JP7620985B2 - How to detect defects in cardboard sheets - Google Patents

Description

The present invention relates to a method for detecting defects such as stains, tears, and wrinkles in a cardboard sheet manufacturing line.

Corrugated cardboard sheets are generally manufactured by a device called a corrugator. In the corrugator, a single facer bonds a corrugated core to a back liner to produce a single-sided cardboard sheet, and a double facer bonds a single-sided cardboard sheet to a front liner or bonds multiple single-sided cardboard sheets to a front liner to produce cardboard sheets with a back liner on one side and a front liner on the other, such as double-sided or multi-sided cardboard sheets. The cardboard sheets produced are scored and cut parallel to the running direction in a slitter-scorer, and then cut to a specified length in a direction perpendicular to the running direction.

When cardboard sheets are manufactured in this manner, defects such as stains caused by the adhesion of machine oil or glue, tears or wrinkles in the back or front liner can occur. Conventionally, one method for detecting such defects has been to photograph the cardboard sheets with a digital camera and analyze the images.

Specifically, the image captured by the digital camera is binarized. In this process, a threshold value for detecting dark color defects that are darker than the background color of the cardboard sheet, such as stains caused by machine oil or shadows from wrinkles, and a threshold value for detecting light color defects that are lighter than the background color of the cardboard sheet, such as stains caused by glue adhesion, are set separately, thereby obtaining two types of binarized images. For example, as shown in FIG. 5(a), an image 100 including a dark color defect 101 is binarized, and pixels with a grayscale value equal to or greater than the threshold are converted to "1" and pixels with a grayscale value less than the threshold are converted to "0", thereby obtaining a binarized image 110 in which a black defect image 111 is displayed, as shown in FIG. 5(b). For light color defects, a binarized image in which the defect is displayed in black is obtained by converting pixels with a grayscale value equal to or greater than the threshold to "0" and pixels with a grayscale value less than the threshold to "1".

In the obtained binary image, if a pixel determined to be defective (a pixel converted to "0") is isolated, it is considered to be noise, and if the defective pixels are grouped together in a predetermined range or more, it can be determined that the defective pixel is one that should be detected. For example, as shown in FIG. 5(c), adjacent pixels among the defective pixels are grouped together. In this example, the pixels constituting each of the groups 201 to 204 are numbered 1 to 4 for identification, and the non-defective pixel 200 is numbered 0. For an image that has been processed in this way, the number of pixels constituting each group, or the length in the X direction, the length in the Y direction, or the area of each group when the two-dimensional coordinate is the XY coordinate, can be compared with a predetermined threshold value to determine that the cardboard sheet has a defect that should be rejected as a defective product. However, in the case of such processing, if the number of defective pixels or the number of groups to be divided are large, the grouping process and the determination process take a long time.

In addition, because the image has already gone through a binarization process, there is a problem in that defects (light defects) that are closer to the background color than the threshold value cannot be detected at the binarization stage. In FIG. 5(b), the light parts of the dark color defect 101 in FIG. 5(a) are shown by dashed lines as being missing in the binarized image 110.

Furthermore, in a cardboard sheet manufacturing line, a running cardboard sheet is held by being sandwiched from above and below by rotatable press rolls, so that a linear indentation caused by the pressure of the press rolls remains on the cardboard sheet. This indentation is called a "press mark." When a cardboard sheet with a press mark is photographed, an image is obtained in which the shadow of the indentation appears as a dark line. Therefore, as shown in FIG. 6(a), when an image 100 having a press mark 105 and a dark color defect 101 is binarized, as shown in FIG. 6(b), an image 115 of the press mark, which is not a defect, is included in the binarized image 110 together with a defect image 111. If a threshold value is set so that the image 115 of the press mark is not included in the binarized image 110, defects as dark as a press mark or lighter than a press mark cannot be detected.

In view of the above, the present invention aims to provide a method for detecting defects in cardboard sheets that can detect defects such as stains, tears, and wrinkles more accurately and in a shorter processing time on a cardboard sheet production line.

In order to solve the above problems, the method for detecting defects in cardboard sheets according to the present invention comprises:
"The back or front liner of the cardboard sheet traveling on the production line is photographed with a digital camera,
The obtained two-dimensional image is divided into a grid to set unit areas each consisting of a plurality of identical pixels;
For each unit area, a determination is made as to whether or not the cardboard sheet has a defect based on an addition process that calculates an addition process value that is either the total sum of the pixel gradation values or the average value obtained by dividing the total sum by the number of pixels.

Conventionally, images captured by a digital camera were binarized to determine whether or not there was a defect in the cardboard sheet, but the present invention determines whether or not there is a defect in the cardboard sheet without binarizing the image.

