WO2023038017A1 - 欠点分類システム - Google Patents
欠点分類システム Download PDFInfo
- Publication number
- WO2023038017A1 WO2023038017A1 PCT/JP2022/033356 JP2022033356W WO2023038017A1 WO 2023038017 A1 WO2023038017 A1 WO 2023038017A1 JP 2022033356 W JP2022033356 W JP 2022033356W WO 2023038017 A1 WO2023038017 A1 WO 2023038017A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- defect
- paper
- classification system
- feature amount
- classification
- Prior art date
Links
- 230000007547 defect Effects 0.000 title claims abstract description 257
- 238000013145 classification model Methods 0.000 claims abstract description 24
- 238000001514 detection method Methods 0.000 claims abstract description 16
- 238000010801 machine learning Methods 0.000 claims abstract description 8
- 238000000034 method Methods 0.000 claims description 36
- 238000003384 imaging method Methods 0.000 abstract description 37
- 238000004364 calculation method Methods 0.000 abstract description 12
- 238000000605 extraction Methods 0.000 abstract description 11
- 239000000123 paper Substances 0.000 description 87
- 238000002360 preparation method Methods 0.000 description 9
- 238000005259 measurement Methods 0.000 description 8
- 238000012544 monitoring process Methods 0.000 description 8
- 238000003860 storage Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 4
- 239000006185 dispersion Substances 0.000 description 4
- 238000012423 maintenance Methods 0.000 description 4
- 239000000126 substance Substances 0.000 description 4
- 241000238631 Hexapoda Species 0.000 description 3
- 239000011087 paperboard Substances 0.000 description 3
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 3
- 238000013527 convolutional neural network Methods 0.000 description 2
- 238000009434 installation Methods 0.000 description 2
- 230000005856 abnormality Effects 0.000 description 1
- 239000000654 additive Substances 0.000 description 1
- 239000002390 adhesive tape Substances 0.000 description 1
- 239000004566 building material Substances 0.000 description 1
- 238000003490 calendering Methods 0.000 description 1
- 239000011111 cardboard Substances 0.000 description 1
- 239000003795 chemical substances by application Substances 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 239000000356 contaminant Substances 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000001035 drying Methods 0.000 description 1
- 239000000428 dust Substances 0.000 description 1
- 239000000945 filler Substances 0.000 description 1
- 239000003292 glue Substances 0.000 description 1
- 239000007788 liquid Substances 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000004806 packaging method and process Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000011002 quantification Methods 0.000 description 1
- 239000002994 raw material Substances 0.000 description 1
- 238000004513 sizing Methods 0.000 description 1
- 238000011144 upstream manufacturing Methods 0.000 description 1
- 239000002023 wood Substances 0.000 description 1
Images
Classifications
-
- 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
-
- D—TEXTILES; PAPER
- D21—PAPER-MAKING; PRODUCTION OF CELLULOSE
- D21F—PAPER-MAKING MACHINES; METHODS OF PRODUCING PAPER THEREON
- D21F7/00—Other details of machines for making continuous webs of paper
-
- 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
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
- G06V10/443—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components by matching or filtering
- G06V10/449—Biologically inspired filters, e.g. difference of Gaussians [DoG] or Gabor filters
- G06V10/451—Biologically inspired filters, e.g. difference of Gaussians [DoG] or Gabor filters with interaction between the filter responses, e.g. cortical complex cells
- G06V10/454—Integrating the filters into a hierarchical structure, e.g. convolutional neural networks [CNN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- 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
- G01N21/8914—Investigating the presence of flaws or contamination in moving material, e.g. running paper or textiles characterised by the material examined
- G01N2021/8917—Paper, also ondulated
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B19/00—Programme-control systems
- G05B19/02—Programme-control systems electric
- G05B19/418—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM]
- G05B19/41875—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM] characterised by quality surveillance of production
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B2219/00—Program-control systems
- G05B2219/30—Nc systems
- G05B2219/32—Operator till task planning
- G05B2219/32222—Fault, defect detection of origin of fault, defect of product
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30124—Fabrics; Textile; Paper
Definitions
- the present invention relates to a defect classification system, and more particularly, for classifying defect information based on defects of paper that has passed through a dry part in a papermaking process into corresponding defect cause items among a plurality of preset defect cause items. about the defect classification system of.
- the dry pulp sheet is fibrillated, and additives such as fillers and sizing agents are added, and the mixture is stirred and mixed to form a pulp dispersion.
- a papermaking process is performed in which the pulp dispersion is passed through various parts such as a wet part including a wire part and a press part, a dry part, and a reel part to form paper.
- an applicator for applying a chemical solution to a portion of a paper machine that comes into contact with paper an operation panel for setting application conditions for the applicator, a monitoring camera for monitoring a monitoring target portion, and the monitoring camera. and a control device connected via a network, wherein the control device uses an image captured by the surveillance camera to digitize the state of the monitored part by binarization;
- a monitoring system is known that has a display unit that displays detection data obtained by digitizing the amount of change in the binarized value during operation with respect to the binarized value in the steady state, and a storage unit that stores the detected data. (See, for example, Patent Document 1).
- Patent Literature 1 Although it is possible to recognize the monitoring target part that causes the defect and to prevent the occurrence of the defect based on it on the paper, the monitoring camera is used in high temperature and high humidity. Since it is necessary to install the monitor at a specific site to be monitored, there is a problem that preparation and maintenance are difficult.
- the present invention has been made in view of the above circumstances, and aims to provide a defect classification system that is easy to prepare and maintain and that can recognize the causes of defects.
- the inventors of the present invention conducted extensive studies to solve the above problems, and found that the defects that occurred in the paper that had undergone the dry part had characteristics. Then, the paper that has undergone the dry part is used as an imaging target, and an extracting means for extracting the feature quantity of the defect, a calculating means for calculating the degree of certainty using a machine-learned classification model, a display means for displaying the degree of certainty, and the like are provided. As a result, the inventors have found that the above problems can be solved, and have completed the present invention.
