, and whether or not each defect was also detected in the original fabric inspection step ST21 (whether or not position information A and position information B match), and the original Let us consider a case in which the type of defect and the execution location of the original fabric inspection process ST21 are as shown in Table 1 below when a defect is also detected in the non-inspection process ST21.
  • Table 1 shows the type of defect and the execution location of the original fabric inspection process ST21.
  • No. 1 in which no defects were detected in the original fabric inspection step ST21. 3 and no.
  • the nine combinations applied) are specified in the combination specifying procedure ST261.
  • extraction condition determination procedure ST262 the control analysis device 4 determines extraction conditions based on the combination of type and execution location specified in combination identification step ST261. Specifically, the control analysis device 4 determines, for example, a majority of the combinations of types and execution locations identified in the combination identification procedure ST261 as extraction conditions. The method of determining the majority of combinations as extraction conditions is preferably used when the number of sheet products S2 in which defects are detected in the sheet product inspection step ST24 is large. In the case shown in Table 1 above, among the nine identified combinations of type and execution location, the number of combinations where the defect type is "foreign object" and the execution location is “stretching process” is 5, which is the majority. Therefore, a combination of "foreign matter" and "stretching process” is determined as the extraction condition.
  • the two sets of combinations applied) are specified in the combination specifying procedure ST261.
  • all of the two combinations of types and execution locations identified in the combination identification step ST261 are determined as extraction conditions. Based on the combination, extraction conditions can be determined.
  • the method of determining all combinations as extraction conditions is suitably used when the number of sheet products S2 in which defects are detected in the sheet product inspection step ST24 is small.
  • extraction conditions are determined using only one of the above-described method of determining the majority of combinations as extraction conditions or the method of determining all combinations as extraction conditions.
  • both methods are switched depending on the number of sheet products S2 in which defects are detected in the sheet product inspection step ST24.
  • Map image generation procedure ST263 In the map image generation procedure ST263, the control analysis device 4 extracts defects (black circles in FIGS. 7(a) and 7(b)) existing in the original fabric S1 detected in the original fabric inspection process ST21. Based on the position information on the original web S1 of the defect that meets the extraction conditions determined in the condition determination step ST262, a map is an image in which the position of the defect on the original web S1 is plotted, as shown in FIG. 7(c). Generate image MP.
  • the control analysis device 4 converts the map image MP as shown in FIG. 7(c) and the template image TE prepared in advance (see FIG. 5(e) to FIG. 5(i)) into a pattern.
  • a matching image MA is created, which is an image in which only the positions of defects that match the template image TE are plotted among the positions of defects plotted on the map image MP. It is generated as an image representing the tendency related to the occurrence of defects.
  • the specific contents of the matching image generation procedure ST264 are the same as the matching image generation procedure ST142 of the analysis method according to the first embodiment, so a detailed explanation will be omitted here.
  • the defects detected in the sheet-like product S2 are A map in which only the positions of defects in the original fabric S1 that meet the same conditions (combination of the type of defect and the execution location of the original fabric inspection step ST21) as the defect (the defect detected on the corresponding original fabric S1) are plotted.
  • Image MP will be generated. Therefore, for example, by visually recognizing the matching image MA generated from the map image MP in the matching image generation step ST264, it is possible to determine the tendency (occurrence of defects) related to the occurrence of defects that may be detected in the sheet product S2. It is possible to understand the trends (occurrence patterns).
  • FIG. 8 is a flow diagram schematically showing the steps of the analysis method according to the third embodiment.
  • the analysis method according to the third embodiment includes an original fabric inspection process ST31, a position information marking process ST32, a linking process ST33, a sheet-like product inspection process ST34, a reading process ST35, analysis step ST36.
  • the analysis method according to the third embodiment is also a method that utilizes the inspection results of the original fabric S1, similarly to the analysis method according to the second embodiment.
  • the contents of the original fabric inspection step ST31, position information marking step ST32, linking step ST33, sheet-like product inspection step ST34, and reading step ST35 of the analysis method according to the third embodiment are the same as those of the analysis method according to the second embodiment.
