JP2011053173A - Defect detection method and defect detection device of thread - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a defect detection method can obtain existence or a kind of a defect of thread comprising a bundle of many continuous single fibers, on an online basis regardless of existence of twist of the thread, and performing speedily quality control of a thread package manufactured by a thread manufacturing process. <P>SOLUTION: In the method for illuminating the thread or a background to generate a brightness difference or a hue difference between the thread and the background, imaging the thread, and processing the obtained image, to thereby detect a defect of the thread, processing of the image and determination are performed by four steps, namely: (a) a binarization step of obtaining a binary image by processing the imaged image based on the brightness difference or the hue difference; (b) a thinning step of performing thinning processing of the binary image; (c) a separation step of separating a thread corresponding thin line from a defect candidate thin line, by removing a branch point of a thin line on the image subjected to the thinning processing; and (d) a determination process for determining match/mismatch of the defect and/or the kind of the defect, by comparing the characteristic quantity of the defect candidate thin line, with a threshold of a defect characteristic quantity set beforehand in a storage means. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a yarn defect detection method and a defect detection apparatus.

  In a manufacturing process of a yarn composed of a bundle of a large number of continuous single fibers such as synthetic fibers, defects such as fuzz and fluff may occur in the yarn due to a single fiber breakage or the like. Furthermore, the cutting of the yarn may occur when all the single fibers forming the yarn are cut.

  Many of these defects (abnormalities) that occur in yarn are based on fluctuations in various process conditions such as friction with equipment during the production of yarns, or fluctuations in tension and heat treatment temperature during the production process. Caused by external or internal structural changes in the yarn. In addition, after the fluff or fluff is once removed from the yarn and floated in the air, it may adhere to the yarn again during the yarn manufacturing process.

  Such yarn defects greatly affect the quality of the yarn itself and the quality of the fiber product formed from the yarn. Therefore, it is very important for the quality control of the yarn itself and the fiber product to accurately detect and grasp the yarn defect. In addition, if the condition of the running yarn in the yarn production process is constantly monitored and changes in the yarn status and frequent occurrences of defects due to process condition fluctuations can be grasped at an early stage, the yarn production process The yield is improved. That is, it is important to grasp the defects of the running yarn online.

  Many of the yarn manufacturing processes employ a method of simultaneously manufacturing a plurality of yarns in order to increase productivity. In such a manufacturing process, it is possible to simultaneously inspect a plurality of yarns, detect defects for each yarn, and grasp information on defects for each yarn, This is important in the quality control of the yarn package wound with the manufactured yarn.

  Conventionally, as a method for simultaneously and inexpensively inspecting defects of a plurality of yarns, a light projecting means is arranged on one side of the yarn, and a light receiving means is arranged on the opposite side across the yarn, so that There has been proposed a method in which a yarn is arranged at the upper part or the lower part, and the pill is detected by light shielding by the pill generated on the yarn (for example, see Patent Document 1). However, in the inspection method described in Patent Document 1, since it is impossible to determine which yarn defect between the light projecting unit and the light receiving unit is detected, it is not possible to obtain defect information for each yarn. Have the problem. Further, when the defect shape is flat and spreads thinly on the surface of the yarn, the amount of light shielding is small, and this inspection method has a problem that it is difficult to detect such a defect.

  On the other hand, as a method for inspecting a yarn defect in a yarn sheet in which a large number of yarns are arranged in parallel, there is a method of imaging a yarn using a camera and detecting a yarn defect by image processing. It has been proposed (see, for example, Patent Document 2). However, the inspection method described in Patent Document 2 has a problem that the number of yarns that can be inspected simultaneously is small due to the problem of resolution in the width direction of the camera. For example, it is difficult to inspect about 100 yarns traveling in parallel at the same time, and when a plurality of yarns are inspected at the same time, it is impossible to determine which yarn defect has been detected. There is a problem that it is impossible to obtain defect information for each article. Also, for example, when the yarn is a twisted yarn that has been twisted during the manufacturing process, the shape of the yarn is not uniform and the yarn has irregularities. Since a part of the light may be emitted in the same manner as the defective portion, there is a problem that the defective portion cannot be detected with high accuracy.