- the present invention provides (1) a method for classifying defect information based on defects of paper that has passed through a dry part in a papermaking process into corresponding defect cause items among a plurality of defect cause items based on preset causes of defects.
- a defect classification system comprising an imaging device that captures an image of paper that has passed through a dry part and acquires captured image data, a detection means that detects a defect of the paper in the image data, and features of the defect. an extracting means for extracting a quantity, a calculating means for calculating a certainty factor of a defect cause item for a classification model in which a reference feature amount is set in advance based on the feature amount of the defect, and displaying the certainty factor. and display means for the defect classification system, wherein the classification model is obtained by learning a reference feature amount by machine learning from the relationship between the feature amount of the defect accumulated in advance and a plurality of defect cause items. exist.
- the calculation means calculates the confidence factor for each defect cause item, and classifies the defect information into the defect cause item having the maximum confidence factor among a plurality of confidence factors. It resides in the defect classification system described in (1) above, further comprising classification means.
- the classification model is further learned by machine learning from the relationship between the feature amount of the defect whose maximum certainty factor is equal to or less than a predetermined value and the defect cause item that classifies it. It resides in the defect classification system described in (2) above.
- defect information of the defect includes coordinate data of the defect when the coordinates are provided on the paper in addition to the feature amount of the defect. It resides in the defect classification system described in the first.
- the present invention resides in the defect classification system according to any one of (1) to (4) above, wherein (5) the defect cause item includes at least an item that causes the defect to adhere to a foreign matter in a dry part. .
- the paper that has passed through the dry part in the papermaking process is taken as an object to be imaged. This eliminates the need to enter the apparatus in which the paper making process is performed in order to install the image pickup apparatus, so preparation and maintenance for using the defect classification system can be performed very safely and easily.
- the defect classification system since the image of the paper that has passed through the dry part is targeted, the defects that occur in the paper are due to the part before the dry part of the stock preparation process or the paper making process.
- the defect classification system of the present invention is provided with the detection means and the extraction means, the defects occurring in the paper are detected and their characteristic quantities are extracted. Further, since the calculation means is provided, the certainty factor for the defect cause item is calculated from the feature amount of the defect. Then, the certainty factor is displayed by the display means. As a result, the defect classification system can recognize the degree of certainty for the defect cause item of the defect. As a result, it becomes possible to classify the defect into defect cause items with a high degree of certainty, and at the same time, it becomes possible to recognize the cause of the defect.
- the defect cause item includes at least an item that causes the defect to adhere to foreign matter in the dry part
- the degree of confidence in this item it is possible to determine whether the cause of the defect is due to the dry part. , it is possible to recognize whether it is due to something other than the dry part.
- the classification means when the classification means is further provided, it becomes possible to recognize the defect cause item including the defect. Further, the accuracy of classification can be improved by changing the reference feature amount of the classification model based on the defect information of the defect and the defect cause item by which the defect is classified.
- the classification model is further trained by machine learning based on the relationship between the characteristic amount of the defect whose maximum certainty factor is equal to or less than a predetermined value and the defect cause item classified by the characteristic amount. be able to. This further improves the accuracy of the confidence calculated by the classification model.
- the defect information when the defect information includes the coordinate data of the defect when the coordinates are provided on the paper, the more detailed position of the part causing the defect is obtained from the coordinate data of the defect. It is possible to infer up to For example, if defects occur repeatedly in the longitudinal direction of the paper, foreign matter may be attached to the rolls that guide the paper, felt, canvas, or other loop-shaped tools. From the distance between them, it is possible to calculate the diameter of the roll when foreign matter adheres to the roll, and the position of the foreign matter on the loop-shaped tool when the foreign matter adheres to the loop-shaped tool.
- FIG. 1 is a block diagram for explaining the configuration of the defect classification system according to this embodiment.
- FIG. 2 is a schematic side view for explaining the papermaking process in which the defect classification system according to the present embodiment is used and the installation position of the imaging device.
- FIG. 3 is a diagram showing defect cause items and images of defects included in the defect cause items in the defect classification system according to the present embodiment.
- FIG. 4 is a flowchart showing the defect classification method according to this embodiment.
- the defect classification system is a program stored in a computer, and is executed on a paper machine, which is hardware, in which an imaging device is arranged.
- a paper machine which is hardware, in which an imaging device is arranged.
- the number of computers is not limited to one, and a plurality of computers may be connected by wire or wirelessly.
- a video camera, a line sensor camera, an area sensor camera, or the like can be used as the imaging device.
- the paper machine is not particularly limited as long as it is a series of machines capable of performing a so-called papermaking process, and conventionally known machines can be appropriately employed.
- target paper is not particularly limited as long as it can be manufactured by a papermaking process.
- the paper includes printing paper, newsprint, coated paper, packaging paper, thin paper, household paper such as toilet paper and tissue paper, so-called western paper such as miscellaneous paper, cardboard base paper, white paperboard, colored paperboard, and paper. So-called paperboard such as base paper for tubes, base paper for building materials, and various mounts can be used.
- the term “defect” means adherence of foreign matter (including inclusion of foreign matter), holes, partial paper breaks, water stains, oil stains, uneven density of pulp, and the like. Also, “foreign matter” includes not only pitch but also insects, paper dust, slime, and the like.
- the term “pitch” refers to contaminants contained in pulp, which is the raw material of paper, and includes sticky substances derived from adhesive tapes and glues, and sticky substances derived from wood.
- the "defect cause item” is an item that distinguishes items that cause defects.
- the “plurality of defect cause items” is a collection of defect cause items, and the defect causes of the defect cause items are different from each other.
- Confidence is a statistical measure (probability) of how certain the prediction is.
- FIG. 1 is a block diagram for explaining the configuration of the defect classification system according to this embodiment.
