Information processing system, information processing method, and information processing program

Abstract

Provided are an information processing system, an information processing method, and an information processing program capable of predicting the state of defects present in a laminated body comprising a plurality of webs. The information processing system acquires pre-lamination feature point information relating to feature points present in each of a first web and a second web, estimates post-lamination feature point information relating to feature points present in a laminated body including the first web and the second web on the basis of the acquired pre-lamination feature point information, and generates output information on the basis of the estimated post-lamination feature point information.

Description

Information processing system, information processing method, and information processing program

[0001] The present invention relates to an information processing system, an information processing method, and an information processing program.

[0002] Optical films are used in displays such as liquid crystal display devices. A web of the optical film or the like is composed of a laminate of multiple webs. In the manufacturing process of such optical films or the like, strict control is required for defects that occur in the web. For example, Patent Document 1 describes a technique related to defects that occur in the web.

[0003] Japanese Patent Application Laid-Open No. 2023-148609

[0004] Defects in a stack of multiple webs tend to be more complex than defects in a single web. Therefore, it is difficult to predict the defect state of a stack of these webs from the defect state of each single web. Therefore, it is desirable to be able to predict the defect state of a stack of multiple webs.

[0005] The present invention has been made in consideration of the above circumstances, and aims to provide an information processing system, an information processing method, and an information processing program that are capable of predicting the state of defects present in a stack of multiple webs.

[0006] The above object of the present invention can be achieved by the following means.

[0007] (1) An information processing system including: an acquisition unit that acquires pre-stacking feature point information regarding feature points present in each of a first web and a second web; an estimation unit that estimates post-stacking feature point information regarding feature points present in a laminate including the first web and the second web based on the acquired pre-stacking feature point information; and a generation unit that generates output information based on the estimated post-stacking feature point information.

[0008] (2) The information processing system described in (1) above, wherein the pre-stacking feature information includes information regarding feature points present on the first web of each of the first production lot and the second production lot and information regarding feature points present on the second web of each of the third production lot and the fourth production lot, and the estimation unit estimates the post-stacking feature information for multiple combinations of the first production lot and the second production lot and the third production lot and the fourth production lot.

[0009] (3) The information processing system described in (2) above, wherein the output information includes information regarding whether or not the characteristic points present in the laminate satisfy a predetermined standard for multiple combinations of the first production lot and the second production lot and the third production lot and the fourth production lot.

[0010] (4) The information processing system described in (2) above, wherein the output information includes information regarding combinations of the first production lot and the second production lot with the third production lot and the fourth production lot, in which the feature points present in the laminate satisfy a predetermined standard.

[0011] (5) The information processing system described in (2) above, wherein the output information includes information regarding combinations of the first production lot and the second production lot with the third production lot and the fourth production lot in which the characteristic points present in the laminate deviate from a predetermined standard, among multiple combinations of the first production lot and the second production lot and the third production lot and the fourth production lot.

[0012] (6) The information processing system described in (1) above, wherein the estimation unit estimates the post-stacking feature point information using a data set in which feature points present on each of the first database web and the second database web are associated with feature points present on the database stack of the first database web and the second database web.

[0013] (7) The information processing system according to (6), wherein the estimation unit estimates the post-stacking feature point information using a machine learning model trained on the dataset.

[0014] (8) An information processing system described in (6) above, in which the feature points present in the database laminate are classified into second-type feature points that are not present in the first database web and the second database web but appear in the database laminate, and third-type feature points that are present in the first database web or the second database web and remain in the database laminate.

[0015] (9) The information processing system according to (1) above, further comprising an output unit that outputs the generated output information.

[0016] (10) The pre-stacking feature information further includes information regarding feature points present in a third web, and the laminate is composed of the first web, the second web, and the third web, in the information processing system described in (1) above.

[0017] (11) An information processing method including: acquiring pre-stacking feature point information regarding feature points present in each of a first web and a second web; estimating post-stacking feature point information regarding feature points present in a laminate including the first web and the second web based on the acquired pre-stacking feature point information; and generating output information based on the estimated post-stacking feature point information.

[0018] (12) An information processing program that causes a computer to execute the information processing method described in (11) above.

[0019] In the information processing system, information processing method, and information processing program according to the present invention, post-stacking feature point information is estimated from pre-stacking feature point information. Output information is then generated based on this post-stacking feature point information. The pre-stacking feature point information is information about feature points present in each of the first web and the second web. The post-stacking feature point information is information about feature points present in a stack of the first web and the second web. In other words, the state of feature points present in the stack of the first web and the second web can be predicted from data on feature points present in each single layer of the first web and the second web. This makes it possible to predict the state of defects present in a stack of multiple webs.

[0020] Advantages and features provided by one or more embodiments of the present invention will be more fully understood from the following detailed description and the accompanying drawings, which are intended for purposes of illustration only and are not intended to define limitations of the present invention.

[0023] FIG. 1 is a schematic diagram illustrating an application example of an information processing system according to a first embodiment.

[0024] FIG. 2 is a cross-sectional view illustrating an example of the configuration of the laminate shown in FIG. 1.

[0025] FIG. 3 is a schematic diagram illustrating an example of the configuration of the inspection device shown in FIG. 1.

[0026] FIG. 3A is another schematic diagram illustrating the configuration of the inspection device shown in FIG. 3A.

[0027] FIG. 3B is a block diagram illustrating the general configuration of the terminal device shown in FIG. 1.

[0028] FIG. 4 is a table for explaining classification of feature points present in the laminate shown in FIG. 2 and feature points present in each web before lamination.

[0029] FIG. 5 is a block diagram illustrating the general configuration of the information processing system shown in FIG. 6.

[0030] FIG. 6 is an example of a user list stored in the memory unit shown in FIG. 7.

[0031] FIG. 7 is an example of a lot list stored in the memory unit shown in FIG. 7.

[0032] FIG. 8 is an example of an inspection data DB stored in the memory unit shown in FIG. 8.

[0033] FIG. 9 is another example of the inspection data DB stored in the memory unit shown in FIG. 9.

[0034] FIG. 10 is another example of the inspection data DB stored in the memory unit shown in FIG. 10.

[0035] 10 is a flowchart showing an example of a comparison process executed by the information processing system shown in FIG. 1; FIG. 11 is a subroutine flowchart of the process of step S33 shown in FIG. 9; FIG. 12 is a diagram showing an example of output information output by the information processing system shown in FIG. 1; FIG. 13 is a flowchart showing an example of an output information generation process executed by the information processing system shown in FIG. 1; FIG. 14 is a flowchart showing a machine learning method for a machine learning model used by the information processing system shown in FIG. 1; FIG. 15 is a subroutine flowchart executed by an information processing system according to a second embodiment; FIG. 16 is a diagram showing an example of a probability density function calculated by kernel density estimation shown in FIG. 14; FIG. 17 is a schematic diagram showing an application example of an information processing system according to a third embodiment; FIG. 18 is a schematic diagram showing an example of a manufacturing apparatus shown in FIG. 1;

[0021] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. However, the scope of the present invention is not limited to the disclosed embodiments. In the description of the drawings, the same elements are denoted by the same reference numerals, and duplicate explanations will be omitted. Furthermore, the dimensional proportions in the drawings are exaggerated for the convenience of explanation and may differ from the actual proportions.

[0022] First Embodiment FIG. 1 is a schematic diagram showing an application example of an information processing system 50 according to a first embodiment. The information processing system 50 is configured by, for example, a server. The information processing system 50 is connected to a terminal device 70 in a factory 100 for communication with each other via a network. The network is a communication line such as a data communication network. Some networks may use a wired LAN or a wireless LAN. The wireless LAN is, for example, a LAN conforming to the IEEE 802.11 standard. The information processing system 50 may be connected to another factory via the network. In the factory 100, a manufacturing apparatus 2000 manufactures a laminate 80. The laminate 80 is, for example, wound into a roll.

