JP6785634B2 - Non-woven fabric inspection equipment - Google Patents

Description

The present invention relates to a non-woven fabric inspection device.

A technique for inspecting a sheet material such as a non-woven fabric in a continuous production line is known.
For example, Patent Document 1 describes that a pattern matching method is used as a surface defect inspection device for a sheet material. Here, the latest pattern information is selected as a reference pattern from the pattern information obtained from the captured image of the sheet material, and the adjacent pattern information is compared with each other to detect surface defects. Patent Document 2 describes a method for determining a roll-induced defect in a processing apparatus that processes a web-like object while traveling it on a plurality of rolls. Further, Patent Document 3 describes that a hole of a material to be inspected such as a steel plate or paper is imaged and inspected by a transmission method.

Japanese Unexamined Patent Publication No. 7-61699 JP-A-2002-162364 JP-A-2009-103615

Nonwoven fabrics are often used in absorbent articles such as diapers. The non-woven fabric is subjected to various processing depending on the purpose. For example, from the viewpoint of enhancing the fit to the body, it is possible to apply a process of imparting elasticity to the non-woven fabric. The stretchable non-woven fabric is obtained, for example, by arranging a plurality of elastic members between two raw fabrics and integrating them, and stretching the non-woven fabric by meshing the tooth grooves between the pair of tooth groove rolls. Be done. Desirable stretchable physical properties can be obtained by adjusting the meshing depth between the tooth groove rolls and the method of applying tension.
Depending on the degree of the adjustment, defects such as holes in the non-woven fabric and breakage of fibers may occur. Depending on the degree of holes and cuts, sufficient product strength may not be obtained and product performance may not be satisfied. In addition, even if there is no problem immediately after processing, the broken part of the fiber may lead to a hole when the product is used. In a continuous production line, if the size and number of holes and the degree of fiber breakage cannot be accurately inspected, there is a possibility that some products should not be produced as a product. On the contrary, there is a problem that a non-defective product is judged to be a defective product and the yield is deteriorated. Therefore, it is desired from the viewpoint of product quality control to accurately inspect the size and number of holes in the non-woven fabric used and the degree of defects such as broken fibers.

However, in the non-woven fabric, holes and fiber breaks are likely to occur in a portion where the non-woven fabric is stretched by the stretching process and the fabric is thinned like uneven plow. Since the thinned part is relatively transparent compared to other parts of the non-woven fabric, it is difficult to distinguish it from a hole or a broken fiber. When the inspection is performed by image processing, it is necessary to accurately fine-tune the image processing such as noise removal on the image.

Further, the stretchable non-woven fabric is stretched in a long state, cut to a length corresponding to a product type such as a diaper, and incorporated into the product. In the continuous production line, the product type (type) may be switched (mold change), and in this case, the division length corresponding to the non-woven fabric product type changes, and the inspection area changes.
When the inspection area is changed, when the inspection is performed by the image processing, the image accuracy may be different, which may lead to a decrease in the inspection accuracy. Even if an attempt is made to improve the inspection accuracy by fine-tuning the image processing, there is a limit, and it is difficult to realize the required accurate inspection. In addition, fine adjustment of image processing requires time for setting, which may lead to a decrease in production efficiency.

In view of the above problems, the present invention relates to a non-woven fabric inspection device capable of accurately detecting defects even if the division length of the non-woven fabric corresponding to the product type changes without lowering the production efficiency.

The present invention is an inspection device that inspects the state of fibers of a long non-woven fabric, and is an imaging processing unit that captures and inspects an image of fibers on the processed surface of the nonwoven fabric, and an image capture signal to the imaging processing unit. The signal output unit has a signal output unit for periodically outputting the image and a switching unit for switching the captured signal of the image, and the signal output units are mutual according to the division length corresponding to the product type of the non-woven fabric. A plurality of signal transmitters having different numbers of signals are provided, and the switching unit provides a non-woven fabric inspection device that selects the signal transmitter according to the division length of the non-woven fabric.

The non-woven fabric inspection device of the present invention can accurately detect defects even if the division length of the non-woven fabric corresponding to the product type changes without lowering the production efficiency.