- a defect classification system 10 includes an imaging device 11 for capturing an image of paper that has passed through a dry part, and acquiring captured image data; means 12, extracting means 13 for extracting the feature quantity of the defect, and calculation of the certainty factor of the defect cause item for the classification model in which the reference feature quantity is set in advance based on the feature quantity of the defect.
- Calculation means 14 display means 15 for displaying certainty, classification means 16 for classifying defect information into defect cause items whose certainty is equal to or greater than a predetermined value, and image data storing means. and storage means 17 .
- defect classification system by providing these means, it is possible to recognize the degree of certainty for each defect cause item of the defect. As a result, as will be described later, it is possible to classify the defect into defect cause items with a high degree of certainty, and at the same time, it is possible to recognize the cause of the defect.
- the imaging means 11 is means for causing the imaging device to image the paper that has passed through the dry part and acquiring the image data of the image.
- the imaging means 11 is connected to an imaging device via a wired or wireless network. Then, a command to image the paper that has passed through the dry part is sent to the imaging device, and the imaging device is caused to start imaging.
- the imaging unit 11 receives image data captured by the imaging device from the imaging device. Note that the resolution of the image data is preferably 10 to 500 MHz from the viewpoint of the size of defects to be detected, which will be described later.
- the image data received by the imaging means 11 is stored in the storage means 17 .
- FIG. 2 is a schematic side view for explaining the papermaking process in which the defect classification system according to the present embodiment is used and the installation position of the imaging device.
- the papermaking process is performed using a paper machine M.
- a dispersion liquid in which pulp is dispersed in water is placed on a papermaking wire, excess water is allowed to fall naturally, and a wire part W1 is passed between a pair of press rolls, and felt is passed through the press rolls.
- It has a wet part W consisting of a press part W2 pressed with a canvas K, a dry part D dried by contacting a heated dryer roll D1 through a canvas K, a calender part C, and a reel part R.
- a doctor blade D2 is in contact with the dryer roll D1 to remove foreign substances adhering to the surface of the dryer roll D1.
- the pulp dispersion obtained in the stock preparation process passes through the wet part W to be in the state of the so-called wet paper P1, and passes through the dry part D to dry the wet paper P1. It becomes the state of the paper P2.
- the paper P2 passes through the calendering part C to smooth the irregularities on the surface of the paper P2.
- the imaging device G is installed above the paper P2 between the calendar part C and the reel part R. As shown in FIG. As a result, the surface of the paper P2 is imaged by the imaging device G between the calendar part C and the reel part R. As shown in FIG. Thus, in the defect classification system 10, it is not necessary to enter the apparatus where the paper making process is performed in order to install the imaging apparatus G, so preparation and maintenance can be performed extremely safely and easily.
- the detection means 12 is a means for detecting defects that have occurred on the paper in the captured image data.
- the image data is for the paper P2 that has passed through the dry part D. In other words, it can be said that the defect detected in the image data occurred in the part before the dry part D of the stock preparation process or the papermaking process.
- the detection means 12 not only detects defects, but also provides virtual coordinates on paper and detects coordinate data of the defects. This makes it possible to estimate the more detailed position of the part that causes the defect from the coordinates of the defect.
- the paper transport direction is the Y axis and the paper width direction is the X axis.
- the value of Y should be increased toward the upstream side along the longitudinal direction of the paper P, and the value of X should be increased toward the left end to the right end.
- the distance between adjacent coordinates in the X-axis direction is 0.001 to 10 mm
- the distance between adjacent coordinates in the Y-axis direction is 0.001 to 10 mm. In this case, even if a very small defect occurs in the paper P2 within a visible range, it can be reliably detected.
- Defect detection is quantified by intensity measurement or RGB measurement.
- the image may be converted into a black and white binary image using a gray scale, and the gradation may be digitized by dividing it into 256 steps from 0 to 255, for example.
- a specific color eg, blue
- the detection means 12 employs intensity measurement or RGB measurement, it is possible to easily identify the presence or absence of defects.
- the image of the defect, the coordinate data obtained by the detection means 12, and the numerical data of the defect are stored in the storage means 17 as defect information.
- the extraction means 13 is a means for extracting the feature amount of the defect using the defect data digitized by the detection means 12 .
- a filter method, a wrapper method, an embedding method, or the like can be appropriately employed as a method for extracting feature amounts.
- specific examples of feature amounts to be extracted include the size, shape, maximum value, minimum value, gradation difference, gradation, and the like of the digitized defect.
- a CNN Convolutional Neural Network
- LeNet LeNet
- AlexNet AlexNet
- VGG VGG16, VGG19
- GoogleLeNet ResNet
- the feature amount of the defect is stored in the storage means 17 as defect information.
- the calculation means 14 is a means for calculating a certainty factor for each defect cause item for a classification model for which a reference feature amount is set in advance, based on the defect feature amount extracted by the extraction means 12 .
- FIG. 3 is an explanatory diagram showing defect cause items and images of defects included in the defect cause items in the defect classification system according to the present embodiment.
- the "pitch” shown in FIG. 3 is caused by adhesion of roll-derived or canvas-derived pitch in the dry part, and the “doctor” is caused by adhesion of doctor-derived pitch in the dry part.
- Paper-making is the paper-making of foreign matter in the paper stock preparation process.
- "hole” is the cause of the defect
- "oil stain” is the cause of the defect due to adhesion of oil
- "insect” is the cause of the defect due to adhesion of insects
- "ear cut” is the cause of the defect.
- the broken ends are the cause of the defect.
- the classification model learns the reference feature amount by machine learning from the defect cause item shown in FIG. 3 and the feature amount extracted from the accumulated defect image included in the defect cause item. It is what I let you do. Note that when new data of defect cause items and defect feature amounts included in the defect cause items are newly obtained, the classification model can be further trained. That is, the reference feature amount can be changed. This further improves the accuracy of the confidence calculated by the classification model.