  • the contents are the same as the original fabric inspection process ST21, position information marking process ST22, tying process ST23, sheet-like product inspection process ST24, and reading process ST25. Therefore, in the following, the differences between the analysis step ST36 and the analysis step ST26 of the analysis method according to the second embodiment will be mainly explained, and detailed explanation of the other steps will be omitted.
  • the position information acquired in the reading step ST35 is input to the control analysis device 4, and the control analysis device 4 acquires it in the reading step ST35. Based on the positional information on the original fabric S1 of the sheet-like product S2, the positional information on the original fabric S1 of the defect existing in the sheet-like product S2 detected in the sheet-like product inspection step ST34 is specified. Then, in the analysis step ST36, similarly to the analysis step ST26 of the analysis method according to the second embodiment, the control analysis device 4 is used to calculate the position information in the original fabric S1 of the defect existing in the identified sheet-like product S2. Based on this, trends related to the occurrence of defects are analyzed. However, the analysis step ST36 differs from the analysis step ST26 in the following points.
  • the control analysis device 4 determines that the number of combinations of the plurality of types and execution locations specified in the combination identification procedure ST361, which is similar to the analysis method combination identification procedure ST261 according to the second embodiment, It is determined whether or not the total number of defects detected in the sheet-shaped product S2 in the sheet-shaped product inspection step ST34 is less than the majority (ST362 in FIG. 8). Then, if the number of combinations of the plurality of specified types and execution locations is not less than the majority (that is, the majority) (“No” in ST362 in FIG. 8), the control analysis device 4 performs the extraction condition determination procedure. ST364, map image generation procedure ST365, and matching image generation procedure ST366 are executed.
  • extraction condition determination procedure ST364, map image generation procedure ST365, and matching image generation procedure ST366 are the contents of extraction condition determination procedure ST262, map image generation procedure ST263, and matching image generation procedure ST264 of the analysis method according to the second embodiment, respectively. It is similar to On the other hand, if the number of identified combinations of types and execution locations is less than the majority (that is, the majority) (“Yes” in ST362 in FIG. 8), the control analysis device 4 determines the extraction conditions. Without executing step ST364, map image generation step ST365, and matching image generation step ST366, an index output step ST363 is executed to output less than a majority as an index representing a tendency related to the occurrence of defects.
  • No. 1 ⁇ No. Defects are detected in each of the 11 sheet-like products S2 of 11, and whether or not each defect was also detected in the original fabric inspection step ST31 (whether or not position information A and position information B match), and the original Let us consider a case in which the type of defect and the execution location of the original fabric inspection process ST31 are as shown in Table 3 below when a defect is also detected in the non-inspection process ST31.
  • Table 3 No. 1 in which no defects were detected in the original fabric inspection step ST31. 2.No. 3.No. 5, No. 7.No. 9 and no. Combinations of the types of defects present in the original fabric S1 corresponding to the sheet-like products S2 excluding No.
  • the control analysis device 4 executes the index output procedure ST363 without executing the extraction condition determination procedure ST364, the map image generation procedure ST365, and the matching image generation procedure ST366, and determines that the number of defects is less than the majority. It will be output as an index representing trends related to.
  • the extraction condition determination step ST364 the map image generation step Without executing ST365 and matching image generation procedure ST366, in index output procedure ST363, the number of combinations of the plurality of types and execution locations specified in combination identification procedure ST361 (in other words, the number of combinations of the plurality of types and execution locations (in other words, sheet-like product S2 and original fabric S1)
  • the number of defects considered to be the same in both sheets S2 is less than the majority of the total number of defects detected in the plurality of sheet products S2 will be output as an indicator representing a tendency related to the occurrence of defects. . Therefore, without needlessly generating map images and matching images, it is possible to accurately understand that a defect is detected in the sheet product S2 but not in the original fabric S1, and to contribute to the investigation of the cause. be.
  • FIG. 9 is a flow diagram schematically showing the steps of the analysis method according to the fourth embodiment.
  • the analysis method according to the fourth embodiment includes an original fabric inspection process ST41, a position information marking process ST42, a linking process ST43, a sheet-like product inspection process ST44, a reading process ST45, analysis step ST46.
  • the analysis method according to the fourth embodiment is also a method that utilizes the inspection results of the original fabric S1, similarly to the analysis method according to the second embodiment.