JP 61-114115 A JP 2004-277738 A

  The purpose of the present invention is to detect the presence or absence of defects in the produced yarn or the state of the defects by processing the data obtained by optical means regardless of whether or not the yarn is twisted. The object is to provide a detection method and a defect detection apparatus.

  Another object of the present invention is a yarn package formed by winding a fiber yarn comprising a bundle of a large number of continuous single fibers inspected by the yarn defect detection method or defect detection device of the present invention, Or it is providing the yarn package which attached the test result.

The yarn defect detection method of the present invention illuminates the yarn or background to cause a brightness difference or a hue difference between the yarn and the background, and then images the yarn together with the background and processes the obtained image. This is a method for detecting a defect of a yarn by performing image processing and determination in the following steps (a) to (d).
(A) a binarization step of processing a captured image based on a brightness difference or a hue difference to obtain a binarized image;
(B) a thinning step of performing thinning processing on the binarized image;
(C) a separation step of removing branch points of the fine lines on the thinned image and separating the yarn-corresponding fine lines and the defect candidate fine lines;
(D) A determination step of comparing the feature quantity of the defect candidate fine line with a threshold value of a defect feature quantity set in advance in the storage means to determine whether or not the defect is present and / or the type of the defect.

  In the yarn defect detection method of the present invention, in the determination step (d), the width of the defect candidate fine line in the direction orthogonal to the orientation direction of the yarn is measured as a feature amount and compared with a threshold value. It is preferable to determine one of the slackness of the yarn.

In the yarn defect detection method of the present invention, for the detected defects,
(E): You may have the recording step which records the image of the defective part determined to be a defect, or the measured value data of a defective part.

The yarn defect detection device of the present invention illuminates a yarn or background to cause a brightness difference or a hue difference between the yarn and the background, and under the illumination by the illumination device, the yarn is combined with the background. An image processing unit that captures an image and an image processing unit that processes the obtained image, and the image processing unit includes image processing and determination in the following stages (A) to (D): It is characterized by.
(A) A binarization step for processing a captured image based on a brightness difference or a hue difference to obtain a binarized image;
(B) A thinning step for thinning the binarized image;
(C) a separation step of removing branch points of the fine lines on the thinned image and separating the yarn-corresponding fine lines and the defect candidate fine lines;
(D) A determination step of comparing the feature quantity of the defect candidate fine line with a threshold value of a defect feature quantity set in advance in a storage unit and determining whether the defect is present and / or the type of the defect.

  In the yarn defect detection apparatus according to the present invention, the illumination unit is disposed on the side facing the imaging unit with the yarn interposed therebetween, and at a position where direct light from the illumination unit does not enter the imaging unit. It is preferable.

  The yarn package of the present invention is obtained by winding a plurality of yarns inspected by the yarn defect detection method or defect detection apparatus of the present invention as a package.

  The yarn package of the present invention may be attached to each package with the result of inspection by the above-described yarn defect detection method or defect detection device of the present invention.

  According to the yarn defect detection method and defect detection apparatus of the present invention, the presence or absence of defects or the state of defects in each of a plurality of yarns composed of a bundle of a large number of continuous single fibers is determined by whether or not the yarn is twisted. Regardless, it is captured online by processing data obtained by optical means. Therefore, by using the yarn defect detection method and apparatus of the present invention, it is possible to appropriately and quickly perform quality control of the yarn manufactured by the yarn manufacturing process.