- the defect cause items include at least an item that the defect is caused by adhesion of foreign matter to the dry part.
- the degree of certainty for this item it is possible to recognize whether the cause of the defect is due to the dry part or due to something other than the dry part.
- the adhesion of pitch shown in FIG. 3 occurs in the dry part, and the paper-making occurs in the stock preparation process. Alternatively, it becomes possible to recognize whether it is due to the dry part.
- the classification model calculates a certainty factor for each defect cause item with respect to the feature amount of the defect extracted by the extraction means 12 . Then, the calculated certainty is displayed by the display means 15 . This makes it possible to recognize defect cause items with high confidence and defect cause items with low confidence.
- a monitor, a touch panel, or the like may be adopted as the display means 15 .
- the classification means 16 stores the defect information of one defect in the folder of the defect cause item having the maximum value of the certainty calculated for each defect cause item. It is a means of classification. This makes it possible to recognize the defect cause item that includes the defect. Further, the accuracy of classification can be improved by changing the reference feature amount of the classification model based on the defect information of the defect and the defect cause item by which the defect is classified.
- the classification model when the confidence of the maximum value is equal to or less than a predetermined value set in advance, from the relationship between the feature amount of the defect and the defect cause item classified by the defect, the classification model is further classified by machine learning. Learning is preferred.
- a classification operation classifies defects whose maximum confidence factor is equal to or less than a preset value into defect cause items based on the maximum confidence factor, and then checks whether or not it is correct. This may be performed by human confirmation, or by human classification of defects with a maximum certainty factor equal to or less than a preset predetermined value as correct defect cause items. This further improves the accuracy of the confidence calculated by the classification model.
- the predetermined value can be set arbitrarily.
- FIG. 4 is a flowchart showing the defect classification method according to this embodiment.
- the defect classification method includes an image capturing step S11 in which the image capturing apparatus G captures an image of the paper that has passed through the dry part and acquires captured image data; Based on step S12, extraction step S13 for extracting the feature amount of the defect, and the feature amount of the defect, the certainty factor of the defect cause item is calculated for the classification model 20 in which the reference feature amount is set in advance.
- display means S15 for displaying the degree of certainty; and a classification step S16 for classifying the defect information into defect cause items whose certainty is equal to or greater than a predetermined value.
- the imaging means 11 and the detection means 12 are stored in the first computer C1, and at least the extraction means 13, the calculation means 14, the display means 15, the classification means 16, and the storage means 17 are stored in the second computer C1. stored in computer C2.
- the first computer C1 and the second computer C2 are connected via a cloud or the like. Therefore, in the defect classification method, the imaging step S11 and detection step S12 are performed using the first computer C1, and the extraction step S13, calculation step S14, display step S15 and classification step S16 are performed using the second computer C2. be done.
- a plurality of first computers C1 that perform the imaging step S11 and the like may be connected to the second computer C2.
- the defect classification method As with the defect classification system described above, since it has a detection step S11 and an extraction step S12, defects occurring in the paper are detected and their characteristic quantities are extracted. Further, since the calculation step S13 is provided, the certainty factor for the defect cause item is calculated from the feature amount of the defect. Then, the certainty is displayed by the display step S15. As a result, in the defect classification method, it is possible to recognize the degree of certainty for the defect cause item of the defect. As a result, it becomes possible to classify the defect into defect cause items with a high degree of certainty, and at the same time, it becomes possible to recognize the cause of the defect. Further, since the classification step S16 is included, the classification accuracy can be improved by changing the reference feature amount of the classification model based on the defect information of the defect and the defect cause item by which the defect is classified. can.
- the imaging device G is installed above the paper P2 between the calendar part C and the reel part R, but is not limited to this position.
- it may be installed below the paper P2, or may be installed both above and below the paper P2.
- the paper making process has a wet part W, a dry part D, a calendar part C and a reel part R, but is not limited to this.
- calendar part C and reel part R may not be included.
- a processing machine or the like that cuts and collects the paper may be installed.
- the imaging device G is installed above the paper P2 between the calendar part C and the reel part R. Not limited. For example, it may be installed above the paper P2 between the dry part D and the calendar part C. Further, if the imaging device G is installed downstream of the dry part D, another imaging device may be provided. In this case, more detailed classification becomes possible.
- defects are detected by digitizing them by intensity measurement or RGB measurement. is also possible.