  • the contents of the original fabric inspection process ST41, position information marking process ST42, sheet-like product inspection process ST44, and reading process ST45 of the analysis method according to the fourth embodiment are respectively the original fabric inspection process ST21 of the analysis method according to the second embodiment.
  • Linking process ST43 In the linking process ST43, the control analysis device 4 uses the position information of the defects present in the raw fabric S1 before cutting, which were detected in the raw fabric inspection process ST41, the types of defects present in the raw fabric S1, and the raw fabric.
  • the execution location of the original fabric inspection step ST41 in which a defect present in S1 was detected (the location of the inspection device 1) and the manufacturing process history information of the original fabric S1 are stored in association with the position information mark M.
  • the defects present in the original fabric S1 detected in the original fabric inspection process ST41 are In addition to the position information, the type of defects present in the original fabric S1, and the execution location of the original fabric inspection step ST41 where the defects present in the original fabric S1 were detected, the manufacturing process history information of the original fabric S1 is included in the location information. It is stored in association with mark M. Examples of the manufacturing process history information include values of various parameters set or measured in the manufacturing processes of the original fabric S1, such as the stretching process, adhesion process, and bonding process.
  • the position information acquired in the reading step ST45 is input to the control analysis device 4, and the control analysis device 4 acquires the information in the reading step ST45. Based on the positional information on the original fabric S1 of the sheet-like product S2, the positional information on the original fabric S1 of the defect existing in the sheet-like product S2 detected in the sheet-like product inspection step ST44 is specified. Then, in the analysis step ST46, similarly to the analysis step ST26 of the analysis method according to the second embodiment, the control analysis device 4 is used to calculate the position information in the original fabric S1 of the defect existing in the identified sheet-like product S2. Based on this, trends related to the occurrence of defects are analyzed. However, the analysis step ST46 differs from the analysis step ST26 in the following points.
  • the position information on the original fabric S1 of the defect existing in the identified sheet-like product S2 and the original fabric stored in association with the position information mark M read in the reading step ST45 are used. It is determined whether or not the position information of the defect existing in S1 matches. This point is similar to the combination identification procedure ST261 of the analysis step ST26 of the analysis method according to the second embodiment. However, in the combination identification procedure ST461 of the fourth embodiment, when the control analysis device 4 matches, the defect existing in the original fabric S1 is stored in association with the position information mark M read in the reading step ST45.
  • the combination identification procedure of the second embodiment is that the combination of the type, the execution location of the original fabric inspection step ST41 where a defect existing in the original fabric S1 was detected, and the manufacturing process history information of the original fabric S1 is specified. Different from ST261. That is, in the combination specifying procedure ST461, unlike the combination specifying procedure ST261, the manufacturing process history information of the original fabric S1 is also specified.
  • the control analysis device 4 determines the type and execution specified in the combination identification step ST461. Determine extraction conditions based on the combination of locations. That is, when determining the extraction conditions, the manufacturing process history information of the identified raw fabric S1 is not used. Then, in the index output procedure ST463 of the analysis step ST46, the control analysis device 4 selects the extraction conditions determined in the extraction condition determination step ST462 from among all the defects present in the original fabric S1 detected in the original fabric inspection process ST41.
  • the manufacturing process history information specified in step ST461 for specifying the combination of the original fabric S1 containing defects that conform to the above is output as an index representing a tendency related to the occurrence of defects.
  • the combination of "foreign matter” and “stretching process” is determined as the extraction condition, and the manufacturing process history information about the raw fabric S1 that has defects that meet this extraction condition is It will be output as an index representing the tendency related to the occurrence of defects.
  • the manufacturing process history of the original fabric S1 in which a defect matching the extraction condition exists among all the defects present in the original fabric S1 is determined.
  • Information is output as an index representing a trend related to the occurrence of defects. Therefore, it can be expected that trends in the manufacturing process history information of the original fabric S1 in which defects may be detected in the sheet-like product S2 can be appropriately grasped.
  • FIG. 10 is a flow diagram schematically showing the steps of the analysis method according to the fifth embodiment.