[0023] FIG. 2 shows an example of a cross-sectional configuration of a laminate 80. The laminate 80 includes a first web 81, a second web 82, and a third web 83. In the laminate 80, the first web 81, the second web 82, and the third web 83 are laminated in this order. An adhesive layer 801 is provided between the first web 81 and the second web 82. An adhesive layer 802 is provided between the second web 82 and the third web 83. The first web 81 and the third web 83 are protective films such as TAC films. TAC is an abbreviation for triacetyl cellulose. The second web 82 is an optically functional film such as a polarizer. The adhesive layer 801 bonds the first web 81 and the second web 82 together. The adhesive layer 802 bonds the second web 82 and the third web 83 together. In the laminate 80, a pressure-sensitive adhesive layer may be provided instead of the adhesive layers 801 and 802.

[0024] [Factory 100] In the factory 100, the first web 81, the second web 82, and the third web 83 are stacked to produce the laminate 80. The first web 81, the second web 82, and the third web 83 may each be produced in the factory 100 or in another factory. The width of the laminate 80 is, for example, in the range of 1,000 mm to 5,000 mm. The thickness of each of the first web 81, the second web 82, and the third web 83 is set to, for example, in the range of 15 μm to 500 μm, taking into consideration quality, handling, and the like. When the first web 81, the second web 82, or the third web 83 contains a metal such as steel, the thickness of the first web 81, the second web 82, or the third web 83 may be 1 mm or more. The length of the roll-shaped laminate 80 is, for example, in the range of 2,000 m to 10,000 m.

[0025] In the factory 100, the laminate 80 is inspected. For example, the surface of the laminate 80 is inspected by an inspection device 90. The inspection device 90 includes, for example, a camera. Inspection data for the laminate 80 is generated by optical inspection by the inspection device 90. This inspection data includes post-stacking feature point information for a database regarding feature points present in the laminate 80. Here, the laminate 80 used to generate the post-stacking feature point information for a database corresponds to a specific example of a database laminate of the present invention. The first web 81 and the second web 82 included in this laminate 80 correspond to a specific example of a database first web and a database second web of the present invention.

[0026] Here, the feature points present in the laminate 80 are optical feature points present in the web, specifically, spots that are optically different from their surroundings. Feature points present in the web may also be referred to as web defects, malfunctions, or failures. Feature points include, for example, defects caused by poor adhesion when bonding multiple webs together and defects caused by axial unevenness. For example, ultrasonic welding is used to bond multiple webs together. For example, tens to tens of thousands of feature points can be detected from image data captured of a single laminate 80. The total length of the laminate 80 is, for example, several hundred meters to several kilometers.

[0027] The post-stacking feature point information for the database includes, for example, information regarding the position, size, etc. of each of the multiple feature points present in the laminate 80. The position of the feature point can be expressed using, for example, XY coordinates. In the post-stacking feature point information for the database, multiple adjacent feature points may be clustered.

[0028] The inspection device 90 that detects feature points present in the laminate 80 is a transmission type or reflection type inspection device. The transmission type inspection device 90 irradiates the laminate 80 with light and receives the light that has passed through the laminate 80. The reflection type inspection device 90 irradiates the laminate 80 with light and receives the light that has been reflected by the laminate 80. The transmission type inspection device and the reflection type inspection device each include a bright-field type inspection device and a dark-field type inspection device. A plurality of inspection devices 90 may be installed in the factory 100.

[0029] 3A and 3B show an example of the configuration of a reflective inspection device 90. Fig. 3A shows the configuration of this inspection device 90 as viewed from the width direction of the laminate 80. Fig. 3B shows the configuration of this inspection device 90 as viewed from the conveyance direction of the laminate 80. The inspection device 90 includes, for example, a light source 91, a camera 92, an analysis unit 93, and a storage unit 94.

[0030] In the inspection device 90, cameras 92 capture images of the surface of the laminate 80 and generate image data. The number of cameras 92, the angle of view, the distance to the surface of the laminate 80, etc. are set appropriately so that the entire width of the laminate 80 is captured. Fig. 3B shows an inspection device 90 having two cameras 92 in the width direction.

[0031] The light source 91 irradiates light onto an inspection area of ​​the laminate 80. The light source 91 irradiates light uniformly in the width direction of the roll-shaped laminate 80. Here, irradiating light uniformly means that the illuminance of the irradiated light is approximately the same.

[0032] The camera 92 is an optical sensor that optically reads the inspection area of ​​the laminate 80. The camera 92 has an imaging element such as a CCD or a CMOS, a lens, etc. CCD is an abbreviation for Charge Coupled Device. CMOS is an abbreviation for Complementary Metal Oxide Semiconductor. The camera 92 generates two-dimensional image data from the output signals of each imaging element. The camera 92 detects diffused light among the light irradiated by the light source 91 and reflected by the surface of the laminate 80. The camera 92 may be a color camera or a monochrome camera. The camera 92 may detect light in the visible light range or infrared light range.

[0033] The camera 92 can read, for example, the entire widthwise range of the laminate 80 at once. The contrast of the image data captured by the camera 92 is preferably equal to or greater than a predetermined value. In other words, it is desirable that the area of ​​the laminate 80 that is irradiated with light from the light source 91 and the non-irradiated area have a contrast equal to or greater than a predetermined value. In order to generate image data that has a contrast equal to or greater than a predetermined value, it is desirable to use a light source 91 that is strong and has high linearity.

[0034] Here, "strong" means that, for example, when the illuminance at an irradiation distance of 50 mm is E50, the illuminance E50 is 50,000 lx or more. Also, "highly directional" means, for example, when the illuminance at an irradiation distance of 50 mm is E50 and the illuminance at an irradiation distance of 100 mm is E100, the relationship (E50-E100) / E50<0.5 is satisfied.

[0035] The camera 92 is disposed, for example, at a position where it receives specularly reflected light of the light emitted from the light source 91. The camera 92 may also be disposed at a position where it avoids specularly reflected light of the light emitted from the light source 91, that is, at a position where it receives diffused light from the laminate 80.

[0036] The analysis unit 93 is composed of a CPU, RAM, etc. The analysis unit 93 reads out various processing programs stored in the storage unit 94, loads them into the RAM, and performs various processes in cooperation with the programs. The analysis unit 93 processes image data captured by the camera 92 to detect feature points present in the laminate 80. For example, the analysis unit 93 performs predetermined image processing on the image data, detects feature points, and quantitatively evaluates each of the detected feature points. For example, known techniques are used for image analysis. Specifically, pixels whose pixel values ​​in the image deviate by a predetermined amount or more from the average value of the surrounding pixels are extracted as feature points. The analysis unit 93 may detect feature points using the following method.

[0037] The analysis unit 93 divides the image data captured by the camera 92 into multiple regions. For example, the analysis unit 93 divides the image data into n regions in the width direction. n may be, for example, several to several tens of regions. Hereinafter, the n regions will be referred to as region a1 to region an.

[0038] Next, the analysis unit 93 acquires image data of one region a1 and performs mathematical processing on the image data of the region a1. The mathematical processing includes, for example, preprocessing, enhancement processing, signal processing, and image feature extraction.

[0039] Preprocessing includes, for example, image cropping, low-pass filtering, high-pass filtering, Gaussian filtering, median filtering, bilateral filtering, morphological transformation, color conversion, contrast adjustment, noise removal, restoration of blurred and blurred images, mask processing, Hough transform, and projective transformation, etc. Color conversion includes, for example, L*a*b*, sRGB, HSV, and HSL.

[0040] The enhancement process includes, for example, a Sobel filter, a Scharr filter, a Laplacian filter, a Gabor filter, and a Canny algorithm.

[0041] Signal processing includes, for example, basic statistics, square root of sum of squares, difference, sum, product, ratio, distance matrix calculation, differential and integral calculus, threshold processing, Fourier transform, wavelet transform, and peak detection. Basic statistics include, for example, maximum, minimum, mean, median, standard deviation, variance, and quartile. Threshold processing includes, for example, binarization and adaptive binarization. Peak detection includes, for example, detection of peak value, peak number, half-width, etc.

[0042] Image feature extraction includes template matching, SIFT features, and the like.

[0043] The analysis unit 93 performs mathematical processing on the image data of the region a1, and then performs threshold processing on the values ​​obtained by this processing. The threshold processing is a process of determining whether or not a point is a feature point based on a predetermined threshold, and determining the size of the feature point, etc.

[0044] The analysis unit 93 performs the same process on areas other than the area a1.