- the defect classification system of the present invention can be used as a system for classifying defect information based on defects of paper that has passed through the dry part in the papermaking process into corresponding defect cause items among a plurality of preset defect cause items. According to the defect classification system of the present invention, preparation and maintenance are easy, and causes of defects can be recognized.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Theoretical Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Immunology (AREA)
- Chemical & Material Sciences (AREA)
- Biochemistry (AREA)
- Analytical Chemistry (AREA)
- Pathology (AREA)
- Evolutionary Computation (AREA)
- Multimedia (AREA)
- Artificial Intelligence (AREA)
- Textile Engineering (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Databases & Information Systems (AREA)
- Computing Systems (AREA)
- Quality & Reliability (AREA)
- Signal Processing (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- Biodiversity & Conservation Biology (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
- Paper (AREA)
- Image Analysis (AREA)
Abstract
Description
このため、紙にピッチ等の異物が付着する等の欠点が発生した場合、歩留まりが大きく低下することになる。
したがって、抄紙工程においては、欠点の発生を極力抑制することを目的として、様々な技術が開発されている。
これにより、撮像装置を設置等するために抄紙工程が行われる装置の内部に侵入する必要がなくなるので、欠点分類システムを用いるための準備やメンテナンスを極めて安全且つ簡単に行うことができる。
なお、欠点分類システムにおいては、ドライパートを経た紙を撮像の対象としているので、紙に発生した欠点は、紙料調成工程若しくは抄紙工程のドライパート以前のパートに起因するものである。
また、算出手段を備えるので、その欠点の特徴量から、欠点原因項目に対する確信度が算出される。
そして、表示手段により当該確信度が表示される。
これらのことにより、欠点分類システムにおいては、その欠点の欠点原因項目に対する確信度を認識することができる。
その結果、その欠点を確信度が高い欠点原因項目に分類することが可能となり、同時に、その欠点の原因を認識することが可能となる。
また、その欠点の欠点情報と、その欠点が分類された欠点原因項目とに基づいて、分類モデルの基準特徴量を変更することにより、分類の精度を向上させることができる。
これにより、分類モデルにより算出される確信度の精度がより一層向上する。
例えば、紙の長さ方向において、欠点の発生が繰り返されている場合、紙を案内するロールや、フェルト、カンバス等のループ状用具に異物が付着している恐れがあり、また、繰り返される欠点同士間の距離から、異物がロールに付着した場合、ロールの径、異物がループ状用具に付着した場合、ループ状用具における異物の位置を算出することができる。
なお、図面中、同一要素には同一符号を付すこととし、重複する説明は省略する。
また、上下左右等の位置関係は、特に断らない限り、図面に示す位置関係に基づくものとする。
更に、図面の寸法比率は図示の比率に限られるものではない。
なお、コンピュータは、1台に限定されず、複数台を有線又は無線により連結して用いてもよい。
上記抄紙機としては、いわゆる抄紙工程を行うことが可能な一連の装置であれば、特に限定されず、従来公知のものを適宜採用することができる。
欠点分類システムにおいて、対象となる紙は、抄紙工程により製造可能であれば特に限定されない。
例えば、当該紙としては、印刷用紙、新聞用紙、塗工用紙、包装用紙、薄葉紙、トイレットペーパーやティッシュペーパー等の家庭紙、雑種紙等のいわゆる洋紙や、段ボール原紙、白板紙、色板紙、紙管原紙、建材原紙、各種台紙等のいわゆる板紙等を採用することができる。
また、「異物」には、ピッチだけでなく、虫、紙粉、スライム等も含まれる。
なお、「ピッチ」とは、紙の原材料であるパルプに含まれる夾雑物のうち、粘着テープや糊に由来する粘着質のものや木材に由来する粘着質のものである。
また、「欠点原因項目」とは、欠点の原因となる事項を区別して項目としたものである。なお、「複数の欠点原因項目」とは、欠点原因項目の集合体であり、当該欠点原因項目の欠点の原因は互いに異なるものとなっている。
また、「確信度」とは、予測がどのくらい確実であるかの統計的な尺度(確率)である。
図1は、本実施形態に係る欠点分類システムの構成を説明するためのブロック図である。
図1に示すように、欠点分類システム10は、撮像装置にドライパートを経た紙を撮像させ、撮像された画像データを取得する撮像手段11と、画像データにおいて紙の欠点を検出するための検出手段12と、該欠点の特徴量を抽出するための抽出手段13と、該欠点の特徴量に基づいて、予め基準特徴量が設定された分類モデルに対して欠点原因項目の確信度を算出させる算出手段14と、確信度を表示するための表示手段15と、確信度が予め設定した所定値以上となる欠点原因項目に、欠点情報を分類する分類手段16と、画像データを格納するための記憶手段17とを備える。
欠点分類システムにおいては、これらの手段を備えることにより、欠点の欠点原因項目毎に対する確信度を認識することができる。
その結果、後述するように、その欠点を確信度が高い欠点原因項目に分類することが可能となり、同時に、その欠点の原因を認識することが可能となる。
撮像手段11は、有線又は無線のネットワークを介して、撮像装置と接続させている。そして、撮像装置に、ドライパートを経た紙を撮像する指令を発信し、撮像装置に撮像を開始させる。
一方で、撮像手段11は、撮像装置により撮像された画像データを撮像装置から受信する。
なお、画像データの解像度は、後述の検出される欠点のサイズの観点から、10~500MHzであることが好ましい。
そして、撮像手段11が受信した画像データは、記憶手段17に格納される。
図2は、本実施形態に係る欠点分類システムが用いられる抄紙工程、及び、撮像装置の設置位置を説明するための概略側面図である。
図2に示すように、抄紙工程は、抄紙機Mを用いて行われる。
抄紙工程は、水中にパルプが分散された分散液を抄紙用のワイヤーに載せ、余分な水を自然落下させるワイヤーパートW1、及び、これを一対のプレスロール間に通し、フェルトを介してプレスロールで押圧するプレスパートW2、からなるウェットパートW、カンバスKを介して、加熱されたドライヤーロールD1に接触させることで乾燥させるドライパートD、カレンダーパートC、並びに、リールパートRを有する。
なお、ドライヤーロールD1には、当該ドライヤーロールD1の表面に付着した異物を除去するためのドクターブレードD2が当接されている。
そして、ドライパートDを経た紙P2は、カレンダーパートCを経ることにより、紙P2の表面の凹凸が平滑化され、リールパートRで、紙P2がスプール等により巻き取られる。
これにより、紙P2は、カレンダーパートC及びリールパートRの間において、紙P2の表面が撮像装置Gにより撮像される。