It is explanatory drawing which shows one preferable embodiment of the inspection apparatus of the nonwoven fabric of this invention. It is an image figure which shows typically the example of the image which image | imaged the fiber of the processed surface of the nonwoven fabric by the division length of the nonwoven fabric. It is a partially cutaway perspective view which shows a specific example of a long non-woven fabric to be inspected. It is a schematic diagram which shows the process of stretching a raw material composite nonwoven fabric by a tooth groove roll. It is an image figure which shows typically the example of the state of the fiber of the processed surface of the non-woven fabric which has been stretched. (A) is an enlarged image diagram schematically showing a portion where a fiber breakage to be detected occurs on the fiber-processed surface of the non-woven fabric shown in FIG. 5, and (B) is an image diagram schematically showing the non-woven fabric shown in FIG. It is an image figure which shows typically the state which the hole to detect is generated on the machined surface. (A) is a graph showing the average shade value of the original images of the three cases, and (B) is a graph showing the shade value deviation of the original images of the three cases shown in (A). (A) is a glass showing the area ratio of holes calculated by image processing the original images of the three cases of FIG. 7, and (B) changes the threshold value for the original images of cases II and III of FIG. It is a graph which shows the area ratio of a hole calculated by.

A preferred embodiment of the long non-woven fabric inspection apparatus according to the present invention will be described below with reference to the drawings.

As shown in FIG. 1, the non-woven fabric inspection device 10 of the present embodiment periodically receives an image capture signal to the image pickup processing unit 1 and the image pickup processing unit 1 for capturing and inspecting an image of fibers on the processed surface of the nonwoven fabric. It has a signal output unit 2 for outputting to, and a switching unit 3 for switching an image capture signal. The image pickup processing unit 1 is connected to the signal output unit 2 via the switching unit 3. This inspection device 10 is used to inspect the state of fibers of a long non-woven fabric.
The "fiber state" and "fiber" here refer to the state of the fibrous structure constituting the non-woven fabric, and for example, the state of the fibrous structure that affects the quality of the non-woven fabric such as the degree of entanglement of fibers, the orientation direction, and the density. .. From the state of such fibers, defects that are problematic in the quality of the non-woven fabric, such as holes and broken fibers, are inspected.

The image processing unit 1 controls the operation of the image pickup unit 11 that captures the processed surface of the non-woven fabric that is continuously conveyed based on the image capture signal output by the signal output section 2, and the operation of the image pickup section 11 to control the operation of the image pickup image. It has an image inspection unit 12 that performs processing and determines the quality of the processed surface of the non-woven fabric. The image pickup processing unit 1 further includes an illumination unit 13 and an illumination control unit 14.

The imaging unit 11 images the processed surface of the non-woven fabric that is continuously conveyed by the line scan method based on the image capture signal. The image capture signal is a pulse signal that is periodically output by the signal transmitter 21. The imaging unit 11 repeats scanning according to the number of pulse signals. Specifically, the imaging unit 11 repeatedly acquires an image (line scan) in the width direction (X direction) for the processed surface of the non-woven fabric that is continuously conveyed at a fixed imaging position by the number of image capture signals. .. At this time, the period (output interval) of the image capture signal is set in the signal output unit 2 described later so that the relationship with the conveying speed of the non-woven fabric is constant. As a result, the image pickup unit 11 captures an image having a length equal to the number of image capture signals on the processed surface of the non-woven fabric that moves in the transport direction (Y direction) at a predetermined speed.

A solid-state image sensor is used as the image sensor of the image pickup unit 11 so that image processing by digital processing can be easily performed. As the solid-state image sensor, it is preferable to use a color image sensor. The color image sensor is preferably an image sensor capable of expressing gradations of 128 gradations or more, and may be a charge coupling element (CCD) image sensor or a CMOS image sensor. Further, the gradation expression may be enhanced by using a plurality of pixels of 2 pixels or 4 pixels or more of the image sensor as one pixel. By enhancing the gradation expression, it becomes possible to recognize the boundary between the bright image and the dark image even when the contrast difference is small.