[0045] After processing each of the regions a1 to an, the analysis unit 93 integrates the results for each of the regions a1 to an. As a result, feature points present in each region of the stack 80 are detected in association with their positions, sizes, etc. The analysis unit 93 then stores the detection results of the feature points of the stack 80 in the storage unit 94.

[0046] The analysis unit 93 may, for example, combine multiple images obtained by continuous shooting with one camera 92. The analysis unit 93 may, for example, generate one image data of the entire surface of the laminate 80 and store it in the storage unit 94. Alternatively, the analysis unit 93 may generate multiple image data in association with the shooting time and store it in the storage unit 94. The analysis unit 93 may combine multiple image data obtained by multiple cameras 92 arranged in the width direction. The analysis unit 93 may, for example, determine the longitudinal position of the laminate 80 based on the shooting time by referring to the stored transport speed.

[0047] The storage unit 94 is composed of an HDD, an SSD, etc. SSD is an abbreviation for Solid State Drive. The storage unit 94 stores various processing programs and data necessary for executing the programs. For example, the storage unit 94 stores image data captured by the camera 92, linking it to the time of capture. The storage unit 94 stores manufacturing conditions such as the winding speed of the manufacturing apparatus 2000. The manufacturing conditions for the laminate 80 may be included in a process list of an inspection DB, which will be described later.

[0048] 3C shows an example of a transmission type inspection device 90. In this inspection device 90, a light source 91 is placed at a position facing a camera 92 with the laminate 80 therebetween.

[0049] The inspection device 90 may have multiple inspection units. For example, the camera 92 may include a camera for detecting scratches on the surface of the laminate 80 and a camera for detecting foreign matter inside the laminate 80. The inspection data may use some of the data from the multiple inspection units, or may use a combination of the results from the multiple inspection units.

[0050] [Configuration of Terminal Device 70] The terminal device 70 is a computer such as a PC, a smartphone, or a tablet terminal. PC is an abbreviation for Personal Computer. The terminal device 70 is configured to be connectable to the information processing system 50. The terminal device 70 transmits and receives various information to and from the information processing system 50. For example, the terminal device 70 is a PC used by an employee of a manufacturing company that operates the factory 100.

[0051] 4 is a block diagram showing a schematic configuration of the terminal device 70. Each terminal device 70 has a CPU 71, a ROM 72, a RAM 73, a storage 74, a communication interface 75, a display unit 76, and an operation reception unit 77. CPU is an abbreviation for Central Processing Unit. ROM is an abbreviation for Read Only Memory. RAM is an abbreviation for Random Access Memory. Each component is connected to each other via a bus 78 so as to be able to communicate with each other.

[0052] The CPU 71 controls the above components and performs various arithmetic processing in accordance with a program recorded in the ROM 72 or the storage 74 .

[0053] The ROM 72 stores various programs and various data.

[0054] The RAM 73 serves as a working area for temporarily storing programs and data.

[0055] The storage 74 stores various programs including an operating system and various data. For example, an application for displaying various information transmitted from the information processing system 50 is installed in the storage 74.

[0056] The communication interface 75 is an interface for communicating with other devices. A wired or wireless communication interface conforming to various standards is used as the communication interface 75. The communication interface 75 is used, for example, when transmitting inspection data from the inspection device 90 to the information processing system 50 and when receiving comparison information (described later) from the information processing system 50.

[0057] The display unit 76 includes, for example, a liquid crystal display or an organic EL display. The display unit 76 displays various information. The display unit 76 may be configured with viewer software, a printer, or the like.

[0058] The operation acceptance unit 77 has, for example, a touch sensor, a pointing device such as a mouse, a keyboard, etc. The operation acceptance unit 77 accepts various operations from the user. Note that the display unit 160 and the operation acceptance unit 77 may form a touch panel by superimposing a touch sensor serving as the operation acceptance unit 77 on the display surface serving as the display unit 76.

[0059] The terminal device 70 may generate the post-stacking feature point information for the database by performing image analysis on the image data of the laminate 80 captured by the inspection device 90. The terminal device 70 transmits, for example, inspection data including the post-stacking feature point information for the database to the information processing system 50.

[0060] The terminal device 70 further transmits, for example, inspection data for each of the first web 81 before lamination, the second web 82 before lamination, and the third web 83 before lamination to the information processing system 50. This inspection data includes pre-lamination feature point information for the database regarding feature points present in each of the first web 81 before lamination, the second web 82 before lamination, and the third web 83 before lamination. The feature points present in each of the first web 81 before lamination, the second web 82 before lamination, and the third web 83 before lamination are optically detected, for example, by an inspection device similar to the inspection device 90 described above.

[0061] The inspection data of each of the first web 81 before lamination, the second web 82 before lamination, and the third web 83 before lamination may be transmitted from another terminal device to the information processing system 50. The other terminal device is, for example, a terminal device in a factory where the first web 81, the second web 82, and the third web 83 are manufactured.

[0062] The information processing system 50, for example, compares the pre-stacking feature point information for database and the post-stacking feature point information for database received from the terminal device 70, and classifies each feature point into three types: first type feature points, second type feature points, and third type feature points.

[0063] 5 is a table for explaining the first, second, and third type characteristic points. The characteristic points present in each of the first web 81 before lamination, the second web 82 before lamination, and the third web 83 before lamination can be classified, for example, into the first type characteristic points and the third type characteristic points. The characteristic points present in the laminate 80 can be classified, for example, into the second type characteristic points and the third type characteristic points.

[0064] The first type characteristic points are characteristic points that are present only in the webs before lamination and are not present in the laminate 80. That is, the first type characteristic points are characteristic points that disappear during the lamination process of the webs. Even if these first type characteristic points are present in the first web 81 before lamination, the second web 82 before lamination, or the third web 83 before lamination, they are characteristic points that are relatively unlikely to affect the processes subsequent to the lamination process.

[0065] The second type characteristic points are characteristic points that do not exist in the web before lamination and that exist in the laminate 80. That is, the second type characteristic points are characteristic points that are newly generated in the lamination process. These second type characteristic points are characteristic points that result from the lamination process.

[0066] The third type feature points are feature points that exist in both the web before lamination and the laminate 80. That is, the third type feature points are feature points that are generated in the web before lamination and remain in the laminate 80. These third type feature points are feature points that are relatively likely to affect processes after the lamination process. Note that the shape and size of the third type feature points may change during the lamination process compared to when they existed in the web before lamination. Depending on the lamination process, additional faults may occur.

[0067] For example, the information processing system 50 classifies and stores the characteristic points present in each web before stacking and the characteristic points present in the laminate 80 as first type characteristic points, second type characteristic points, and third type characteristic points.

[0068] 6 is a block diagram showing a schematic configuration of the information processing system 50. The information processing system 50 includes, for example, a control unit 51, a storage unit 52, and a communication unit 53.

[0069] The control unit 51 has, for example, a CPU and memories such as RAM and ROM. The CPU is configured with a multi-core processor or the like that controls the above-mentioned units and executes various arithmetic processing in accordance with a program. Each function of the information processing system 50 is realized by the CPU executing the corresponding program. The specific functions of the control unit 51 will be described later. The communication unit 53 is an interface that connects to an external device such as a terminal device 70 via a network.

[0070] The memory unit 52 is a large-capacity auxiliary storage device that stores various programs including an operating system and various data. For example, a hard disk, a solid-state drive, a flash memory, or a ROM is used as the storage. For example, the memory unit 52 stores a user list, a lot list, an inspection data DB, and the like. For example, the manager of the factory 100 manages each of the user list and the lot list. The manager of the factory 100 is, for example, an employee of the manufacturer that operates the factory 100.

[0071] 7A shows an example of a user list. The user list includes information such as the user ID, user name, and contact information of each user. Each user may be assigned access rights to the test data DB, and each user may have access to various data related to a specific web site.

[0072] 7B shows an example of a lot list, which includes information such as the lot ID, product name, delivery destination user ID, manufacturing conditions, size, and manufacturing date for each lot.

[0073] 8A to 8F each show an example of the inspection data DB. The inspection data DB includes, for example, a process list for each manufacturing process of the laminate 80 and inspection data for each inspection performed in each manufacturing process.