It is the schematic side view which showed one form of the defect detection apparatus used for implementation of the defect detection method of the thread | yarn of this invention. It is the schematic side view which showed another one form of the defect detection apparatus used for implementation of the defect detection method of the thread | yarn of this invention. It is the schematic side view which showed another one form of the defect detection apparatus used for implementation of the defect detection method of the thread | yarn of this invention. It is the schematic side view which showed another one form of the defect detection apparatus used for implementation of the defect detection method of the thread | yarn of this invention. It is a schematic diagram of the image obtained by binarizing the image imaged by the imaging means in the defect detection device used for carrying out the yarn defect detection method of the present invention. FIG. 6 is a schematic diagram of an image obtained by thinning the image shown in FIG. 5 in the defect detection apparatus used for carrying out the yarn defect detection method of the present invention. FIG. 7 is a schematic diagram of an image showing a branch point of a line in the image shown in FIG. 6 in the defect detection device used for carrying out the yarn defect detection method of the present invention. FIG. 8 is a schematic diagram of an image obtained by erasing a branch point in the image shown in FIG. 7 in the defect detection apparatus used for carrying out the yarn defect detection method of the present invention. FIG. 3 is a schematic diagram of an image obtained by separating a yarn-corresponding fine line and a defect candidate fine line and extracting only the defect candidate fine line in the defect detection apparatus used for carrying out the yarn defect detection method of the present invention. . In the defect detection apparatus used for carrying out the yarn defect detection method of the present invention, an image obtained by imaging a yarn group including yarn slack by an imaging means, binarizing the captured image, and thinning the image FIG. It is a schematic side view of the defect detection apparatus used for implementation of the yarn defect detection method of the present invention.

The yarn defect detection method of the present invention illuminates the yarn or background to cause a brightness difference or a hue difference between the yarn and the background, and then images the yarn together with the background and processes the obtained image. This is a method for detecting a defect of a yarn by performing image processing and determination in the following steps (a) to (d).
(A) a binarization step of processing a captured image based on a brightness difference or a hue difference to obtain a binarized image;
(B) a thinning step of performing thinning processing on the binarized image;
(C) a separation step of removing a branch point of the fine line on the thinned image and separating the yarn-corresponding fine line and the defect candidate fine line;
(D) A determination step of comparing the feature quantity of the defect candidate fine line with a threshold value of a defect feature quantity set in advance in the storage means to determine whether or not the defect is present and / or the type of the defect.

  Examples of the yarn to which the yarn defect detection method and apparatus of the present invention are applied include synthetic fiber yarns such as acrylic fiber, polyamide fiber, polyester fiber, and aramid fiber, and fiber yarns such as glass fiber and carbon fiber. There is.

  Since the yarn defect detection method of the present invention can easily expand the brightness difference between the yarn portion and the background portion formed between the yarns, the yarn defect detection method of the present invention is made of a material. It can also be applied to inspection of yarns with low lightness by themselves or by being processed such as dyeing. In addition, since the yarn defect detection method of the present invention is hardly affected by unevenness in brightness on the surface layer of the yarn, it can also be applied to yarn twisted in the production process.

  The yarn to which the yarn defect detection method and apparatus of the present invention is applied is composed of a bundle of many continuous single fibers. Usually, the yarn is preferably composed of about 10 to 1,000,000 single fibers.

  The yarn defect detection method and apparatus of the present invention (hereinafter sometimes simply referred to as the method of the present invention) should be used in various yarn manufacturing processes that require understanding of yarn defects. Can do. As such a manufacturing process, in the manufacturing process of a synthetic fiber, there exist an oil agent provision process, an extending process, a heat treatment process, etc., for example. When the intermediate product is sent to the next process, in the final step during the manufacturing process of the intermediate product, and when the product is shipped to the customer, the yarn according to the method of the present invention is used in the final process during the manufacturing process. It is preferable to inspect the presence or absence of defects and the state of defects.

  In the present invention, the yarn or background is illuminated to produce a brightness difference or hue difference between the yarn and the background, and then the yarn is imaged together with the background, and the obtained image is processed. That is, by illuminating the yarn or background with illumination to make it brighter, the brightness difference or hue difference from the unilluminated background or yarn is increased, and the yarn is imaged. Specifically, lighting the yarn brightens the yarn, causing a difference in brightness or hue from the unilluminated background, and conversely, illuminating the background and unilluminated yarn. Can produce a lightness difference or a hue difference.