The image inspection unit 12 receives an image capture signal from the signal output unit 2 and issues an image pickup instruction to the image pickup unit 11. Further, the image inspection unit 12 receives and records the image captured by the image pickup unit 11, performs image processing necessary for the inspection, and determines the quality of the processed surface of the non-woven fabric. The contents of the necessary image processing and the quality judgment described above differ depending on the processing contents and inspection contents of the non-woven fabric, and can be appropriately set. The case of inspecting the state of the fibers of the non-woven fabric which have been subjected to the tooth groove stretching process, which is mentioned as a specific embodiment, will be described later.
The determination result is fed back to the non-woven fabric processing and manufacturing process by various methods. For example, the image inspection unit 12 may transmit a signal for automatically adjusting the processing conditions to the control unit of the processing line based on the data of the determination result. Further, the result of the pass / fail judgment may be displayed on the screen to issue a warning, and the machining line may be manually adjusted based on the warning.

The illumination unit 13 is arranged so as to illuminate the image pickup region by the image pickup unit 11 on the processed surface of the non-woven fabric. When inspecting the state of the fibers of the non-woven fabric that have been stretched in the tooth groove in the embodiment described later, it is preferable that the illumination unit 13 is arranged on the same surface side of the non-woven fabric as the imaging unit 11. That is, it is preferable to take an image by a light reflection method. The illuminance of the illumination unit 13 is adjusted by the illumination control unit 14.

The signal output unit 2 has a plurality of signal transmitters 21 and a drive unit 22 of the signal transmitter 21. The signal output unit 2 periodically outputs the above-mentioned image capture signal to the image pickup processing unit 1 by the signal transmitter 21.

A plurality of signal transmitters 21 are arranged in the signal output unit 2 according to the number of product types to be manufactured. In FIG. 1, two signal transmitters 21 (21A and 21B) are arranged corresponding to the two product types. However, the number of signal transmitters 21 is not limited to two, and a required number corresponding to the number of product types can be arranged.

The plurality of signal transmitters 21 have different numbers of image capture signals depending on the division length corresponding to the product type of the non-woven fabric to be inspected. Each signal transmitter 21 transmits a set number of image capture signals at regular intervals during a fixed period in which the non-woven fabric for the division length passes through the imaging unit 11. Each signal transmitter 21 has a period of outputting a set number of image capture signals at equal intervals as one cycle, and the signal output of the cycle can be repeated while the non-woven fabric is continuously conveyed. The signal transmitter 21 is preferably a rotary encoder.

The classified length of the non-woven fabric is a length that is cut according to the standard of the product type of the finished product in which the processed non-woven fabric is finally incorporated. The classified length of the non-woven fabric corresponds to the length at which the non-woven fabric continuously conveyed passes through the imaging unit 11 in a certain period of time.
The division length of this non-woven fabric changes depending on the product type switching. In the signal output unit 2 of the present embodiment, the transport speed of the non-woven fabric and the output cycle of the signal are set so that a set number of image capture signals are transmitted at equal intervals with respect to the changing division length of the non-woven fabric. The relationship is constant. In this relationship, the resolution of the image captured by the imaging unit 11 is determined by the division length corresponding to the product type of the non-woven fabric and the number of signals of the signal transmitter 21. The above "resolution" refers to pixel information per pixel, and is expressed as, for example, "μm / pixel". The smaller this value is, the finer the area of information is in one pixel, and the higher the accuracy of the image.

As described above, the signal output unit 2 includes a plurality of signal transmitters 21 having a number of signals corresponding to the division length of the non-woven fabric, corresponding to the product type. As a result, the signal output unit 2 can select the number of signals whose resolution of the acquired image is constant even if the product type is changed.

The term "constant resolution" as used herein means a range in which the resolution of each signal transmitter 21 is at an appropriate level that can withstand image processing for inspection. It means that the difference in resolution falls within the range of ± 5 μm / pixel. Further, the standard of the resolution to be "constant" is determined by the image accuracy required for the inspection content on the processed surface of the non-woven fabric. For example, in the case of inspecting the state of the fibers of the non-woven fabric that have been stretched in the tooth groove in the embodiment described later, the size of the holes and the cuts of the fibers generated by the stretched tooth groove is used as a reference. In this case, based on the distance between the cut fibers (for example, about 165 μm), which is smaller than the hole, the range of the resolution “constant” is set so as to be within the above range before and after the resolution (for example, 160 μm / pixel). Can be set.