[0074] 8A shows an example of a process list. The process list includes information such as the lot ID, the width and length of the web at each process, the process name, the stretch rate of the web, the number of layers of the web, the width direction and the length direction. The process list also includes, for example, information about the inspections performed at each manufacturing process. The information about each inspection includes, for example, the inspection ID of each inspection, the inspection device ID of the inspection device used for each inspection, the inspection data, and the inspection date and time. The information about the width direction and the length direction of the web is expressed, for example, by comparing the inspection of each web before stacking with the inspection of the laminate 80.

[0075] 8B, 8C, 8D, 8E, and 8F each show an example of inspection data for each inspection included in the process list. FIG. 8B shows an example of inspection data with an inspection ID of i0101. This inspection data is, for example, inspection data for the first web 81. FIG. 8C shows an example of inspection data with an inspection ID of i0102. This inspection data is, for example, inspection data for the second web 82. FIG. 8D shows an example of inspection data with an inspection ID of i0103. This inspection data is, for example, inspection data for the third web 83. These inspection data include, for example, information regarding the feature point IDs, positions, areas, lengths, widths, maximum brightness, minimum brightness, classifications, and the presence or absence of concentrated dots of feature points present in the first web 81, the second web 82, and the third web 83 before lamination. That is, these inspection data include pre-stacking feature point information for the database regarding feature points present in the first web 81, the second web 82, and the third web 83 before lamination.

[0076] The positions of the feature points are expressed, for example, by X and Y coordinates based on a predetermined position on each web. The X coordinate is, for example, the coordinate in the width direction of each web and can range from 0 to 3000 mm. The Y coordinate is, for example, the coordinate in the length direction of each web and can range from 0 to 10000 m.

[0077] The classification of feature points represents, for example, the shape and brightness distribution of each feature point. Each feature point is classified into, for example, about 5 to 20 classifications according to its shape and brightness distribution. The presence or absence of concentrated dots represents, for example, whether or not multiple feature points exist around the feature point.

[0078] 8E shows an example of inspection data with an inspection ID of i0104. This inspection data is, for example, inspection data for the laminate 80. This inspection data includes, for example, information about the feature point IDs, positions, areas, lengths, widths, maximum brightness, minimum brightness, divisions, and the presence or absence of concentrated dots of the feature points present in the laminate 80. In other words, this inspection data includes post-stacking feature point information for the database regarding the feature points present in the laminate 80.

[0079] 8F shows an example of a comparison result between the pre-stacking database feature point information and the post-stacking database feature point information. The information processing system 50, for example, integrates and compares feature points present in the first web 81 before stacking, the second web 82 before stacking, and the third web 83 before stacking with feature points present in the laminate 80. The information processing system 50, for example, assigns new feature point IDs to all of the integrated feature points. The information processing system 50, for example, classifies the feature points for each feature point ID into first, second, or third type feature points. The comparison result may include information regarding the accuracy of the classification of the first, second, and third type feature points.

[0080] 9 is a flowchart showing an example of a processing procedure for comparing pre-lamination feature point information for database with post-lamination feature point information for database, which is executed in the information processing system 50. The processing of the information processing system 50 shown in the flowchart in FIG. 9 is stored as a program in the storage unit 52 of the information processing system 50, and is executed by the CPU controlling each unit.

[0081] (Step S31) The information processing system 50 acquires inspection data for each web before stacking and inspection data for the laminate 80, for example, in response to instructions from a user via the terminal device 70. This inspection data includes, for example, inspection data for each of the first web 81 before stacking, the second web 82 before stacking, and the third web 83 before stacking, as well as inspection data for the laminate 80. This allows the information processing system 50 to acquire pre-stacking feature point information for the database regarding feature points present in each web before stacking, and post-stacking feature point information for the database regarding feature points present in the laminate 80. The information processing system 50 may acquire the inspection data for each web before stacking and the inspection data for the laminate 80 at a predetermined timing.

[0082] (Step S32) The information processing system 50 performs preprocessing on each piece of inspection data in order to align the coordinate systems of each web before stacking with the laminate 80. For example, the information processing system 50 aligns the XY coordinate system of each web before stacking with the XY coordinate system of the laminate 80.

[0083] For example, as preprocessing for each inspection data, the information processing system 50 inverts the Y coordinates of one or more webs before lamination. The information processing system 50 may also invert the X coordinates of one or more webs before lamination. The information processing system 50 may also convert the X and Y coordinates of one or more webs before lamination depending on the elongation rate of each web before lamination, etc.

[0084] The information processing system 50 further performs noise removal processing as preprocessing of each inspection data. The noise removal processing includes, for example, removal of low-intensity feature points, removal of extremely small feature points, and removal of continuous dots. The noise removal processing may also include removal of concentrated dots in the width direction. Concentrated dots in the width direction occur, for example, at the leading and trailing ends of each web and the laminate 80 before stacking.

[0085] (Step S33) After performing preprocessing on each piece of inspection data, the information processing system 50 performs alignment processing between the coordinate system of the first web 81 before lamination and the coordinate system of the laminate 80.

[0086] FIG. 10 is a subroutine flowchart showing the alignment process in step S33.

[0087] (Steps S401 to S403) The information processing system 50 first roughly adjusts the XY coordinates of the characteristic points of the first web 81 and the laminate 80 before lamination as follows. For example, the information processing system 50 first shifts the coordinate position of a predetermined characteristic point of the first web 81 before lamination by a predetermined amount. Next, the information processing system 50 calculates distances L1 to Lm between the predetermined characteristic point of the first web 81 before lamination and the corresponding characteristic point of the laminate 80, and selects the shift amount (x1, y1) whose sum is the smallest. The information processing system 50 may use an average value instead of the sum. For example, the information processing system 50 determines the characteristic point of the laminate 80 whose coordinate position is closest to the coordinate position of the predetermined characteristic point of the first web 81 before lamination as the characteristic point of the laminate 80 corresponding to the predetermined characteristic point of the first web 81 before lamination. If the information processing system 50 cannot determine a feature point of the laminate 80 that corresponds to a specified feature point of the first web 81 before lamination, i.e., if the specified feature point is a first-type feature point, it may exclude this feature point and calculate the sum of the distances L1 to Lm.

[0088] The information processing system 50 sequentially shifts the coordinate positions of predetermined feature points of the first web 81 before lamination from (-shift_x, -shift_y) to (+shift_x, +shift_y) around the central shift amount (0, 0) in increments of a fixed coarse adjustment shift amount a. From the coordinate positions of the predetermined feature points of the first web 81 before lamination, distances L1 to Lm to feature points 1 to m of the laminate 80 are calculated. Then, the shift amount (x1, y1) that minimizes the sum of distances L1 to Lm is selected from among (-shift_x, -shift_y) to (+shift_x, +shift_y).

[0089] For example, when the coarse adjustment shift amount a = 1.0 mm, (-shift_x, -shift_y) = (-10 mm, -10 mm), (-shift_x, +shift_y) = (+10 mm, +10 mm). The information processing system 50 may use different units for the X direction and the Y direction for the coarse adjustment shift amount a. For example, the X direction may be in millimeters and the Y direction in meters, so that (-shift_x, -shift_y) = (-10 mm, -10 m), (-shift_x, +shift_y) = (+10 mm, +10 m).

[0090] (Steps S404 to S406) Next, the information processing system 50 fine-tunes the X and Y coordinates of the characteristic points of each of the first web 81 and the laminate 80 before lamination, and selects a shift amount (x2, y2). The information processing system 50 selects the shift amount (x2, y2) in substantially the same manner as in steps S401 to S403 described above. Steps S404 to S406 differ from steps S401 to S403, for example, in the following respects. The fine adjustment shift amount b in step S404 is smaller than the coarse adjustment shift amount a. Furthermore, the shift amount (x1, y1) selected in step S403 is used as the center shift amount in step S405. For example, the fine adjustment shift amount b is sufficiently smaller than the coarse adjustment shift amount a, for example, 0.1 mm, which is one order of magnitude smaller.

[0091] (Step S407) The information processing system 50 performs coordinate transformation processing on all the feature points of the first web 81 before lamination, using the shift amount (x2, y2) selected in step S406.

[0092] (Step S408) The information processing system 50 calculates the distances L1 to Lm after the coordinate transformation in step S407 and checks whether the sum of the distances L1 to Lm is less than a predetermined threshold. If the sum is equal to or greater than the predetermined threshold, the information processing system 50 may determine that the coordinate transformation process in step S407 is inappropriate.