  In the case of illuminating the yarn, the illumination means is not limited in both intensity and wavelength of the illumination light as long as sufficient reflected light or scattered light from the defect generated on the yarn can be obtained. In particular, when simultaneously inspecting defects of a plurality of yarns at a place where a plurality of yarns are running in parallel, each yarn is arranged in a direction across the row of parallel yarns, that is, in the width direction. Any device that can illuminate uniformly is acceptable. It is preferable that the illumination means at this time can illuminate each yarn with a light amount difference within 20% in the width direction.

  As the illumination means for illuminating the yarn, a high-frequency fluorescent lamp, a metal haloloid lamp, LED illumination, or the like can be used. Further, the light from a light source such as halogen or LED may be guided by an optical fiber for illumination. In particular, when a plurality of yarn defects are simultaneously inspected in a place where a plurality of yarns run in parallel, it is preferable to use a high-frequency fluorescent lamp or LED illumination that is long in the width direction from the viewpoint of cost and maintainability.

  As for the positional relationship between the illumination and the yarn when illuminating the yarn, a position where only the yarn can be illuminated and the brightness difference or hue difference from the background can be increased is preferable. For example, as shown in FIG. There is a method of illuminating the yarn 1 with the illumination means 2 from an oblique direction indicated by the arrow 5 from the same surface side. In this case, it is preferable not to place anything in the field of view of the imaging means on the opposite side of the imaging means 3 across the yarn 1 in order to avoid the background being illuminated by the reflection of the illumination means 2. When some device is arranged due to the installation space or the like, a light absorbing plate may be placed between the device and the yarn to prevent reflection of illumination light from the device. As the light absorption plate, for example, a black acrylic plate whose surface is not a mirror surface can be used. Further, as shown in FIG. 2, the illuminating unit 2 may illuminate the yarn 1 from the side opposite to the imaging unit 3 with the yarn 1 interposed therebetween. In this case, if the illumination light is directly incident on the imaging means 3, the background is illuminated more strongly than the yarn 1, so that the illumination means 2 is not directly incident on the imaging means 3, and It is preferable that the scattered light or diffracted light from the yarn or the defect is arranged at a position where it can enter the imaging means 3. Further, in order to prevent direct illumination light from entering the image pickup means, a slit or a lens may be placed on the front surface of the illumination to prevent the illumination light from spreading.

  On the other hand, when illuminating the background, the illumination means is sufficient for detecting defects by the illumination light being incident on the imaging means and the illumination light being shielded by the yarn and the defects generated on the yarn. The intensity and wavelength of the illumination light are not limited as long as a brightness difference or hue difference between the yarn and the background can be obtained. As in the case of illuminating the yarn, when simultaneously inspecting defects of a plurality of yarns in a place where a plurality of yarns are running in parallel, in the direction across the row of parallel yarns, that is, Any device that can uniformly illuminate in the width direction may be used. The illumination means at this time is preferably capable of illuminating with a light amount difference within 20% in the width direction. As specific illumination means, a high-frequency fluorescent lamp, a metal haloloid lamp, LED illumination, or the like can be used as in the case of illuminating the yarn. Further, the light from a light source such as halogen or LED may be guided by an optical fiber for illumination. In particular, when a plurality of yarn defects are simultaneously inspected in a place where a plurality of yarns run in parallel, it is preferable to use a high-frequency fluorescent lamp or LED illumination that is long in the width direction from the viewpoint of cost and maintainability.