2 (A) and 2 (B) show an image example of an image in which the processed surface of the non-woven fabric is captured by the division length of the non-woven fabric while the number of signals is constant (3600 pulses) as in the conventional case. FIG. 2 (A) shows a non-woven fabric division length (Y direction length) of 580 mm, which is required according to the product type standard, and FIG. 2 (B) shows a non-woven fabric division length (Y direction length). Is 660 mm. In addition, in FIG. 2, the rough state of the fiber of the portion stretched by the tooth groove stretching process is schematically shown, and the dense state of the fiber of the portion not stretched is omitted.
The inspection device 10 of the present embodiment includes a number of signals that keeps the resolution of the image constant according to the above-mentioned different division lengths, and can be appropriately selected. The following are examples. The inspection device 10 includes a signal transmitter 21A and a signal transmitter 21B corresponding to the above two division lengths. The signal transmitter 21A includes 3600 pulses of signals corresponding to the length of the non-woven fabric of 580 mm, and has a resolution of about 161 μm / pixel. The signal transmitter 21B includes 4000 pulses of signals corresponding to the length of the non-woven fabric of 660 mm, and has a resolution of about 165 μm / pixel. In this way, the signal transmitter 21A and the signal transmitter 21B have a number of signals that keep the resolution of the image constant.

As described above, the inspection device 10 can perform a highly accurate inspection on a product-by-product basis by performing an inspection based on an image having a constant resolution in units of division lengths of the non-woven fabric according to the product type. As a result, it is possible to accurately prevent the shipment of products that are determined to be defective. In addition, by accumulating the inspection results obtained for each product and various information accompanying the inspection results, it is possible to more quickly investigate the cause of the product judged to be defective in the manufacturing process.

The drive unit 22 is connected in the signal output unit 2 so as to drive all the signal transmitters 21. For example, when there are two signal transmitters 21A and 21B shown in FIG. 1, the drive unit 22 drives the signal transmitter 21A via the timing belt 23, and the timing belt 23 and the timing belt 24 are driven. The signal transmitter 21B is driven via the signal transmitter 21B.

The drive unit 22 rotates once each time the non-woven fabric is conveyed for the division length. The signal transmitter 21 rotates at a constant speed in accordance with the rotation of the drive unit 22, and outputs an image capture signal to the image pickup unit 11. At this time, only the signal transmitter 21 selected by the switching unit 3 described later is connected to the image inspection unit 12, and the signal transmitter 21 not selected is disconnected from the image inspection unit 12. Based on this, only the selected signal transmitter 21 (connected to the image inspection unit 12) operates, and the image capture signal to the image pickup unit 11 is output at a constant speed.
Further, the drive unit 22 receives a processing speed signal in the processing process of the non-woven fabric, and changes the rotation speed according to the received processing speed signal. As a result, each signal transmitter 21 determines the output interval so that the plurality of image capture signals are contained within the division length of the target non-woven fabric at equal intervals. Each signal transmitter 21 makes one rotation for each section length of the target non-woven fabric in accordance with one rotation of the drive unit 22.
The drive unit 22 is preferably an encoder servomotor.

The switching unit 3 selects the signal transmitter 21 from the plurality of signal output units 2 according to the division length of the non-woven fabric. That is, among the signal output units 2 set so that the resolution of the image is desired, the one suitable for the division length of the non-woven fabric corresponding to the product type is selected. As a result, the resolution of the image actually captured by the imaging unit 11 can be kept constant even if the product type is switched on the production line.