[0093] (Step S409) If the alignment is inappropriate, i.e., if the answer is YES, the information processing system 50 ends the process. If the alignment is inappropriate, the information processing system 50 may display an error message on the display unit 76 or record a message in the test data DB indicating that calculation is not possible. On the other hand, if the alignment is appropriate, i.e., if the answer is NO, the information processing system 50 ends the process of FIG. 10, returns to the process of FIG. 9, and executes the processes from step S34 onwards.

[0094] (Step S34) The information processing system 50 performs a process of aligning the coordinate system of the second web 82 before stacking with the coordinate system of the laminate 80. The information processing system 50 performs a process of aligning the coordinate system of the second web 82 before stacking with the coordinate system of the laminate 80, for example, in the same manner as described in step S33 above.

[0095] (Step S35) The information processing system 50 performs a process of aligning the coordinate system of the third web 83 before lamination with the coordinate system of the laminate 80. The information processing system 50 performs a process of aligning the coordinate system of the third web 83 before lamination with the coordinate system of the laminate 80, for example, in the same manner as described in step S33 above. The information processing system 50 may perform the processes of steps S33 to S35 in a different order, or may perform the processes of steps S33 to S35 simultaneously.

[0096] (Step S36) The information processing system 50 compares the pre-stacking feature point information with the post-stacking feature point information. Specifically, the information processing system 50 compares each feature point present in each web before stacking with each feature point present in the laminate 80. For example, the coordinates of the feature points of each web before stacking have been transformed by the processes of steps S33 to S35. For example, through this comparison, the information processing system 50 classifies the feature points present in each web before stacking and the feature points present in the laminate 80 into one of first type feature points, second type feature points, or third type feature points.

[0097] (Step S37) The information processing system 50 stores the comparison result of step S36 in the inspection data DB of the storage unit 52, and ends the process.

[0098] [Functions of Information Processing System 50] As shown in FIG. 6 , the information processing system 50 functions as an acquisition unit 511, an estimation unit 512, a generation unit 513, and an output unit 514 by the control unit 51 reading a program stored in the storage unit 52 and executing processing.

[0099] The acquisition unit 511 acquires pre-stacking feature point information. The pre-stacking feature point information is information about feature points present in each of the first web 81, the second web 82, and the third web 83 before stacking. The pre-stacking feature point information includes, for example, information similar to the pre-stacking feature point information for the database. The pre-stacking feature point information includes, for example, information about the feature point ID, position, area, length, width, maximum brightness, minimum brightness, classification, and presence or absence of concentrated dots for each of the multiple feature points present in each web before stacking.

[0100] The pre-stacking feature point information preferably includes at least information regarding the positions of feature points present on the first web 81 before stacking, the second web 82 before stacking, and the third web 83 before stacking. The positions of the feature points are expressed, for example, by XY coordinates based on a predetermined position on each web. The acquisition unit 511 acquires the pre-stacking feature point information from, for example, the terminal device 70. The acquisition unit 511 may also acquire the pre-stacking feature point information from the storage unit 52.

[0101] The acquiring unit 511 acquires pre-stacking feature information for each of a plurality of production lots, for example. The pre-stacking feature information includes information on feature points present in each of the first webs 81 having lot IDs a002 to a005, the second webs 82 having lot IDs b002 to b005, and the third webs 83 having lot IDs c002 to c005.

[0102] The estimation unit 512 estimates post-stacking feature point information based on the pre-stacking feature point information acquired by the acquisition unit 511. The post-stacking feature point information is information about feature points that may be present in the laminate 80 when the pre-stacking first web 81, the pre-stacking second web 82, and the pre-stacking third web 83 are stacked. The estimation unit 512 estimates post-stacking feature point information for, for example, all laminates 80 that may be formed by combining multiple production lots. The multiple production lots are, for example, the first web 81 with lot IDs a002 to a005, the second web 82 with lot IDs b002 to b005, and the third web 83 with lot IDs c002 to c005.

[0103] The estimation unit 512 estimates the post-stacking feature point information using a data set. In this data set, for example, feature points present in each of the first database web and the second database web are associated with feature points present in the database laminate of the first database web and the second database web. The estimation unit 512 estimates the post-stacking feature point information using, for example, a machine learning model formed using an inspection data DB. As shown in FIG. 8F , this inspection data DB includes information that associates feature points present in each of the first web 81, the second web 82, and the third web 83 before stacking with feature points present in the laminate 80.

[0104] The post-stacking feature point information estimated by the estimation unit 512 includes, for example, information about each of the multiple feature points present in the laminate 80, such as the feature point ID, position, area, length, width, maximum brightness, minimum brightness, division, and presence or absence of concentrated dots. The post-stacking feature point information preferably includes at least information about the position of the feature point present in the laminate 80. The position of the feature point is expressed, for example, by XY coordinates based on a predetermined position on the laminate 80. The post-stacking feature point information further includes, for example, information about the process that causes the feature point. Specifically, the estimation unit 512 can estimate whether a predetermined feature point present in the laminate 80 corresponds to a second-type feature point or a third-type feature point. When a predetermined feature point present in the laminate 80 is a third-type feature point, the estimation unit 512 can estimate whether the feature point originates from the first web 81, the second web 82, or the third web 83.

[0105] The generating unit 513 generates output information based on the post-stacking feature point information estimated by the estimating unit 512. The output unit 514 outputs the output information generated by the generating unit 513. The output unit 514 outputs the output information by, for example, displaying the output information on the display unit 76 of the terminal device 70.

[0106] The output information includes, for example, information regarding the feature point ID, position, area, and originating process of each of the multiple feature points present in the laminate 80. The output information may be the post-stacking feature point information itself. The output information further includes, for example, information regarding whether the feature points present in the laminate 80 satisfy a predetermined standard. The predetermined standard may be, for example, the number of feature points contained in the laminate 80 per unit length. The predetermined standard may be the size, density, periodicity, or strength of the feature points present in the laminate 80. The predetermined standard may be a combination of multiple conditions. The output information includes, for example, information regarding whether the feature points present in the laminate 80 satisfy a predetermined standard for multiple combinations of production lots of the first web 81, the second web 82, and the third web 83.

[0107] FIG. 11 shows an example of output information output by the output unit 514. For example, the combination of a production lot consisting of a first web 81 with a lot ID of a002, a second web 82 with a lot ID of b002, and a third web 83 with a lot ID of c002 meets the predetermined standard. For example, the combination of a production lot consisting of a first web 81 with a lot ID of a003, a second web 82 with a lot ID of b003, and a third web 83 with a lot ID of c003 does not meet the predetermined standard. For example, the combination of a production lot consisting of a first web 81 with a lot ID of a003, a second web 82 with a lot ID of b003, and a third web 83 with a lot ID of c004 meets the predetermined standard. For example, the combination of a production lot consisting of a first web 81 with a lot ID of a004, a second web 82 with a lot ID of b004, and a third web 83 with a lot ID of c003 meets the predetermined standard. By outputting such output information, the production manager of the laminate 80 can combine the production lots of the first web 81, the second web 82, and the third web 83 so that the specified standards are met.

[0108] The output information may include, for example, information regarding a combination of production lots of the first web 81, the second web 82, and the third web 83, such that the feature points present in the laminate 80 satisfy a predetermined standard. The display unit 76 may display, for example, a combination of the first web 81, the second web 82, and the third web 83, such that the feature points present in the laminate 80 satisfy the predetermined standard, among all combinations of production lots of the first web 81, the second web 82, and the third web 83.

[0109] The output information may include, for example, information regarding combinations of production lots of the first web 81, the second web 82, and the third web 83 that result in feature points in the laminate 80 that deviate from a predetermined standard. The display unit 76 may display, for example, combinations of the production lots of the first web 81, the second web 82, and the third web 83 that result in feature points in the laminate 80 that deviate from a predetermined standard, among all combinations of the production lots of the first web 81, the second web 82, and the third web 83. By outputting such output information, a production manager or the like of the laminate 80 can deliberately combine production lots of the first web 81, the second web 82, and the third web 83 that do not meet the standard, thereby improving the yield of the laminate 80 throughout the entire production process.