  As for the positional relationship between the illumination and the yarn when illuminating the background, for example, as shown in FIG. 3, the imaging means 3, the field of view of the imaging means on the yarn 1, and the illumination means 2 are arranged in a straight line. Is preferred. If the intensity of the illumination light is too strong and transmitted light from the yarn or defect enters the imaging means, or if the yarn or defect is not normally imaged by diffracted light from the yarn or around the defect, the yarn A translucent diffusion plate may be disposed between the strip and the illumination means to reduce the intensity of the illumination light, or the reflector 6 may be viewed as shown in FIG. 4 by the imaging means 3 and the imaging means on the yarn. And the background may be indirectly illuminated by illuminating the reflector 6. A translucent acrylic plate or the like can be used as the translucent diffusion plate, and a white acrylic plate that is not a mirror surface can be used as the reflection plate.

  As described above, there are several applicable configurations for the positional relationship among the illumination means, the imaging means, and the yarn. Especially when the yarn has a certain degree of transparency, the transmitted light from the yarn is used. Since defects on the surface or inside of the yarn may be imaged as a change in the amount of transmitted light, the yarn is illuminated from the opposite side of the imaging means with the yarn sandwiched as shown in FIG. Most preferably, the light is not incident on the image pickup means and scattered light or diffracted light from the yarn or the defect is arranged at a position where the light can enter the image pickup means.

  As an image pickup means for picking up an image of the yarn along with the background, a light receiving element (pixel) such as a CCD that receives light is linearly or two-dimensionally arranged and a sensor for obtaining data on brightness or hue is adopted. Can do. In particular, in the case of simultaneously inspecting defects of yarns running in parallel over a wide range, a line sensor camera that has excellent resolution in the width direction and can perform a wide range of inspection is preferable.

  The light receiving sensitivity of the imaging means is preferably about 10 to 1,000 V / lx · s so that transmitted light or scattered light from the yarn or the defect can be received when the defect is detected. Specifically, manufacturer products such as Nippon Electrosensory Device Co., Ltd. and Keyence Corporation can be used.

In the method of the present invention, the step of processing the image obtained by the imaging means by the data processing means includes the following steps (a) to (d).
(A) a binarization step of processing a captured image based on a brightness difference or a hue difference to obtain a binarized image;
(B) a thinning step of performing thinning processing on the binarized image;
(C) a separation step of removing branch points of the fine lines on the thinned image and separating the yarn-corresponding fine lines and the defect candidate fine lines;
(D) A determination step of comparing the feature amount of the defect candidate thin line with a threshold value of a defect feature amount set in advance in a storage unit and determining whether the defect is present and / or the type of the defect.

  In the binarization stage of (a), the image obtained by the imaging means is binarized with a preset threshold value. For example, the background is white as shown in FIG. , Fuzz defect 8 and the like) are distinguished as black. At this time, binarization may be performed so that the background is black and the yarn and defective portions are white. The threshold for binarization is determined after investigating values that can distinguish the yarn and the defect from the background in the image obtained by imaging the yarn illuminated by the configuration of the illumination unit and the imaging unit in advance. Is done.

  In the thinning step (b), a process of thinning a portion including a yarn and a defect is performed on the binarized image. In the thinning process, as described in Digital Image Processing (issued by the Foundation for Image Information Education (CG-ARTS Association) July 2004), pixels are trimmed while maintaining the connectivity of the connected components in the image. The process is repeated, and the process of removing pixels is performed until the width becomes one pixel. For example, when thinning processing is performed on the image of FIG. 5, an image as shown in FIG. 6 is obtained. At the time of thinning, a defect having a large area and a complicated shape as indicated by reference numeral 7 in FIG. 6 is not converted to a single line after thinning processing, but is converted into a figure in which a plurality of lines are intricate. There are many.

  In the image binarized in the binarization stage, the yarn and the defect have a width, and the width and shape of the yarn are twisted or shaken during running or mechanical vibration of the device. There are cases where the image fluctuates, the shape thereof is uneven, or the image is slightly meandered. As described above, when the yarn that is originally a normal portion is captured while being fluctuated in the image, it is difficult to extract a defective portion generated in the yarn with high accuracy as it is. Therefore, in order to extract a defective portion with high accuracy without being affected by fluctuations in the yarn, the yarn is converted into a line having a width of one pixel by thinning the yarn. At this time, even if the width of the yarn fluctuates and is captured, it is converted into a line of one pixel width by the thinning process, so that the normal portion of the yarn is equal and long in the yarn running direction. Therefore, it is easy to distinguish from the defective portion.