Specifically, a two-dimensional image obtained by a digital camera is divided into a grid pattern. The two-dimensional image may be an image captured by a two-dimensional camera, or may be an image captured by a one-dimensional camera synthesized into a two-dimensional image. Each unit area divided into a grid pattern is made to contain multiple and equal numbers of pixels. An addition process is then performed for each unit area. The addition process is a process of finding the total value of the gradation values of the pixels contained in the unit area, or a process of finding the average value by dividing the total value by the number of pixels. Whether the addition process value is the total value or the average value is standardized for all unit areas.

Because the summation process value (total value or average value) differs between normal unit areas with no defects and unit areas with defects, defects can be detected based on this difference. With conventional binarization methods, it was inevitable that some of the gradation information (shade information) would be lost due to the threshold setting, but with this invention, the information (gradation values) of all pixels is used in the summation process, allowing for more accurate judgments.

In addition, while the above-mentioned conventional technology uses pixels as the unit of processing, the present invention uses unit areas consisting of multiple pixels as the unit of processing, which allows for a short processing time. Furthermore, the above-mentioned conventional technology performs processing to group adjacent pixels that are determined to be defective together, so if there are a large number of pixels determined to be defective or a large number of groups to be divided, the grouping processing and judgment processing take a long time, whereas the present invention performs addition processing for each unit area, which has a constant area, so the time required for processing can always be constant regardless of the number of defects.

The method for detecting defects in cardboard sheets according to the present invention further comprises the steps of:
"The judgment is performed by comparing the added value of the unit area to be judged with the added value of the other unit areas."

The sum processing value of each unit area without defects will be the same. On the other hand, if there is a defect in a unit area, the sum processing value of that unit area will be a different value from the sum processing value of a unit area without defects. Here, in the manufacturing site of cardboard sheets, in most cases, normal cardboard sheets are manufactured without defects, and defects rarely occur. Therefore, the sum processing value of the unit area to be judged is compared with the sum processing value of the other unit areas. Usually, there is no difference between the two sum processing values, and it can be determined that there is no defect. On the other hand, if there is a difference between the two sum processing values, it is determined that one of the unit areas has a defect. If there are multiple other unit areas to compare the sum processing value of the unit area to be judged with, the presence or absence of a defect in the unit area to be judged can be more accurately determined.

The method for detecting defects in cardboard sheets according to the present invention further comprises the steps of:
"The cardboard sheet has linear press marks which are impressions made by a press roll, and in the judgment, the unit area to be judged and the unit area to be compared are set so that their positional relationship with the image of the press mark in the two-dimensional image is the same."

Although press marks are not defects, the addition processing values differ between unit areas with and without press marks, even if both are normal and defect-free. Therefore, the conditions related to press marks must be the same between the unit areas being compared. Press marks are impressions made by press rolls, and the positions where the press rolls are installed in the production line for corrugated board sheets and the positions where the camera takes pictures are known, so it is possible to know where the image of the press mark will appear in the image, and it is possible to set the unit area to be judged and the unit area to be compared to it so that their positional relationship with the image of the press mark in the two-dimensional image is the same. For example, since press marks are formed in a straight line as the corrugated board sheet travels, the unit area to be judged and the unit area to be compared to it are set so that they are on the same line in the direction of travel of the corrugated board sheet. Alternatively, if the unit areas divided into a lattice pattern are considered to be "rows" that are continuous in the direction of travel of the corrugated board sheet, and multiple rows are considered to be lined up in the width direction (direction perpendicular to the direction of travel) of the corrugated board sheet, it is possible to identify the rows that contain the image of the press mark and the rows that do not. In this way, when a unit area in which a press mark is present is judged, the sum processed value is compared with a unit area belonging to a column in which a press mark is also present, and when a unit area in which a press mark is not present is judged, the sum processed value is compared with a unit area belonging to a column in which a press mark is also not present, thereby making it possible to detect defects without being affected by the presence of press marks.

As mentioned above, conventional techniques that used binarization were unable to detect defects that were as dark as press marks or that were lighter than press marks, whereas the present invention is not affected by press marks and can therefore detect defects with greater accuracy than conventional image analysis using binarization.

As described above, the present invention provides a method for detecting defects in cardboard sheets that can detect defects such as stains, tears, and wrinkles with greater accuracy and in a shorter processing time on a cardboard sheet production line.