[0110] The output information may include, for example, information about an area of ​​the laminate 80 in which the existing feature points satisfy a predetermined standard. For example, when a portion of the area of ​​the laminate 80 satisfies a predetermined standard, the output information includes information about the width and length positions of this portion of the area of ​​the laminate 80. The output information may include information about the size or proportion of the area of ​​the laminate 80 in which the existing feature points satisfy the predetermined standard.

[0111] Fig. 12 is a flowchart showing an example of the procedure for generating output information executed in the information processing system 50. The processing of the information processing system 50 shown in the flowchart of Fig. 12 is stored as a program in the storage unit 52 of the information processing system 50, and is executed by the CPU controlling each unit.

[0112] (Step S101) The information processing system 50 first acquires pre-stacking feature point information regarding feature points present on each of the first web 81, the second web 82, and the third web 83. The information processing system 50 acquires the pre-stacking feature point information by receiving inspection data for each of the first web 81, the second web 82, and the third web 83 from the terminal device 70, for example.

[0113] (Step S102) The information processing system 50 estimates post-stacking feature point information related to feature points present in the laminate 80 based on the pre-stacking feature point information acquired in the processing of step S101. The information processing system 50, for example, inputs the pre-stacking feature point information to a classifier that has been trained in advance by machine learning. This makes it possible to estimate, for example, the position, area, length, width, maximum brightness, minimum brightness, classification, presence or absence of concentrated points, and contributing process of each of multiple feature points that may be present in the laminate 80. For example, the classifier is trained by machine learning using a large amount of past inspection data that has been prepared in advance by a learning method such as that described below. Specifically, the classifier is trained in machine learning using the pre-stacking feature point information for the database as input data and comparison data between the pre-stacking feature point information for the database and the post-stacking feature point information for the database as output data. The comparison data between the pre-stacking feature point information for the database and the post-stacking feature point information for the database is, for example, the data shown in FIG. 8F.

[0114] (Step S103) The information processing system 50 generates output information based on the post-stacking feature point information estimated in the process of step S102.

[0115] (Step S104) The information processing system 50 outputs the output information generated in the process of step S103. For example, the information processing system 50 outputs the output information by displaying the output information on the display unit 76 of the terminal device 70.

[0116] Next, a machine learning method for a trained model used in a classifier will be described.

[0117] FIG. 13 is a flowchart showing a machine learning method for a trained model. In the process of FIG. 13 , pre-stacked database feature point information prepared in advance is used as input, and comparison data between pre-stacked database feature point information and post-stacked database feature point information is used as output. Machine learning is performed using a large number of data sets as training sample data. The large number is, for example, i sets, where i is, for example, several thousand to several hundred thousand. The learning device functioning as a classifier may be, for example, a standalone high-performance computer using a CPU and GPU processor, or a cloud computer. Below, a learning method using a neural network configured by combining perceptrons such as deep learning in the learning device will be described, but this is not limited to this, and various other techniques may be applied. For example, random forests, decision trees, support vector machines, logistic regression, k-nearest neighbors, topic models, etc. may be applied.

[0118] (Step S111) The learning device reads learning sample data, which is teacher data. If it is the first time, the first set of learning sample data is read, and if it is the i-th time, the i-th set of learning sample data is read.

[0119] (Step S112) The learning device inputs the input data from the read learning sample data to the neural network.

[0120] (Step S113) The learning device compares the estimation result of the neural network with the correct answer data.

[0121] (Step S114) The learning device adjusts the parameters based on the comparison result, for example, by performing a process based on backpropagation, so as to reduce the difference in the comparison result.

[0122] (Step S115) If the learning device has completed processing of all data from the first to i-th sets, i.e., if the answer is YES, the process proceeds to step S116; if the processing is not completed, i.e., if the answer is NO, the process returns to step S111, where the learning device reads the next learning sample data and repeats the process from step S111 onwards.

[0123] (Step S116) The learning device determines whether or not to continue learning. If it continues, that is, if the answer is YES, the process returns to step S111, and the processes from the first set to the i-th set are executed again in steps S111 to S115. If it does not continue, that is, if the answer is NO, the process proceeds to step S117.

[0124] (Step S117) The learning device stores the trained model constructed in the processing up to this point, and then ends the processing. The storage destination includes the internal memory of the information processing system 50. In the processing of FIG. 12 described above, the stacked feature point information is estimated using the trained model generated in this manner.

[0125] [Operation and Effect of Information Processing System 50] In the information processing system 50 according to the present invention, post-stacking feature point information is estimated from pre-stacking feature point information. Then, output information is generated based on this post-stacking feature point information. That is, in the information processing system 50, the state of feature points present in the laminate 80 can be predicted from data on feature points present in each single layer of the first web 81, the second web 82, and the third web 83. Therefore, it becomes possible to predict the state of defects present in the laminate 80. The operation and effect of this will be described below.

[0126] Webs such as optical films require strict quality control. For example, defects in the web are strictly controlled. Defects include foreign matter, scratches, and dents. Such webs are manufactured through multiple processes, and quality inspections are conducted at each process.

[0127] A laminate including multiple webs has defects that existed in the webs before lamination and defects that occurred in the webs during the lamination process. Therefore, the defects present in a laminate including multiple webs are more complex than those present in a single-layer web. For example, even if each of the single-layer webs meets the specified defect specifications, a large number of defects may be generated during lamination depending on the combination of production lots. This may cause the laminate to fail to meet the specified defect specifications.

[0128] In contrast, the information processing system 50 can estimate the characteristic points present in the laminate 80 from the characteristic points present in each of the first web 81, the second web 82, and the third web 83. This allows, for example, a production manager of the laminate 80 to predict the state of defects present in the laminate 80 and determine whether or not the laminate 80 satisfies predetermined standards before manufacturing the laminate 80.

[0129] Furthermore, the information processing system 50 may output information regarding whether defects present in the laminate 80 satisfy predetermined standards for multiple combinations of production lots of the first web 81, the second web 82, and the third web 83. Alternatively, the information processing system 50 may output information regarding combinations of multiple production lots of the first web 81, the second web 82, and the third web 83 in which defects present in the laminate 80 satisfy predetermined standards. This allows, for example, a production manager of the laminate 80 to select a combination that satisfies the predetermined standards from among the production lots of the first web 81, the second web 82, and the third web 83. Therefore, the yield of the laminate 80 can be improved.

[0130] The information processing system 50 may output information regarding combinations of the first web 81, the second web 82, and the third web 83 in which defects in the laminate 80 deviate from predetermined specifications, among multiple combinations of production lots. This allows, for example, a production manager of the laminate 80 to select a combination from the production lots of the first web 81, the second web 82, and the third web 83 that will deviate from predetermined specifications when laminated. In this way, by intentionally combining the first web 81, the second web 82, and the third web 83 that do not meet the specifications, it is possible to improve the yield of the laminate 80 throughout the entire production process. In this way, use of the information processing system 50 reduces waste of raw materials, etc., leading to favorable results in terms of the environment, costs, etc.

[0131] Furthermore, the information processing system 50 estimates the post-stacking feature point information using an inspection data DB. This inspection data DB contains information that associates feature points present on the first web 81 before stacking, the second web 82 before stacking, and the third web 83 before stacking with feature points present on the laminate 80. This information processing system 50 can estimate the post-stacking feature point information with high accuracy even for webs in which the size, shape, orientation, brightness, etc. of feature points present on the webs can change before and after stacking.

[0132] In this inspection data DB, predetermined feature points present on the first web 80A and predetermined feature points present on the second web 80B are associated with each other based on their positions. Therefore, even if the size of feature points present on the webs changes before and after lamination, the information processing system 50 can associate the feature points of the webs before and after lamination.

[0133] Below, other embodiments of the information processing system 50 described in the first embodiment will be described. Note that, in order to avoid duplication of explanation, detailed explanations of components similar to those of the information processing system 50 described in the first embodiment will be omitted.

[0134] Second Embodiment Fig. 14 is a subroutine flowchart showing the alignment process of an information processing system 50 according to a second embodiment. Fig. 14 corresponds to Fig. 10 described in the first embodiment. This information processing system 50 uses kernel density estimation to align the X and Y coordinates of each web before stacking with the X and Y coordinates of the stack 80. In this respect, the information processing system 50 according to the second embodiment differs from the information processing system 50 according to the first embodiment. Except for this point, the information processing system 50 according to the second embodiment has the same configuration as the information processing system 50 according to the first embodiment and achieves the same effects.