このように、欠点分類システム10においては、撮像装置Gを設置等するために抄紙工程が行われる装置の内部に侵入する必要がなくなるので、準備やメンテナンスを極めて安全且つ簡単に行うことができる。
なお、上述したように、撮像装置Gが、ドライパートDよりも紙の搬送方向における下流側に設置されるので、画像データは、ドライパートDを経た紙P2が対象となっている。
すなわち、画像データにおいて検出された欠点は、紙料調成工程若しくは抄紙工程のドライパートD以前のパートで発生したものといえる。
これにより、欠点の座標から、欠点の原因となるパートの更に詳細な位置まで推測することが可能となる。
欠点の座標データは、例えば、紙の搬送方向をY軸、紙の幅方向をX軸とする。
ここで、原点とする位置としては、例えば、撮像装置Gによる撮像の開始位置に到達した紙P2の前端をY軸の原点(Y=0)とし、紙P2の左端をX軸の原点(X=0)とすればよい。
すなわち、紙Pの前端の左端を(X,Y)=(0,0)とすればよい。
そして、Yの値は、紙Pの長手方向に沿って、上流側に行くほど数字が大きくなり、Xの値は、左端から右端に行くほど値が大きくなるものとすればよい。
このとき、X軸方向における隣り合う座標同士間の距離が0.001~10mmであり、Y軸方向における隣り合う座標同士間の距離が0.001~10mmであることが好ましい。
この場合、紙P2に、目視できる範囲で、極めて小さい欠点が生じた場合であっても、確実に検出することが可能となる。
この場合、欠点の座標から、繰り返されている欠点同士の間の距離を算出することができる。
そして、異物がロールに付着した場合、その距離は、ロールの周長に相当することから、異物が付着したロールの径を推測することができ、異物がループ状用具に付着した場合、その距離から、異物が付着した位置を推測することができる。
すなわち、所定の座標において、数値化した値に異常が発生した場合、その座標に欠点が発生しているといえる。
具体的には、数値化が明暗度測定によるものである場合、上記映像をグレースケールにより白黒二値画像とし、その濃淡を例えば0~255の256段階に分けて数値化すればよい。
また、数値化がRGB測定によるものである場合、特定の色(例えば、青)を数値化すればよい。
このように、検出手段12は、明暗度測定又はRGB測定を採用しているので、欠点の有無を簡単に識別することができる。
なお、欠点の画像、検出手段12による座標データ及び欠点を数値化したデータは、欠点情報として記憶手段17に格納される。
抽出手段13において、特徴量を抽出する手法としては、フィルタ法、ラッパー法、組み込み法等を適宜採用することができる。
欠点分類システムにおいて、抽出される特徴量の具体例としては、数値化された欠点のサイズ、形状、最大値、最小値、濃淡差、濃淡等が挙げられる。
深層学習のモデルとしては、例えば、CNN(Convolutional Neural Network)モデルが用いられ、ネットワークモデルとしては、LeNet、AlexNet、VGG(VGG16,VGG19)、GoogLeNet、ResNet等を採用することができる。
なお、欠点の特徴量は、欠点情報として記憶手段17に格納される。
図3に示す「ピッチ」は、ドライパートにおけるロール由来又はカンバス由来のピッチの付着が欠点の原因であり、「ドクター」は、ドライパートにおけるドクター由来のピッチの付着が欠点の原因であり、「抄込み」は、紙料調成工程における異物の抄込みである。
また、「穴」は穴が欠点の原因であり、「油汚れ」は油の付着が欠点の原因であり、「虫」は虫の付着が欠点の原因であり、「耳切れ」は紙の端が切れたことが欠点の原因である。
なお、欠点原因項目と、その欠点原因項目に含まれる欠点の特徴量とのデータを新たに得た場合、分類モデルに、これを更に学習させることができる。すなわち、基準特徴量を変更することができる。
これにより、分類モデルが算出する確信度の精度がより向上する。
この場合、この項目への確信度を認識することにより、その欠点の原因が、ドライパートによるものなのか、ドライパート以外によるものなのかを認識することが可能となる。
なお、図3に示すピッチの付着はドライパートで発生し、抄込みは紙料調成工程で発生するものであるので、これらを区別することにより、欠点の原因が、紙料調成工程、又は、ドライパートによるものなのかを認識することが可能となる。
そして、算出された確信度は、表示手段15により表示される。
これにより、確信度が高い欠点原因項目や確信度が低い欠点原因項目を認識することが可能となる。
なお、表示手段15としては、モニター、タッチパネル等を採用すればよい。
これにより、その欠点が含まれる欠点原因項目を認識することが可能となる。
また、その欠点の欠点情報と、その欠点が分類された欠点原因項目とに基づいて、分類モデルの基準特徴量を変更することにより、分類の精度を向上させることができる。
なお、かかる分類の操作は、最大値の確信度が予め設定した所定値以下の欠点を、その最大値の確信度に基づいて、欠点原因項目へ分類し、その後、それが正しいか否かを人が追認することで行ってもよく、最大値の確信度が予め設定した所定値以下の欠点を、人が正しい欠点原因項目に分類することで行ってもよい。
これにより、分類モデルが算出する確信度の精度がより向上する。
なお、所定値は、任意に設定することができる。
図4は、本実施形態に係る欠点分類方法を示すフロー図である。
図4に示すように、欠点分類方法は、撮像装置Gにドライパートを経た紙を撮像させ、撮像された画像データを取得する撮像ステップS11と、画像データにおいて紙の欠点を検出するための検出ステップS12と、該欠点の特徴量を抽出するための抽出ステップS13と、該欠点の特徴量に基づいて、予め基準特徴量が設定された分類モデル20に対して欠点原因項目の確信度を算出させる算出ステップS14と、確信度を表示するための表示手段S15と、確信度が予め設定した所定値以上となる欠点原因項目に、欠点情報を分類する分類ステップS16とを有する。
そして、第1コンピュータC1と第2コンピュータC2とは、クラウド等を介して、接続されている。
したがって、欠点分類方法においては、撮像ステップS11及び検出ステップS12が第1コンピュータC1を用いて実効され、抽出ステップS13、算出ステップS14、表示ステップS15及び分類ステップS16が第2コンピュータC2を用いて実効される。
なお、かかる第2コンピュータC2に対し、撮像ステップS11等を行う第1コンピュータC1が複数接続されていてもよい。
また、算出ステップS13を有するので、その欠点の特徴量から、欠点原因項目に対する確信度が算出される。
そして、表示ステップS15により当該確信度が表示される。
これらのことにより、欠点分類方法においては、その欠点の欠点原因項目に対する確信度を認識することができる。
その結果、その欠点を確信度が高い欠点原因項目に分類することが可能となり、同時に、その欠点の原因を認識することが可能となる。
また、分類ステップS16を有するので、その欠点の欠点情報と、その欠点が分類された欠点原因項目とに基づいて、分類モデルの基準特徴量を変更することにより、分類の精度を向上させることができる。
例えば、紙P2の下方に設置されていてもよく、上方下方の両方に設置されていてもよい。
例えば、カレンダーパートC及びリールパートRは有していなくてもよい。
また、リールパートRの代わりに、紙を切断して回収する加工機等を設置してもよい。
例えば、ドライパートDとカレンダーパートCの間の紙P2の上方に設置してもよい。
また、撮像装置GがドライパートDの下流側に設置されていれば、更に別の撮像装置を備えていてもよい。
この場合、より詳細な分類が可能となる。
本発明の欠点分類システムによれば、準備やメンテナンスが簡単であり、且つ、欠点の原因を認識することが可能となる。
11・・・撮像手段
12・・・検出手段
13・・・抽出手段
14・・・算出手段
15・・・表示手段
16・・・分類手段
17・・・記憶手段
20・・・分類モデル
C・・・カレンダーパート
C1,C2・・・コンピュータ
D・・・ドライパート
D1・・・ドライヤーロール
D2・・・ドクターブレード
G・・・撮像装置
K・・・カンバス
M・・・抄紙機
P1・・・湿紙
P2・・・紙
R・・・リールパート
S11・・・撮像ステップ
S12・・・検出ステップ
S13・・・抽出ステップ
S14・・・算出ステップ
S15・・・表示ステップ