The inspection device 10 of the present embodiment may further include a signal control unit 4 to control the operations of the switching unit 3 and the drive unit 22 described above.
In this case, the signal control unit 4 is connected to the switching unit 3 and outputs the switching signal of the product type of the finished product manufactured on the continuous production line to the switching unit 3. The switching unit 3 determines the division length of the non-woven fabric corresponding to the product type switching signal, and selects the signal transmitter 21 corresponding to the division length of the non-woven fabric as described above. Further, the signal control unit 4 is connected to the drive unit 22 and outputs a drive command signal to the drive unit 22. The drive unit 22 drives the signal transmitter 21 described above based on the drive command signal. Further, the signal control unit 4 may receive the processing speed signal in the above-mentioned non-woven fabric processing process from the apparatus in the processing process and output it to the drive unit 33.
In the inspection device 10, the operation control of the switching unit 3 and the driving unit 22 is not limited to the case where the signal control unit 4 is used, but can be performed by various commonly used mechanisms. For example, each of the switching unit 3 and the driving unit 22 may be directly connected to an apparatus in the non-woven fabric processing process so as to operate in synchronization with the processing process.

The signal control unit 4 may have a mechanism usually used in a continuous manufacturing process. For example, it may be composed of a computer having a display unit and an input unit, an interface, a predetermined program for causing the computer to perform necessary operations, and the like. Alternatively, a programmable logic controller (PLC) may be used.

Examples of the inspection target of the inspection device 10 include various forms of non-woven fabric to which various processing is performed, and may be a single-layer non-woven fabric or a multi-layer non-woven fabric. Further, it may be a composite member in which another member is combined with the non-woven fabric.
A preferred embodiment is to inspect the tooth groove stretching processed state of the stretchable long non-woven fabric 70. The non-woven fabric 70 is made of a composite member in which a plurality of elastic members 63 are arranged between two non-woven fabrics 61 and 62 (see FIG. 3), and is stretched by meshing a pair of tooth groove rolls. .. As shown in FIG. 1, this tooth groove stretching process is performed by using the tooth groove rolls 81 and 82, the upstream in-feed nip rolls 86 and 86, and the downstream out-feed nip rolls 87 and 88 to combine the raw materials before processing. Tension is applied to the non-woven fabric 70A. The tooth groove extension is performed, for example, as shown in FIG. That is, the teeth 81A of the tooth groove rolls 81 and 82 and the tips of the teeth 82A push the raw material composite nonwoven fabric 70A in opposite directions, and the raw composite nonwoven fabric 70A is partially stretched in the region (83-84) between the tips of both teeth. Will be done.

As shown in FIG. 5, for example, as shown in FIG. 5, the processed surface of the non-woven fabric 70 that has been subjected to the tooth groove stretching process has a thin cloth in the stretched portion 71 and a transparent background, and the non-stretched portion 72 has dense fibers and the cloth is white. appear. Note that FIG. 5 schematically shows the rough state of the fibers in the stretched portion 71 on the processed surface of the non-woven fabric 70, and omits the dense state of the fibers in the unstretched portion 72. At this time, the elastic member 63 extending in the Y direction is shown only in the stretched portion 71 that is easily visible, and is omitted in the portion 72 in which the fibers are not densely stretched. In this regard, also in FIGS. 6A and 6B, the dense state of the fibers in the unstretched portion 72 is omitted.
Defects to be detected by inspection are broken fibers as shown in FIG. 6 (A) and holes as shown in FIG. 6 (B). From the viewpoint of the strength of the non-woven fabric, the holes to be defective are usually grasped as those having a thickness of 1 mm ² or more. Further, from the same viewpoint, the fiber breakage is usually grasped as having a thickness of about 165 μm. However, this value is an example, and the standard of the size to be a defect varies depending on the material and characteristics of the non-woven fabric to be processed, the degree of processing, the required performance, and the like.
Further, defects are not limited to holes and broken fibers, and may occur in various ways depending on the type of non-woven fabric to be inspected, the processing content, and the like.

Next, the inspection method using the inspection device 10 of the present embodiment will be described based on the inspection of the tooth groove stretching processed state of the nonwoven fabric 70 mentioned as the above embodiment.

In the inspection device 10, the imaging unit 11 of the imaging processing unit 1 is arranged on the downstream side of the outfeed nip rolls 74, 74, which is the outlet side of the tooth groove stretching process. The illumination unit 13 is arranged on the same surface side as the image pickup unit 11 with respect to the non-woven fabric 70. As a result, the imaging unit 11 images the processed surface of the non-woven fabric 70 that is continuously conveyed at the arranged fixed position by a light reflection method. The illumination unit 13 can set a preferable imaging environment by adjusting the irradiation angle and the illuminance.