  • the analysis method according to the fifth embodiment includes an original fabric inspection process ST51, a position information marking process ST52, a linking process ST53, a sheet-like product inspection process ST54, a reading process ST55, analysis step ST56.
  • the analysis method according to the fifth embodiment is also a method that utilizes the inspection results of the original fabric S1, similarly to the analysis method according to the second embodiment.
  • the contents of the original fabric inspection process ST51, position information marking process ST52, sheet-like product inspection process ST54, and reading process ST55 of the analysis method according to the fifth embodiment are respectively the original fabric inspection process ST21 of the analysis method according to the second embodiment.
  • Linking process ST53 In the linking process ST53, the control analysis device 4 uses the position information of the defects present in the original fabric S1 before cutting, which were detected in the original fabric inspection process ST51, the types of defects present in the original fabric S1, and the original fabric.
  • the execution location of the original fabric inspection step ST51 in which a defect present in S1 was detected (the location of the inspection device 1) and the manufacturing process history information of the original fabric S1 are stored in association with the position information mark M.
  • the defects present in the original fabric S1 detected in the original fabric inspection process ST51 are In addition to the position information, the type of defects present in the original fabric S1, and the execution location of the original fabric inspection step ST51 where the defects present in the original fabric S1 were detected, the manufacturing process history information of the original fabric S1 is included in the position information. It is stored in association with mark M. This point is similar to the linking step ST43 of the analysis method according to the fourth embodiment.
  • the position information acquired in the reading step ST55 is input to the control analysis device 4, and the control analysis device 4 acquires it in the reading step ST55. Based on the positional information on the original fabric S1 of the sheet-like product S2, the positional information on the original fabric S1 of the defect existing in the sheet-like product S2 detected in the sheet-like product inspection step ST54 is specified. Then, in the analysis step ST56, similarly to the analysis step ST26 of the analysis method according to the second embodiment, the control analysis device 4 is used to calculate the position information in the original fabric S1 of the defect existing in the identified sheet-like product S2. Based on this, trends related to the occurrence of defects are analyzed. However, the analysis step ST56 differs from the analysis step ST26 in the following points.
  • the position information on the original fabric S1 of the defect existing in the identified sheet-like product S2 and the original fabric stored in association with the position information mark M read in the reading step ST55 are used. It is determined whether or not the position information of the defect existing in S1 matches. This point is similar to the combination identification procedure ST261 of the analysis step ST26 of the analysis method according to the second embodiment. However, in the combination identification procedure ST561 of the fifth embodiment, when the control analysis device 4 matches, the defect existing in the original fabric S1 is stored in association with the position information mark M read in the reading step ST55.
  • the combination identification procedure of the second embodiment is that the combination of the type, the execution location of the original fabric inspection step ST51 where a defect existing in the original fabric S1 was detected, and the manufacturing process history information of the original fabric S1 is specified. Different from ST261. That is, in the combination specifying procedure ST561, unlike the combination specifying procedure ST261, the manufacturing process history information of the original fabric S1 is also specified. This point is similar to the analysis method combination identification procedure ST461 according to the fourth embodiment.
  • the control analysis device 4 determines the type and execution specified in the combination identification step ST561. Determine extraction conditions based on the combination of locations. That is, when determining the extraction conditions, the manufacturing process history information of the identified raw fabric S1 is not used.
  • the control analysis device 4 performs a determination step ST563 in which a learning model is used to determine a tendency related to the occurrence of defects. have The learning model is stored in the control analysis device 4 in an updatable manner.
  • the learning model may be generated by supervised learning using a combination of known inputs and outputs as training data, or may be generated by unsupervised learning. As the learning model, any known learning model such as a neural network or support vector machine may be used as appropriate.
  • the determination step ST563 among all the defects present in the original fabric S1 detected in the original fabric inspection step ST51, those that have defects that match the extraction conditions determined in the extraction condition determination step ST562 are selected.