S16・・・分類ステップ
W・・・ウェットパート
W1・・・ワイヤーパート
W2・・・プレスパート
Claims (5)
- 抄紙工程におけるドライパートを経た紙の欠点に基づく欠点情報を、予め設定された欠点の原因に基づく複数の欠点原因項目のうち、相当する欠点原因項目に分類するための欠点分類システムであって、
撮像装置にドライパートを経た紙を撮像させ、撮像された画像データを取得する撮像手段と、
前記画像データにおいて紙の欠点を検出するための検出手段と、
該欠点の特徴量を抽出するための抽出手段と、
該欠点の特徴量に基づいて、予め基準特徴量が設定された分類モデルに対して前記欠点原因項目の確信度を算出させる算出手段と、
前記確信度を表示するための表示手段と、
を備え、
前記分類モデルが、予め蓄積された欠点の特徴量と前記複数の欠点原因項目との関係から、機械学習により、前記基準特徴量を学習させたものである欠点分類システム。 - 前記算出手段において、前記欠点原因項目毎の前記確信度がそれぞれ算出されるものであり、
複数の前記確信度のうち最大値の前記確信度となる前記欠点原因項目に、前記欠点情報を分類する分類手段を更に備える請求項1記載の欠点分類システム。 - 前記分類モデルが、前記最大値の確信度が予め設定した所定値以下の前記欠点の特徴量と、それを分類した前記欠点原因項目との関係から、機械学習により、更に学習させたものである請求項2記載の欠点分類システム。
- 前記欠点の欠点情報には、該欠点の特徴量に加え、前記紙に座標を設けた場合の該欠点の座標データが含まれる請求項1~3のいずれか1項に記載の欠点分類システム。
- 前記欠点原因項目が、少なくとも、前記ドライパートにおける異物の付着を欠点の原因とする項目を含む請求項1~4のいずれか1項に記載の欠点分類システム。
Priority Applications (6)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
EP22867339.8A EP4400832A1 (en) | 2021-09-07 | 2022-09-06 | Defect classification system |
CN202280060009.4A CN117916582A (zh) | 2021-09-07 | 2022-09-06 | 缺陷分类系统 |
AU2022342915A AU2022342915A1 (en) | 2021-09-07 | 2022-09-06 | Defect classification system |
KR1020247009317A KR20240054300A (ko) | 2021-09-07 | 2022-09-06 | 결점 분류 시스템 |
CA3230180A CA3230180A1 (en) | 2021-09-07 | 2022-09-06 | Defect classification system |
JP2023515256A JP7390085B2 (ja) | 2021-09-07 | 2022-09-06 | 欠点分類システム |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2021-145767 | 2021-09-07 | ||
JP2021145767 | 2021-09-07 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2023038017A1 true WO2023038017A1 (ja) | 2023-03-16 |
Family
ID=85507651
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2022/033356 WO2023038017A1 (ja) | 2021-09-07 | 2022-09-06 | 欠点分類システム |
Country Status (8)
Country | Link |
---|---|
EP (1) | EP4400832A1 (ja) |
JP (1) | JP7390085B2 (ja) |
KR (1) | KR20240054300A (ja) |
CN (1) | CN117916582A (ja) |
AU (1) | AU2022342915A1 (ja) |
CA (1) | CA3230180A1 (ja) |
TW (1) | TW202318224A (ja) |
WO (1) | WO2023038017A1 (ja) |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH05101164A (ja) * | 1991-09-13 | 1993-04-23 | Fujitsu Ltd | 疵検査装置 |
JPH10302049A (ja) * | 1997-04-28 | 1998-11-13 | Kumamoto Techno Porisu Zaidan | 画像識別装置および方法および画像識別装置を備えた画像検出識別装置ならびに画像識別用プログラムを記録した媒体 |
JP2018524804A (ja) * | 2015-06-05 | 2018-08-30 | ケーエルエー−テンカー コーポレイション | 反復的欠陥分類方法及びシステム |
US20190010005A1 (en) * | 2017-07-06 | 2019-01-10 | Honeywell Limited | Continuous Web Sheet Defect Analytics, Classification and Remediation for Enhancing Equipment Efficiency and Throughput |
JP2019159499A (ja) * | 2018-03-08 | 2019-09-19 | 株式会社Jvcケンウッド | 学習用データ作成装置、学習用モデル作成システム、学習用データ作成方法、及びプログラム |
JP2019212073A (ja) * | 2018-06-06 | 2019-12-12 | アズビル株式会社 | 画像判別装置および方法 |
JP6697132B1 (ja) | 2018-10-01 | 2020-05-20 | 株式会社メンテック | 監視システム |
JP2021025161A (ja) * | 2019-08-06 | 2021-02-22 | 東芝三菱電機産業システム株式会社 | データ推定制御装置 |
-
2022
- 2022-09-06 CN CN202280060009.4A patent/CN117916582A/zh active Pending
- 2022-09-06 CA CA3230180A patent/CA3230180A1/en active Pending
- 2022-09-06 KR KR1020247009317A patent/KR20240054300A/ko unknown
- 2022-09-06 AU AU2022342915A patent/AU2022342915A1/en active Pending
- 2022-09-06 WO PCT/JP2022/033356 patent/WO2023038017A1/ja active Application Filing
- 2022-09-06 JP JP2023515256A patent/JP7390085B2/ja active Active
- 2022-09-06 EP EP22867339.8A patent/EP4400832A1/en active Pending
- 2022-09-07 TW TW111133921A patent/TW202318224A/zh unknown
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH05101164A (ja) * | 1991-09-13 | 1993-04-23 | Fujitsu Ltd | 疵検査装置 |
JPH10302049A (ja) * | 1997-04-28 | 1998-11-13 | Kumamoto Techno Porisu Zaidan | 画像識別装置および方法および画像識別装置を備えた画像検出識別装置ならびに画像識別用プログラムを記録した媒体 |