The inspection device 10 identifies the type of product type in which the non-woven fabric 70 is incorporated in the signal control unit 4, and outputs a product type switching signal based on this to the switching unit 3. Based on the switching signal, the switching unit 3 selects a plurality of signal transmitters 21 (signal transmitters 21A and 21B in FIG. 1) corresponding to the division length of the non-woven fabric 70, and sets the image inspection unit 12 and the switching unit 3. Connect the wiring. For example, as described above, when the non-woven fabric section length is 580 mm, the signal transmitter 21A having a pulse number of 3600 is selected, and when the non-woven fabric section length is 660 mm, the signal transmitter 21B having a pulse number 4000 is selected. Can be selected. At this time, the resolution of the captured image based on the signal of the signal transmitter 21A is about 161 μm / pixel, and the resolution of the captured image based on the signal of the signal transmitter 21B is 165 μm / pixel, and both have a certain resolution. It becomes.

The signal control unit 4 outputs a drive command signal to the drive unit 22 of the signal output unit 2 at the same time as the start of the tooth groove stretching process of the nonwoven fabric 70. The drive unit 22 operates based on the drive command signal, and causes the signal transmitter 21 selected above to periodically output the image capture signal. At that time, the drive unit 22 receives the processing speed signal of the non-woven fabric directly or via the signal control unit 4, and determines the rotation speed according to the received processing speed signal.

In the image pickup processing unit 1, the image pickup unit 11 continuously images (line scans) the processed surface of the non-woven fabric 70 that is continuously conveyed based on the image capture signal periodically output by the signal transmitter 21. The captured image is recorded at any time in the image inspection unit 12.

The image inspection unit 12 sets a processing division window corresponding to the division length of the non-woven fabric for the recorded image. Specifically, a processing division window is set in which the length direction (Y direction) of the image is the division length of the non-woven fabric and the width direction (X direction) is the length matching the width of the non-woven fabric to be cut. A plurality of these processing division windows are continuously set in the length direction (Y direction) of the image. At that time, it is preferable to partially overlap each treatment division window so that defects (holes and broken fibers) on the sides of the treatment division windows can be inspected without being overlooked. In this case, it is preferable to set the length of the processing division window in the image inspection unit 12 and the number of signals of the signal transmitter 21 in consideration of the overlapping portion.

The image inspection unit 12 performs preprocessing on the image of the processing division window portion.
This pretreatment refers to a processing process on an image that enables appropriate identification of holes, broken fibers, etc. on the processed surface of the non-woven fabric 70. Specific processing includes shading and binarization processing in which a threshold value is set. The shading correction is a process for removing shading unevenness, and various commonly used methods can be adopted. The threshold value can be appropriately set based on the characteristics of the non-woven fabric so that a specific gradation or higher is not processed as 0 gradation (black) when performing the binarization processing.
In the non-woven fabric 70, there are stretched portions 71 (see FIG. 5) and unevenness in which the background is transparent, and these are regions where the fibers are thin. Since the imaging is performed by the light reflection method, the shade value of the image is smaller in the thin region of these fibers than in the non-stretched portion 72 in which the fibers are dense, similar to the region where the fiber is cut or the hole is cut. The threshold value in the shading correction and the binarization process is appropriately set so that the stretched portion 71, unevenness, and various other noises are not set to 0 gradation (black). Further, since the measurement of the size of each hole has a great influence on the inspection result, the pixel region having a low shading value as a hole is appropriately grasped by the above pretreatment.

In this way, the preprocessing is performed on the shading of the image. Therefore, if the accuracy of the information per pixel in the image does not satisfy the appropriate level as in the conventional case, proper preprocessing and proper inspection cannot be performed. In particular, when the product type is changed on the production line, if the division length of the non-woven fabric 70 corresponding to the product type changes, the resolution (pixel information per pixel) changes as described above. .. By changing the resolution, for example, the unevenness may be treated as 0 gradation and misidentified as a "hole", and a defect determination may be made even though it is not a defect. Alternatively, it is possible that holes and fiber breaks cannot be properly treated as 0 gradation, and what should be originally defective may be overlooked. Further, if binary processing is performed based on a value deviating from the reference shading average value due to a change in resolution, appropriate preprocessing cannot be performed.