  • the manufacturing process history information identified in the combination identification step ST561 of the fabric S1 is input to the learning model, and it is determined whether or not defects existing in the fabric S1 that meet the extraction conditions are detected as defects in the sheet product S2.
  • the determination result (probability of detection, etc.) is output from the learning model.
  • the determination step ST563 out of all the defects present in the original fabric S1, the original fabric S1 having a defect that meets the extraction condition is identified.
  • the manufacturing process history information obtained is input into the learning model stored in the control analysis device 4, and it is determined whether or not defects present in the original fabric S1 that meet the extraction conditions are detected as defects in the sheet-like product S2.
  • the determination result is output from the learning model. Therefore, it is possible to evaluate (determine) the possibility that a defect will be detected (occurred) in the sheet-like product S2 according to the manufacturing process history information of the original fabric S1.
  • a large sheet-like intermediate (a sheet-like product larger in size than the plurality of sheet-like products S2) is It is also possible to apply this method to the case where a plurality of sheet-like products S2 are manufactured by cutting out an intermediate (intermediate) and cutting this intermediate. Note that the intermediate body may be cut while being transported by a conveyor R3 or the like, or may be cut while being placed on a predetermined cutting table.
  • the present invention is not limited to this, and the intermediate material cut out from the original fabric S1 is It is also possible to mark the position information mark M on the body or on the plurality of sheet-like products S2 after cutting.
  • FIG. 11 is a diagram showing a schematic configuration of a manufacturing apparatus for executing an analysis method according to a modification of the present invention (a configuration in which position information marks M are marked on a plurality of sheet-like products S2 after cutting).
  • the same components as those in the manufacturing apparatus 100a shown in FIG. 1 are given the same reference numerals.
  • the marking device 2 is arranged on the conveyor R3 downstream of the cutting device 3 (downstream in the conveying direction of the sheet-like product S2), and Before collection, the marking device 2 marks the sheet-like product S2 with a position information mark M.
  • control analysis device 4 calculates the separation distance L2' in the X direction between the feed roller R1 and the marking device 2, and the conveyance of the original fabric S1 measured by an encoder (not shown) attached to the nip roller R2.
  • a predetermined portion of the original web S1 is cut into sheet-like products S2 based on the conveyance amount of the sheet-like products S2 measured by an encoder (not shown) attached to a roller included in the conveyor R3. After that, the timing of reaching the marking device 2 can be calculated.
  • the control analysis device 4 determines a part having predetermined XY coordinates based on the planned cutting line CL of the original fabric S1 (the cutting line that becomes the edge of the sheet-like product S2 after cutting into the sheet-like product S2).
  • the marking device 2 transparent ink is ejected from the nozzle of the marking device 2 corresponding to the Y coordinate of the predetermined XY coordinates.
  • the marking device 2 is controlled so as to inject.
  • position information marks M are marked within each rectangle (inside the sheet-like product S2).
  • an identification indicating the position of the original fabric S1 in the transport direction is attached to the width direction (Y direction) end (the edge not used as the sheet-like product S2) of the original fabric S1.
  • Information is recorded and, for example, this identification information is read by a predetermined reading device placed immediately before the cutting device 3 and input into the control analysis device 4.
  • the control analysis device 4 determines the cut sheet product S2 based on the read identification information and the conveyance amount of the sheet product S2 measured by an encoder (not shown) attached to a roller included in the conveyor R3.
  • the control analysis device 4 can also control the marking device 2 to eject transparent ink from the nozzle of the marking device 2 corresponding to the Y coordinate of the predetermined XY coordinates at the above timing.
  • the position information marking process ST1, the cutting process ST2, and the collecting process ST3 are executed in a single manufacturing apparatus 100a (in other words, a single manufacturing line) has been described as an example.
  • the present invention is not limited to this.
  • the original fabric S1 marked with the positional information mark M is once wound up on a take-up roller (not shown), transported to another production line, and then subjected to the cutting step ST2 and It is also possible to adopt a mode in which the recovery step ST3 is executed.
  • the present invention is not limited to this, and it is also possible to adopt a mode in which the position information marking process is performed first, and then the original fabric inspection process is performed.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Textile Engineering (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Biochemistry (AREA)
  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Immunology (AREA)
  • Pathology (AREA)
  • Optics & Photonics (AREA)
  • Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
  • Length Measuring Devices By Optical Means (AREA)
  • Polarising Elements (AREA)