1 is a diagram showing the arrangement of lighting and cameras relative to a cardboard sheet in a defect detection method according to one embodiment of the present invention; 1 is a configuration diagram of a defect detection device used in a defect detection method according to an embodiment of the present invention. 4A is an explanatory diagram of a process for setting a unit area in a two-dimensional image in the defect detection method according to one embodiment of the present invention, and FIG. 4B is an explanatory diagram of an image after addition processing. 13A and 13B are diagrams illustrating the setting of another unit area to be compared with the unit area to be determined. 1A and 1B are diagrams illustrating a conventional defect detection method using binarization and problems associated therewith. 1A and 1B are diagrams for explaining the problem of press marks in a conventional defect detection method using binarization.

Below, a method for detecting defects in cardboard sheets, which is one embodiment of the present invention, and a defect detection device 20 used in this method are described with reference to Figures 1 to 4.

First, the configuration of the defect detection device 20 will be described. As shown in Figures 1 and 2, the defect detection device 20 includes a light 21 that projects light onto the cardboard sheet 30 as it travels along the production line, a camera 22 that captures the cardboard sheet 30 onto which the light is projected, and a computer 40 that determines whether or not the cardboard sheet has a defect based on the image captured by the camera 22.

The cardboard sheets to be subjected to defect detection can be double-sided or double-sided cardboard sheets in which one or more single-sided cardboard sheets and a front liner are bonded together, cardboard sheets with a back liner on one side and a front liner on the other side, or single-sided cardboard sheets in which a corrugated core and a back liner are bonded together. When a cardboard sheet with a back liner on one side and a front liner on the other side is to be subjected to the defect detection, the lighting 21 and the camera 22 are installed between the double facer and the slitter scorer in the production line. In this case, both a set of lighting 21 and camera 22 that projects light onto the back liner and takes an image, and a set of lighting 21 and camera 22 that projects light onto the front liner and takes an image can be provided. When a single-sided cardboard sheet is to be subjected to the defect detection, the lighting 21 and the camera 22 are installed between the single facer and the double facer, and light is projected onto the back liner and taken.

The lighting 21 projects light from a direction inclined with respect to the perpendicular to the back liner surface or the front liner surface of the cardboard sheet 30. This allows the shadow of a tear or wrinkle to be captured if the defect is one of the defects.

In this embodiment, a line sensor camera is used as the camera 22, and one-dimensional images are acquired in the width direction of the cardboard sheet, i.e., in the direction perpendicular to the running direction of the cardboard sheet (Y direction in the figure). The one-dimensional images acquired as the cardboard sheet runs are sent to the computer 40. The scanning period of the line sensor camera (the time it takes to store one line's worth of signals) is generally much shorter than the scanning period of a two-dimensional camera (area camera). One-dimensional images are continuously acquired as the cardboard sheet runs at high speed, and a two-dimensional image can be formed at high speed by arranging the acquired one-dimensional images. In addition, the camera 22 used has light receiving elements arranged at a length sufficient to capture one-dimensional images across the entire width of the cardboard sheet 30, even if the width length changes due to a change in the type of cardboard sheet 30.

The computer 40 has a hardware configuration including a storage device consisting of a main storage device and an auxiliary storage device, and a central processing unit (CPU) that performs processing according to the programs stored in the storage device.

The storage device stores an image processing program that causes the computer 40 to function as image processing means 41 that forms a two-dimensional image from the one-dimensional image acquired by the camera 22, and a judgment processing program that causes the computer 40 to function as judgment means 42 that judges whether or not the cardboard sheet has a defect based on the formed two-dimensional image.

The storage device can store one-dimensional image data transmitted from the camera 22, two-dimensional image data after image processing, information on unit areas set on the two-dimensional image, reference values such as thresholds used for judgment, information on defects detected as a result of the judgment, etc.

The defect inspection device 20 also includes an output device 25, such as a monitor or printer, that displays the process and results of processing by the computer 40, and an input device 24, such as a keyboard or pointing device, that inputs various commands and reference values such as thresholds to the computer 40. In addition, the defect inspection device 20 includes an alarm device 26 that notifies the occurrence of an abnormality based on the determination that a defect exists in the cardboard sheet 30.

The computer 40 of this embodiment is connected to the production management device 27 of the manufacturing line and the manufacturer's office computer 28 for wired or wireless communication. Information about the cardboard sheets, such as width and length, can be input from the office computer 28 to the computer 40, and the process and results of processing by the computer 40 can be sent from the computer 40 to the office computer 28. Alternatively, the configuration can be such that information about the cardboard sheets and the standard values for judgment are sent from the office computer 28 to the production management device 27, stored in the production management device 27, and then sent from the production management device 27 to the computer 40.