On the other hand, when the inspection is performed using the inspection device 10, the signal transmitter 21 that makes the image resolution appropriate and constant is selected according to the division length of the non-woven fabric 70 corresponding to the product type. As a result, it is possible to acquire a highly accurate image and appropriately detect the number and size of holes and fiber breaks.
Further, in the inspection device 10, the setting for the fine adjustment is made because the signal transmitter is switched instead of the fine adjustment of the image processing as in the preprocessing described above according to the model change of the product type. It is possible to avoid wasting time and prevent a decrease in production efficiency. In addition, since the inspection device 10 sets the original pixel information of the captured image to an appropriate level by switching the signal transmitter, a processed image with higher accuracy can be obtained as compared with the case where the decrease in accuracy is compensated by fine adjustment of image processing. It is obtained and highly accurate inspection becomes possible.

The following is a comparison between the case where the above-mentioned product type switching is dealt with by switching the signal transmitter and the case where the image accuracy is supplemented by fine-tuning the image processing. Here, the description will be made based on the following three cases.
Case I: Non-woven fabric division length: 580 mm, signal transmitter pulse number 3600
Resolution: Approximately 161 μm / pixel
Case II: Non-woven fabric division length: 660 mm, signal transmitter pulse number 3600
Resolution: Approximately 183 μm / pixel
Case III: Non-woven fabric division length: 660 mm, signal transmitter pulse number 4000
Resolution: Approximately 165 μm / pixel

FIG. 7A shows the shading averages of the original images before the preprocessing for the above three cases. FIG. 7B shows the shading value deviation. Both are calculated with N = 10.
As can be seen from FIG. 7A, when the resolution is reduced (from about 161 μm / pixel to about 183 μm / pixel) due to the type change from Case I to Case II (division length 580 mm to 660 mm), the average shade value is 238.5. The gradation is as high as 240.5 gradations. Looking at this in terms of the shading deviation in FIG. 7B, Case II has a lower numerical value than Case I and there is no variation. That is, it can be seen that the number of pixels having a high shading value is increased as a whole due to the decrease in resolution due to the type change. That is, the number of pixels with low shading values is reduced. It is considered that this is because the pixel information of the portion thinly stretched by the tooth groove stretching including the hole (see the portion of reference numeral 71 in FIG. 5) is missing. In this case, it can be said that it is difficult to properly capture defects such as holes and perform binarization processing.
On the other hand, in Case III, the number of pulses of the signal transmitter is increased and the resolution is set to about 165 μm / pixel for the type change. In this case, the resolution is the same as that of Case I, and the difference between Case I and the average shade value is almost eliminated. In addition, the shading deviation is large and the accuracy of pixel information is improved. It can withstand the inspection of broken fibers.

FIG. 8A shows the number of holes detected based on a processed image obtained by performing image processing on the above original image. Specifically, the image processing is performed by performing shading correction or the like using an image processing machine "XG-8500" (trade name) manufactured by Keyence Corporation, setting a threshold value of 75, and performing binarization processing. In the processed image, the values obtained by dividing the pixels remaining as holes by binarization by the inspection area are shown as the amount of information of the holes in Cases I, II and III. Both were calculated with N = 10.
As can be seen from FIG. 8 (A), in Case II, the proportion of regions detected as holes is kept low due to the fact that the number of small shade values is smaller than that in Case I. This means that the image accuracy of Case II is lower than that of Case I due to the mold change, and the accuracy of product detection that should be judged as defective is lower.
On the other hand, in Case III, which corresponds to the mold change by changing the number of pulses, hole detection (1.6 times that of Case II) is possible, which is comparable to that of Case I, and the accuracy of product inspection is improved. doing.
Further, FIG. 8B shows a case where the inspection of Case II is handled by fine-tuning the image processing. At this time, Case III also shows the case where the same image processing is performed. Specifically, the threshold value is raised from 75 to 77 by two gradations to expand the range of processing as black pixels. These two gradations are set based on the difference in the average shade value between Case I and Case II in FIG. 7A. As can be seen from FIG. 8B, the accuracy and inspection accuracy of the processed image and the inspection accuracy of the case II even if the case II is dealt with by fine-tuning the image processing instead of switching the number of pulses as the correspondence from the case I. Cannot make up for the decline in In this regard, as shown in FIG. 8 (B), in Case III, the type change from Case I was performed by switching the number of pulses, so that the accuracy of the processed image and the accuracy of the processed image were changed even if the image processing conditions were changed. Inspection accuracy is maintained at a high level.