Abstract

Le problème à résoudre par la présente invention est de fournir un procédé d'analyse de tendance d'occurrence de défaut apte à examiner de manière adéquate des causes d'occurrences de défaut dans un produit de type feuille, et apte à contribuer de manière adéquate à une amélioration du rendement du produit de type feuille. La solution de la présente invention porte sur un procédé d'analyse d'une tendance associée à des occurrences de défauts en fonction de défauts se produisant dans une pluralité de produits de type feuille (S2) fabriqués, par exemple, par découpe d'un matériau d'origine allongé (S1), le procédé comprenant : une étape ST11 consistant à marquer des marques d'informations de position (M), représentant des informations de position dans une direction longitudinale et une direction de largeur du matériau d'origine, sur le matériau d'origine (S1) avant la coupe, par exemple, de telle sorte que les marques d'informations de position (M) soient présentes sur chaque produit de la pluralité de produits de type feuille ; une étape ST12 consistant à inspecter les produits de type feuille ; une étape de lecture ST13 consistant à acquérir des informations de position des produits de type feuille dans le matériau d'origine, par la lecture des marques d'informations de position marquées sur les produits de type feuille ; et une étape ST14 consistant à identifier les informations de position dans le matériau d'origine de défauts présents dans les produits de type feuille, et à analyser une tendance associée à l'occurrence de défauts en fonction des informations de position identifiées.
PCT/JP2023/001281 2022-03-30 2023-01-18 Procédé d'analyse de tendance d'occurrence de défaut WO2023188711A1 (fr)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202380030286.5A CN118871775A (zh) 2022-03-30 2023-01-18 缺陷的产生倾向分析方法

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2022-056736 2022-03-30
JP2022056736A JP2023148609A (ja) 2022-03-30 2022-03-30 欠点の発生傾向解析方法

Publications (1)

Publication Number Publication Date
WO2023188711A1 true WO2023188711A1 (fr) 2023-10-05

Family

ID=88200258

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2023/001281 WO2023188711A1 (fr) 2022-03-30 2023-01-18 Procédé d'analyse de tendance d'occurrence de défaut

Country Status (4)

Country Link
JP (1) JP2023148609A (fr)
CN (1) CN118871775A (fr)
TW (1) TW202405411A (fr)
WO (1) WO2023188711A1 (fr)

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH08285558A (ja) * 1995-04-17 1996-11-01 Ricoh Co Ltd 表面検査装置
JP2002258263A (ja) * 2001-02-28 2002-09-11 Toray Ind Inc デバイス基板製造におけるデータ管理システム及びこれを用いた製造方法
JP2006189402A (ja) * 2005-01-07 2006-07-20 Hamamatsu Photonics Kk 透明体検査装置および透明体検査方法
JP2007298505A (ja) * 2006-04-05 2007-11-15 Hitachi High-Technologies Corp 欠陥検査方法およびその装置
JP2009294645A (ja) * 2008-05-07 2009-12-17 Nitto Denko Corp 光学表示装置製造システムに適用される製造管理システム及び製造管理方法
JP2013522595A (ja) * 2010-03-10 2013-06-13 スリーエム イノベイティブ プロパティズ カンパニー ウェブ製造プロセスにおける用途固有の繰り返し欠陥検出
US20150292155A1 (en) * 2014-04-15 2015-10-15 Georgia-Pacific Consumer Products Lp Methods and apparatuses for controlling a manufacturing line used to convert a paper web into paper products by reading marks on the paper web
JP2019212393A (ja) * 2018-05-31 2019-12-12 住友化学株式会社 電子デバイスの製造方法及びインクジェット印刷装置の駆動方法