[0135] (Step S451) The information processing system 50 obtains a probability density function by performing kernel density estimation on the feature points of the first web 81. The kernel density estimation is performed two-dimensionally, and a Gaussian kernel is used as the kernel function. For example, a predetermined value is used as the bandwidth. For example, a table correlating web product names with bandwidths is stored in the storage unit 52. The information processing system 50 may use a bandwidth value corresponding to the web product name, or may use different bandwidth values ​​depending on the number of feature points present on the first web 81. When estimating the kernel density of a given feature point, the information processing system 50 takes into account data surrounding the feature point. Then, the information processing system 50 sums the densities of each feature point to obtain a probability density function.

[0136] Fig. 15 shows an example of a probability density function calculated by kernel density estimation, in which the vertical and horizontal axes represent X and Y coordinates, and the intensity of the color represents the density.

[0137] (Step S452) The information processing system 50 obtains a probability density function of each feature point of the laminate 80 in the same manner as in step S451.

[0138] (Steps S453 to S455) The information processing system 50 compares the two obtained probability density functions and performs correspondence based on the density distribution. Then, the information processing system 50 calculates a transformation matrix based on the correspondence result and performs coordinate transformation of the X and Y coordinates for the feature points of the first web 81.

[0139] (Steps S456 to S457) The information processing system 50 performs the processes of S456 to S457 in the same manner as steps S406 to S407 in FIG.

[0140] The information processing system 50 may use kernel density estimation for all or part of the alignment of the XY coordinates of each web before stacking with the XY coordinates of the stack 80 .

[0141] In the information processing system 50 according to the second embodiment, similarly to the first embodiment, post-stacking feature point information is estimated from pre-stacking feature point information. Then, output information is generated based on this post-stacking feature point information. This makes it possible to predict the state of defects present in the stack 80.

[0142] The information processing system 50 may use other methods to align the X and Y coordinates of each web before stacking with the X and Y coordinates of the laminate 80. For example, the information processing system 50 may perform kernel density estimation of the feature points of one of the first web 81 and the laminate 80 before stacking. At this time, the information processing system 50 compares the obtained probability density function with the feature points of the other. This allows the information processing system 50 to align the X and Y coordinates of the first web 81 and the laminate 80 before stacking.

[0143] Third Embodiment FIG. 16 illustrates an application example of an information processing system 50 according to a third embodiment. The laminate 80 may be manufactured through multiple factories. For example, in factory 100A, an inspection device 90A inspects a first web 81 for characteristic features. The first web 81 is transported to factory 100B. The first web 81 may then be transported to another factory. In factory 100B, a second web 82 and a third web 83 are laminated on the first web 81 to manufacture the laminate 80. The manufacturing and characteristic inspection of the second web 82 and the third web 83 may be performed in factory 100A or 100B, or may be performed in another factory. For example, a manufacturing device 2000A is provided in factory 100A, and a manufacturing device 2000B is provided in factory 100B.

[0144] The information processing system 50 acquires the pre-stacking feature point information from, for example, a terminal device 70A provided in the factory 100A and a terminal device 70B provided in the factory 100B. The information processing system 50 may acquire the pre-stacking feature point information from either the terminal device 70A or the terminal device 70B.

[0145] In the information processing system 50 according to the third embodiment, similarly to the first embodiment, post-stacking feature point information is estimated from pre-stacking feature point information. Then, output information is generated based on this post-stacking feature point information. This makes it possible to predict the state of defects present in the stack 80.

[0146] (Manufacturing Apparatus 2000) FIG. 17 shows an example of a schematic configuration of the manufacturing apparatus 2000. The manufacturing apparatus 2000 laminates, for example, a first protective film 12 and a second protective film 13 on both sides of a polarizer 1c. This results in a laminated polarizing film having a layer structure of, for example, first protective film 12 / adhesive layer 31 / polarizer 1c / adhesive layer 32 / second protective film 13. For example, the first protective film 12 corresponds to a specific example of the first web 81. For example, the adhesive layer 31 corresponds to a specific example of the adhesive layer 801. For example, the polarizer 1c corresponds to a specific example of the second web 82. For example, the adhesive layer 32 corresponds to a specific example of the adhesive layer 802. For example, the second protective film 13 corresponds to a specific example of the third web 83.

[0147] The manufacturing apparatus 2000 includes, for example, in order from the upstream side, a wet treatment apparatus 204, a drying apparatus 205, and a laminating apparatus 206. The manufacturing apparatus 2000 further includes a feeding section 202.

[0148] The wet-treatment device 204 has a first roll unit 41, a conveying unit 42, and a processing unit. A long strip of untreated hydrophilic polymer film 1a is wound around the first roll unit 41. The conveying unit 42 conveys the hydrophilic polymer film 1a. In the processing unit, the hydrophilic polymer film 1a is transformed into a polarizer 1b.

[0149] The transport unit 42 has a plurality of guide rolls, etc. The transport unit 42 unwinds the hydrophilic polymer film 1a wound around the first roll unit 41 and transports it to a processing unit.

[0150] The processing section includes, for example, a swelling processing tank 4A, a dyeing processing tank 4B, a crosslinking processing tank 4C, a stretching processing tank 4D, and a washing processing tank 4E in this order from the upstream side.

[0151] The swelling treatment tank 4A is a treatment tank containing a swelling treatment liquid. The swelling treatment liquid swells the hydrophilic polymer film 1a. The dyeing treatment tank 4B is a treatment tank containing a dyeing treatment liquid. The dyeing treatment liquid dyes the hydrophilic polymer film 1a. The crosslinking treatment tank 4C is a treatment tank containing a crosslinking treatment liquid. The crosslinking treatment liquid crosslinks the dyed hydrophilic polymer film 1a. The stretching treatment tank 4D is a treatment tank containing a stretching treatment liquid. The stretching treatment liquid is not particularly limited, but for example, a solution containing a boron compound as an active ingredient can be used. The cleaning treatment tank 4E is a treatment tank containing a cleaning treatment liquid. The cleaning treatment liquid cleans the hydrophilic polymer film 1a after stretching. The cleaning treatment liquid is a treatment liquid for cleaning treatment liquids such as the dyeing treatment liquid and the crosslinking treatment liquid that have adhered to the hydrophilic polymer film 1a. Typical cleaning treatment liquids used are water such as ion-exchanged water, distilled water, and pure water.

[0152] The drying device 205 has a conveying section 501 and a heating section. There may be one drying device 205, or two or more drying devices 205 may be provided side by side in the polarizer conveying direction. In the illustrated example, for example, one drying device 205 is provided on the polarizer conveying path. The conveying section 501 conveys the long strip-shaped polarizer 1b. The heating section applies heat to the polarizer 1b to dry it. In this way, a polarizer 1c is formed.

[0153] The heating section of the drying device 205 has, for example, a chamber 502 and a heat source. The chamber 502 has a space 503 therein in which the polarizer can be transported. For example, an inspection device 90A is disposed on the transport path between the drying device 205 and the laminating device 206. The inspection device 90A inspects the polarizer 1c before lamination.

[0154] The laminating apparatus 206 includes a conveying unit 61, an adhesive coating unit 64, a bonding unit 67, and a chamber 69. The conveying unit 61 conveys the polarizer 1c and the first protective film 12. The chamber 69 surrounds the adhesive coating unit 64 and the bonding unit 67.

[0155] The conveying unit 61 has guide rolls and the like. The conveying unit 61 conveys the long strip-shaped polarizer 1c dried by the drying device 205 to the bonding unit 67. The conveying unit 61 also conveys the long strip-shaped first protective film 12 and the like to the bonding unit 67. For example, an inspection device 90B is disposed on the conveying path downstream of the bonding unit 67. The inspection device 90B inspects the laminate of the first protective film 12 / adhesive layer 31 / polarizer 1c / adhesive layer 32 / second protective film 13.