In the mechanical configuration of the production line, an encoder 23 is attached to a point that rotates in sync with the mechanism that drives the travel of the cardboard sheet. An electrical signal from the encoder 23 is sent to the computer 40, and the coordinates in the two-dimensional image are associated with the actual position on the cardboard sheet 30 based on the positional relationship between the point photographed by the camera 22 and the encoder 23.

Next, a method for detecting defects in cardboard sheets using the defect detection device 20 configured as described above will be described. In the method for detecting defects in cardboard sheets in this embodiment, light is projected from a light source 21 onto a cardboard sheet 30 traveling on a production line, a one-dimensional image taken by a camera 22 is converted into a two-dimensional image through image processing, image 1 of the cardboard sheet is obtained as shown in FIG. 3(a), and a determination is made as to whether or not the cardboard sheet has a defect through processing based on a determination processing program. Image 1 of the cardboard sheet can be, for example, an 8-bit (256 gradations) image.

Here, FIG. 3(a) shows a schematic diagram of a case where an image 1 of a cardboard sheet includes an image of a light-color defect 2 that is lighter than the background color of the cardboard sheet, an image of a dark-color defect 3 that is darker than the background color of the cardboard sheet, and an image of a press mark 5 that is a mark made by a press roll.

First, as a determination process, the image 1 of the cardboard sheet is divided into a grid pattern by multiple lines parallel to the running direction Y of the cardboard sheet and multiple lines perpendicular to this, as shown by the dashed lines in Figure 3(a). Each unit area 4 divided into a grid pattern contains multiple identical pixels.

Next, an addition process is performed to find an addition process value for each of the unit areas 4 partitioned into a grid pattern. The addition process value is the total grayscale value of the pixels contained in each unit area 4, or the average value obtained by dividing this total value by the number of pixels that make up the unit area. When the total value is used as the addition process value, the total value is found for all unit areas, and when the average value is used as the addition process value, the average value is found for all unit areas. By finding the addition process value in this way, the image 1 of the cardboard sheet is converted into a post-addition process image 10 as shown in Figure 3(b).

In the image 10 after the addition process, the addition process value of the unit area 4 without defects is different from the addition process value of the unit area 4 with a defect, but even in the unit area 4 without defects, the addition process value is different between the unit area 4 that contains the image of the press mark 5 and the unit area 4 that does not contain the image of the press mark 5. If we consider that the unit areas 4 are "rows" in which the unit areas 4 are continuous in the running direction Y of the cardboard sheet 30, and are lined up in multiple rows in the width direction of the cardboard sheet 30, in the image 10 after the addition process shown in Figure 3 (b), the unit areas 4 belonging to rows L1, L3, and L5 contain the image of the press mark 5, and therefore have a smaller addition process value than the unit areas 4 belonging to rows L2, L4, and L6 that do not contain the image of the press mark 5.

In addition, unit area 12(4) containing the image of light-colored defect 2 has a larger additive processing value than unit area 4 without a defect, and unit area 13a(4) and unit area 13b(4) containing the image of dark-colored defect 3 have a smaller additive processing value than unit area 4 without a defect.

The judgment of whether or not a defect exists in a certain unit area can be performed by comparing the sum processing value of the unit area to be judged with that of other unit areas. In this case, the other unit areas to be compared can be unit areas adjacent to the unit area to be judged. For example, when the unit area 4a(4) is judged as shown in FIG. 4(a), the sum processing value is compared with that of the adjacent unit areas 4b(4) and 4c(4) belonging to the same row L6. Since these three unit areas have the same positional relationship with the image of the press mark 5, the sum processing value is the same when there is no defect. Therefore, it can be judged that there is no defect in the unit area 4a(4) when the absolute values of the difference between the sum processing value of the unit area 4a(4) and the sum processing value of the unit area 4b(4), and the difference between the sum processing value of the unit area 4a(4) and the sum processing value of the unit area 4c(4) are smaller than the predetermined threshold value.

On the other hand, when a unit area having a defect such as unit area 12(4) is to be judged, the absolute value of the difference between the summed values of unit areas 4d(4) and 4e(4) on either side of the unit area belonging to the same column L2 is greater than or equal to a predetermined threshold, so that it can be judged that a defect exists in unit area 12(4).

Depending on the size of the unit area, the eight unit areas surrounding the unit area to be judged may be the same in terms of whether they contain an image of a press mark 5. In such a case, the difference between the summation value of the unit area to be judged and the summation value of each of the eight surrounding unit areas is calculated, and if the number of times that the absolute value is equal to or greater than a predetermined threshold (the number of combinations of the unit area to be judged and the surrounding unit areas) is smaller than a predetermined number, it is judged that there is no defect in the unit area to be judged, and if the number is equal to or greater than the predetermined number, it is judged that there is a defect in the unit area to be judged. In this way, by increasing the number of unit areas to be compared, it is possible to prevent the judgment result from being affected by a difference in the summation value between the two due to a defect in the unit area to be compared, even if there is no defect in the unit area to be judged.