The image inspection unit 12 performs various data processing based on the preprocessed processed image, collates it with a preset standard, and determines the quality. For example, in the hole inspection, the area ratio of the holes in the processing division window portion, the maximum hole area, the number of holes, and the like are calculated, and the quality is judged based on whether or not these values exceed the standard. Further, in the inspection of fiber breakage, the continuous fiber breakage region is counted in the width direction, the ratio of the fiber breakage to the width direction is obtained, and the quality is judged based on whether or not this value exceeds the reference value. ..

The information of the inspection result by the image inspection unit 12 may be fed back to the signal control unit 4 as shown in FIG. 1, for example. In this case, when the product is defective, the signal control unit 4 may transmit an emission signal for discharging the product, or may stop the continuous production line depending on the content of the inspection result. Further, the information of the inspection result may be displayed on the screen to send a warning, or the information may be fed back to the manufacturing process to make fine adjustments. In addition, the inspection result information may be recorded in a specific recording device together with the time information, and may be used for investigating the cause in the manufacturing process for each defective product.

As described above, when the product type is changed in the non-woven fabric processing process as described above, the inspection device 10 determines the division length of the non-woven fabric corresponding to the product type switching signal, and the above-mentioned As described above, the signal transmitter 21 corresponding to the division length of the non-woven fabric is selected. Then, the connection between the signal transmitter 21 and the image inspection machine 1 is switched. As a result, even if the product type is switched, it is possible to acquire a certain level of image without delay and continue to perform high-precision inspection.

1 Imaging processing unit 11 Imaging unit 12 Image inspection unit 13 Lighting unit 14 Lighting control unit 2 Signal output unit 21 Signal transmitter 22 Drive unit 3 Switching unit 4 Signal control unit 10 Inspection device 70 Non-woven fabric

Claims (7)

An inspection device that inspects the condition of fibers in long non-woven fabrics.
Switching between an image pickup processing unit that captures and inspects an image of fibers on the processed surface of the non-woven fabric, a signal output section that periodically outputs an image capture signal to the image pickup processing unit, and a switching for switching the image capture signal. Has a part and
The signal output unit includes a plurality of signal transmitters having different numbers of signals depending on the division length of the non-woven fabric corresponding to the product type.
The switching unit is a non-woven fabric inspection device that selects the signal transmitter according to the division length of the non-woven fabric. The non-woven fabric inspection device according to claim 1, which inspects the state of the fibers of the non-woven fabric that have been stretched. The non-woven fabric inspection device according to claim 1 or 2, wherein the switching unit selects a signal transmitter having a number of signals that makes the image resolution constant with respect to the division length of the non-woven fabric. It has a signal control unit that outputs the switching signal of the product type.
The non-woven fabric inspection device according to any one of claims 1 to 3, wherein the switching unit selects the signal output unit based on the switching signal of the product type. The signal output unit has a drive unit connected to a plurality of the signal transmitters.
The non-woven fabric inspection device according to claim 4, wherein the signal control unit outputs a drive command signal to the drive unit, and the drive unit rotates once each time the non-woven fabric is conveyed by the division length. The non-woven fabric inspection device according to any one of claims 1 to 5, wherein the signal transmitter is a rotary encoder. The non-woven fabric inspection device according to any one of claims 1 to 6, wherein the image pickup processing unit captures an image by a light reflection method.

Images