[0156] The adhesive coating unit 64 has a gravure roll 641. The gravure roll 641 is a coating roll that coats an adhesive on a film. The adhesive coating unit 64 is arranged upstream of the laminating unit 67. For example, the adhesive coating unit 64 is arranged on one side of the first protective film 12 and one side of the second protective film 13. The adhesive coating unit 64 may also be arranged on one side of the polarizer 1c and the other side of the polarizer 1c.

[0157] The adhesive coating unit 64 includes, for example, a gravure roll 641 which is a coating roll, a container 642 in which an adhesive is stored, and a doctor blade 643. The adhesive coating unit 64 may also include a backup roll as necessary. The backup roll is disposed opposite the gravure roll 641 with the film sandwiched therebetween.

[0158] A plurality of cells are formed on the surface of the gravure roll 641. The plurality of cells are recesses into which adhesive is placed. The gravure roll 641 rotates so that its surface comes into contact with the adhesive 65 stored in a container 642. As the gravure roll 641 rotates, the adhesive 65 adheres to the surface of the gravure roll 641, including the cells. Excess adhesive 65 is scraped off into the container 642 by a doctor blade 643. When the gravure roll 641 comes into contact with the film, the adhesive 65 in the cells is transferred to one side of the first protective film 12 and the second protective film 13. In this way, the adhesive 65 is solidly coated from the gravure roll 641 onto one side of each of the first protective film 12 and the second protective film 13.

[0159] The adhesive for bonding the polarizer 1c to the first protective film 12 and the second protective film 13 is not particularly limited, but it is preferable to use an active energy ray-curable adhesive as described above. Any conventionally known active energy ray-curable adhesive can be used. The active energy ray-curable adhesive generally contains an active energy ray-curable component and a polymerization initiator, and optionally contains various additives.

[0160] The unwinding unit 202 has an easy-adhesion treatment tank 21, a cleaning treatment tank 22, and a heat treatment tank 23. The easy-adhesion treatment tank 21 performs easy-adhesion treatment on the surface of the second protective film 13 to which the polarizer 11 is bonded. For example, the easy-adhesion treatment tank 21 performs corona discharge treatment, plasma treatment, or the like. In corona discharge treatment, a high voltage is applied to a wire or sawtooth electrode inside a chamber. The cleaning treatment tank 22 has a configuration similar to the above-described cleaning treatment tank 4E. A cleaning treatment liquid is stored in the cleaning treatment tank 22. The second protective film 13 is cleaned by the cleaning treatment liquid. The heat treatment tank 23 has a configuration similar to the drying device 205. The heat treatment tank 23 heats and dries the second protective film 13. The drying temperature of the heat treatment tank 23 is changed, for example, by the manufacturing management system 3000.

[0161] Manufacturing apparatus 2000 further includes a second roll unit 62 and a third roll unit 63. A long strip of first protective film 12 is wound around second roll unit 62. A long strip of second protective film 13 is wound around third roll unit 63. First protective film 12 and second protective film 13 are transported from second roll unit 62 and third roll unit 63 to bonding unit 67. First protective film 12 and second protective film 13 are inspected by an inspection device before being wound around second roll unit 62 and third roll unit 63, respectively.

[0162] The configuration of the information processing system 50 described above is a description of the main configuration in explaining the features of the above embodiment, but is not limited to the above configuration and can be modified in various ways within the scope of the claims. Furthermore, configurations included in general information processing devices or information processing systems are not excluded. For example, the information processing system 50 may include an inspection device 90. Furthermore, the feature point generation function of the analysis unit 93 of the inspection device 90 may be performed by the control unit 51 of the information processing system 50.

[0163] For example, in the above embodiment, a laminate 80 composed of a first web 81, a second web 82, and a third web 83 is exemplified, but the laminate 80 may have at least two layers, or may have four or more layers.

[0164] The pre-stacking feature point information may also include information about feature points present in a stack of multiple webs. For example, the pre-stacking feature point information may include information about feature points present in a stack of the first web 81 and the second web 82, and information about feature points present in the third web 83.

[0165] The pre-stacking feature information may also be information relating to feature points present in some of the webs that make up the laminate 80 .

[0166] In the above embodiment, an example was described in which the first production lot and the second production lot of the present invention are production lots with different lot IDs. A lot ID is assigned to each web cut to a predetermined length and width, for example. A lot ID may also be assigned for each web production time. The first production lot and the second production lot of the present invention do not have to be consecutive.

[0167] In the above embodiment, an example has been described in which the information processing system 50 estimates the post-stacking feature point information using a machine learning model, but the information processing system 50 may estimate the post-stacking feature point information using other methods. For example, the information processing system 50 may estimate the post-stacking feature point information by statistical processing using an inspection data DB. Furthermore, the terminal device 70, the terminal device 70A, or the terminal device 70B may estimate the post-stacking feature point information.

[0168] Furthermore, the means and methods for performing various processes in the information processing system 50 according to the above-described embodiment can be realized by either a dedicated hardware circuit or a programmed computer. The program may be provided, for example, by a computer-readable recording medium such as a USB memory or a DVD-ROM, or may be provided online via a network such as the Internet. In this case, the program recorded on the computer-readable recording medium is typically transferred and stored in a storage unit such as a hard disk. The program may also be provided as standalone application software, or may be incorporated into the software of a device as a function of that device. DVD is an abbreviation for Digital Versatile Disc.

[0169] While embodiments of the present invention have been described and illustrated in detail, the disclosed embodiments are made for purposes of illustration and example only and are not intended to be limiting, and the scope of the present invention should be construed by the language of the appended claims.

[0170] This application is based on a Japanese patent application (Patent Application No. 2024-65965) filed on April 16, 2024, the disclosure of which is incorporated herein by reference in its entirety.

[0171] 50 Information processing system 51 Control unit 511 Acquisition unit 512 Estimation unit 513 Generation unit 514 Output unit 52 Storage unit 53 Communication unit 90 Inspection device 2000 Manufacturing device

Claims

1. An information processing system comprising: an acquisition unit that acquires pre-stacking feature point information regarding feature points present in each of a first web and a second web; an estimation unit that estimates post-stacking feature point information regarding feature points present in a laminate including the first web and the second web based on the acquired pre-stacking feature point information; and a generation unit that generates output information based on the estimated post-stacking feature point information.

2. The information processing system of claim 1, wherein the pre-stacking feature information includes information regarding feature points present on the first web of each of the first and second production lots and information regarding feature points present on the second web of each of the third and fourth production lots, and the estimation unit estimates the post-stacking feature information for multiple combinations of the first and second production lots and the third and fourth production lots.

3. An information processing system as described in claim 2, wherein the output information includes information regarding whether or not the characteristic points present in the laminate satisfy a predetermined standard for multiple combinations of the first production lot and the second production lot and the third production lot and the fourth production lot.

4. An information processing system as described in claim 2, wherein the output information includes information regarding combinations of the first and second production lots and the third and fourth production lots in which characteristic points present in the laminate satisfy predetermined standards.

5. An information processing system as described in claim 2, wherein the output information includes information regarding combinations of the first and second production lots and the third and fourth production lots in which characteristic points present in the laminate deviate from predetermined standards.

6. An information processing system as described in claim 1, wherein the estimation unit estimates the post-stacking feature point information using a data set in which feature points present on each of the first database web and the second database web are associated with feature points present in the database stack of the first database web and the second database web.

7. The information processing system according to claim 6, wherein the estimation unit estimates the post-stacking feature point information using a machine learning model trained on the dataset.

8. An information processing system as described in claim 6, wherein the feature points present in the database laminate are classified into second type feature points that are not present in the first database web and the second database web but appear in the database laminate, and third type feature points that are present in the first database web or the second database web and remain in the database laminate.

9. The information processing system according to claim 1, further comprising an output unit that outputs the generated output information.

10. An information processing system as described in claim 1, wherein the pre-stacking feature information further includes information regarding feature points present in a third web, and the laminate is composed of the first web, the second web, and the third web.

11. An information processing method comprising: acquiring pre-stacking feature point information relating to feature points present in each of a first web and a second web; estimating post-stacking feature point information relating to feature points present in a laminate including the first web and the second web based on the acquired pre-stacking feature point information; and generating output information based on the estimated post-stacking feature point information.

12. An information processing program that causes a computer to execute the information processing method according to claim 11.

Images