Furthermore, depending on the relationship between the size of the unit area and the size of the defect, if a defect exists in a unit area to be judged, there is a high probability that a defect also exists in an adjacent unit area. In such a case, a unit area that is the same as the unit area to be judged in terms of whether it contains an image of a press mark 5 or not and is separated from the unit area to be judged is set as a comparison target based on a certain rule. An example of the rule in this case is "unit areas that belong to the same column and are separated by N units." For example, in FIG. 4(b), when a unit area 13b(4) having a defect is to be judged, the addition processing value is compared with that of unit area 4f(4) that belongs to the same column L4 and is separated, rather than the adjacent unit area 13a(4) being used as the comparison target. In this way, the defect can be accurately detected even when there is a high probability that the defect spans multiple unit areas.

As described above, the judgment process by comparing the sum processing value of the unit area to be judged with the sum processing value of other unit areas is based on the assumption that in most cases, normal cardboard sheets are manufactured without defects at the manufacturing site of cardboard sheets, and defects rarely occur, that is, the number of unit areas without defects is overwhelmingly greater than the number of unit areas with defects. If this assumption is not made, defects can also be detected by comparing the sum processing value of the unit area to be judged itself with a predetermined threshold value, without comparing it with other unit areas. In this case, the threshold value can be an upper threshold value and a lower threshold value obtained by plus or minus an allowable numerical range to the sum processing value of a normal unit area without a defect. In addition, a threshold value is separately determined for a normal unit area including an image of a press mark 5 and a normal unit area not including an image of a press mark 5, and the corresponding threshold value is used depending on whether the unit area to be judged is located in a position including an image of a press mark 5 or not.

In this way, by comparing the sum of the unit area being judged with the threshold value, it is possible to accurately detect defects, even when the defects are present over a relatively wide range.

In addition, when comparing the sum of the unit area being judged with the threshold in this way, if the sum is an average value rather than a total value, the same threshold can be used regardless of the size of the unit area.

When a unit area containing a defect is detected, the number of unit areas containing defects within a certain range set in advance on the cardboard sheet is counted. Then, if the number reaches or exceeds a predetermined threshold, that range on the cardboard sheet is determined to be a defective product containing an unacceptable defect, and a defect determination signal is sent from the computer 40 to the alarm device 26. The alarm device 26, which receives the signal, issues an alert by turning on or blinking a warning light or emitting an alarm sound. At the same time, the range of the cardboard sheet 30 that has been determined to be defective is identified based on the signal from the encoder 23, and is output from the output device 25.

As described above, the cardboard sheet defect detection method of this embodiment makes it possible to detect defects such as stains, tears, and wrinkles more accurately and in a shorter processing time on a cardboard sheet production line.

The present invention has been described above with reference to preferred embodiments, but the present invention is not limited to the above embodiments, and various improvements and design changes are possible without departing from the spirit of the present invention, as described below.

For example, the shape of the unit area set by dividing the two-dimensional image into a grid is illustrated as being square, but is not limited to this and can also be rectangular. The dimensions and shape of the unit area can be set by taking into consideration the balance between the time it takes for the Y-direction length of the two-dimensional image formed from the one-dimensional image captured by the line sensor camera as the cardboard sheet travels to become equal to the Y-direction length of the unit area, and the time required to perform the determination process across the entire width of the cardboard sheet.

Reference Signs List 1: Imaging of cardboard sheet 2: Light-colored defect 3: Dark-colored defect 4: Unit area 5: Press mark 10: Image after addition processing 21: Lighting 22: Camera 30: Cardboard sheet

Claims (1)

The back or front liner of the cardboard sheet traveling on the production line is photographed with a digital camera,
The obtained two-dimensional image is divided into a grid to set unit areas each consisting of a plurality of identical pixels;
For each of the unit areas, a determination is made as to whether or not the cardboard sheet has a defect based on an addition process that obtains an addition process value that is either a total sum of pixel gradation values or an average value obtained by dividing the total sum by the number of pixels ;
The cardboard sheet has linear press marks which are impressions made by a press roll,
When the determination is made by comparing the sum-processed value of the unit area to be determined with the sum-processed value of the other unit areas,
The unit area to be judged and the unit area to be compared are set so that their positional relationship with the image of the press mark in the two-dimensional image is the same.
A method for detecting defects in cardboard sheets.

Images