JP2018524804A (ja) * | 2015-06-05 | 2018-08-30 | ケーエルエー−テンカー コーポレイション | 反復的欠陥分類方法及びシステム |
US20190010005A1 (en) * | 2017-07-06 | 2019-01-10 | Honeywell Limited | Continuous Web Sheet Defect Analytics, Classification and Remediation for Enhancing Equipment Efficiency and Throughput |
JP2019159499A (ja) * | 2018-03-08 | 2019-09-19 | 株式会社Jvcケンウッド | 学習用データ作成装置、学習用モデル作成システム、学習用データ作成方法、及びプログラム |
JP2019212073A (ja) * | 2018-06-06 | 2019-12-12 | アズビル株式会社 | 画像判別装置および方法 |
JP6697132B1 (ja) | 2018-10-01 | 2020-05-20 | 株式会社メンテック | 監視システム |
JP2021025161A (ja) * | 2019-08-06 | 2021-02-22 | 東芝三菱電機産業システム株式会社 | データ推定制御装置 |
Also Published As
Publication number | Publication date |
---|---|
CN117916582A (zh) | 2024-04-19 |
JPWO2023038017A1 (ja) | 2023-03-16 |
EP4400832A1 (en) | 2024-07-17 |
KR20240054300A (ko) | 2024-04-25 |
JP7390085B2 (ja) | 2023-12-01 |
AU2022342915A1 (en) | 2024-02-29 |
CA3230180A1 (en) | 2023-03-16 |
TW202318224A (zh) | 2023-05-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
EP4019948B1 (en) | A machine vision method and system for monitoring manufacturing processes | |
EP3580720B1 (en) | Real-time, full web image processing method and system for web manufacturing supervision | |
US10585046B2 (en) | Apparatus and method for tracking defects in sheet materials | |
CN102007367A (zh) | 幅材上的优先缺陷标记 | |
US10501274B2 (en) | Continuous web sheet defect analytics, classification and remediation for enhancing equipment efficiency and throughput | |
US20230169642A1 (en) | Inspecting Sheet Goods Using Deep Learning | |
US20050008765A1 (en) | Method for monitoring quality of paper web | |
US8506758B2 (en) | System for guiding web patching using a re-reeler | |
CN116770620A (zh) | 监控系统 | |
WO2023038017A1 (ja) | 欠点分類システム | |
JP6828652B2 (ja) | 衛生用紙の製造方法及び欠陥検査装置 | |
US20050254701A1 (en) | Method for time synchronization of information measured by means of imaging | |
JP2021091991A (ja) | 検査システム | |
JP2003213585A (ja) | 紙の品質モニタリング装置及び抄紙機並びに紙の製造方法 | |
JP4018962B2 (ja) | 帯状体の表面疵検査方法およびその装置 | |
JP4344312B2 (ja) | 帯状体又は柱状体の表面疵検査方法およびその装置 | |
US20240003677A1 (en) | A device and a method to measure and monitor physical properties of moving web of slit plastic film tapes | |
JP2006268142A (ja) | 用紙の検査方法及び検査装置 | |
CN117825387A (zh) | 一种ps版瑕疵检测系统、方法和设备 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
WWE | Wipo information: entry into national phase |
Ref document number: 2023515256 Country of ref document: JP |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 22867339 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: AU2022342915 Country of ref document: AU |
|
WWE | Wipo information: entry into national phase |
Ref document number: 202447011794 Country of ref document: IN |
|
WWE | Wipo information: entry into national phase |
Ref document number: 12024550512 Country of ref document: PH |
|
WWE | Wipo information: entry into national phase |
Ref document number: 3230180 Country of ref document: CA |
|
ENP | Entry into the national phase |
Ref document number: 2022342915 Country of ref document: AU Date of ref document: 20220906 Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 202280060009.4 Country of ref document: CN Ref document number: 2401001458 Country of ref document: TH |
|
WWE | Wipo information: entry into national phase |
Ref document number: 18689637 Country of ref document: US |
|
ENP | Entry into the national phase |
Ref document number: 20247009317 Country of ref document: KR Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2022867339 Country of ref document: EP |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
ENP | Entry into the national phase |
Ref document number: 2022867339 Country of ref document: EP Effective date: 20240408 |