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH08285558A (ja) * 1995-04-17 1996-11-01 Ricoh Co Ltd 表面検査装置
JP2002258263A (ja) * 2001-02-28 2002-09-11 Toray Ind Inc デバイス基板製造におけるデータ管理システム及びこれを用いた製造方法
JP2006189402A (ja) * 2005-01-07 2006-07-20 Hamamatsu Photonics Kk 透明体検査装置および透明体検査方法
JP2007298505A (ja) * 2006-04-05 2007-11-15 Hitachi High-Technologies Corp 欠陥検査方法およびその装置
JP2009294645A (ja) * 2008-05-07 2009-12-17 Nitto Denko Corp 光学表示装置製造システムに適用される製造管理システム及び製造管理方法
JP2013522595A (ja) * 2010-03-10 2013-06-13 スリーエム イノベイティブ プロパティズ カンパニー ウェブ製造プロセスにおける用途固有の繰り返し欠陥検出
US20150292155A1 (en) * 2014-04-15 2015-10-15 Georgia-Pacific Consumer Products Lp Methods and apparatuses for controlling a manufacturing line used to convert a paper web into paper products by reading marks on the paper web
JP2019212393A (ja) * 2018-05-31 2019-12-12 住友化学株式会社 電子デバイスの製造方法及びインクジェット印刷装置の駆動方法

Also Published As

Publication number Publication date
TW202405411A (zh) 2024-02-01
JP2023148609A (ja) 2023-10-13
CN118871775A (zh) 2024-10-29

Similar Documents

Publication Publication Date Title
CN204957003U (zh) 用于保证折叠盒质量的装置
JP2009244064A5 (fr)
CN105882142A (zh) 印刷品缺陷检测系统及其方法
JP2009133741A (ja) 検査群データ管理システム
CN108535259B (zh) 缺陷标注方法及装置、卷料和片的制造方法及卷料和片
WO2023188711A1 (fr) Procédé d'analyse de tendance d'occurrence de défaut
US20230330780A1 (en) Method and apparatus for automated quality control for cutting machines of flexible material parts
US11981520B2 (en) Apparatus and method for web twist defect correction
US20220318985A1 (en) Training data generation device, inspection device and program
JP2007196129A (ja) 検査情報管理システム
JP2007071789A (ja) 栗の品質検査方法
CN204882397U (zh) 不停机自动标识纺织面料疵点的系统
CN107618283A (zh) 基于视觉传感器的激光打标打码方法
WO2022107756A1 (fr) Procédé de fabrication de produit de type feuille, produit de type feuille, tissu original et corps intermédiaire
JPH10297063A (ja) 印刷物の品質検査選別方法
KR101872934B1 (ko) 직기정대 및 제직불량 위치정보로 편성되는 롤맵을 이용한 정단위치 결정방법
EP4400832A1 (fr) Système de classification de défauts
CN219285035U (zh) 一种用于半固化片外观缺陷的甄别剔除装置
US20240317444A1 (en) A method for assessing quality of a transversal sealing of a food package
EP4327791A1 (fr) Système et procédé d'inspection de contrôle de qualité de saillies unitaires dans un substrat
CN110775374B (zh) 检查装置、ptp包装机和ptp片的制造方法
KR20090112421A (ko) 미립 품위 판별 시스템 및 그 방법
JP4128572B6 (ja) Ｘ線検査システム
JP4128572B2 (ja) Ｘ線検査方法、システムおよび装置
Gawde Fabric Inspection in Folding: Current Scenario in India.

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23778750

Country of ref document: EP

Kind code of ref document: A1