  In the separation step (c), a process of separating the yarn converted into a long line in the yarn running direction and the defective portion in the thinned image is performed. At the time of separation, firstly, as shown by reference numeral 9 in FIG. 7, an intersection point between the yarn candidate thin line and the defect candidate thin line, in other words, a branch point where a long line in the yarn running direction branches in the middle of the defective portion is extracted. As a branch point extraction method, as described in the digital image processing (issued by the Foundation for Image Information Education (CG-ARTS Association) July 2004), a predetermined 3 × 3 logic filter is used. A method of extracting a thinned image by performing a raster scan is known.

  After the branch point is extracted, the branch point portion is deleted from the thinned image to obtain an image in which the yarn-corresponding fine line and the defect candidate fine line are separated as shown in FIG. Here, by deleting the yarn-corresponding fine line, an image in which only the defect candidate fine line is extracted as shown in FIG. 9 can be obtained. A specific method for deleting the yarn-corresponding thin line is performed by designating only the yarn-corresponding thin line by using a characteristic feature of the yarn-corresponding thin line. Specific features of the yarn-corresponding fine line include, for example, “long in the yarn running direction in the image” and “exists in parallel with the yarn running direction in the image”. There is no particular limitation as long as the fine line corresponding to the yarn can be designated.

  When the yarn-corresponding thin line is separated, the obtained defect-only line is a line that is interrupted as shown by reference numerals 7 and 8 in FIG. 8 because the branch point with the yarn-corresponding thin line is deleted. It has become. Since it may be impossible to accurately measure the extracted defect size and other features if the middle of the line is interrupted, the broken candidate thin line should be stored and connected. Is preferred. Specifically, there are a method of expanding and connecting discontinuous lines, and a method of connecting by deleting a deleted branch point portion to a defect candidate thin line.

  In the determination step (d), after obtaining an image in which only defect candidate fine lines are extracted as shown in FIG. 9, the feature quantity of the extracted defect candidate fine lines is compared with a threshold value of a defect feature quantity preset in the storage means. Then, it is determined whether or not there is a defect. The feature amount of the defect candidate thin line to be compared with the threshold value of the defect feature amount includes “the diameter of the smallest circle circumscribing the defect candidate thin line”, “the smallest rectangle size circumscribing the defect candidate thin line”, “defect candidate thin line” Area ”,“ ferret diameter ”and the like.

  In addition, some of the defects include yarns caused by fluctuations in tension and heat treatment temperature during the production process, in addition to fuzz or fluff, etc., caused by cutting part of the single fiber group constituting the yarn. A part of the single fiber group constituting the strip may sag, and may occur on the yarn as a yarn sag as indicated by reference numeral 10 in FIG. The slack of the yarn may be treated as a defect because it causes the quality of the yarn to deteriorate as with the fluff or fluff. Moreover, since the cause of occurrence of slack yarn may be different from that of fluff or fluff, it may be detected separately from fluff or fluff. As a detection method, as shown in FIG. 9, by comparing the defect candidate fine line with the feature amount of the thread slack for the image obtained by extracting only the defect candidate fine line, whether the defect candidate fine line is fluff or fluff, Alternatively, it is possible to determine whether the yarn is slack and it is possible to determine the type of defect, which is preferable. The feature amount of thread slack includes the direction of the defect candidate fine line, the number of contact points between the defect candidate fine line and the yarn, the distance between the defect candidate fine line and the yarn, and the orientation of the yarn of the defect candidate fine line. The width of the direction orthogonal to the direction "or the like can be used. In particular, it is preferable to use the feature of “width in a direction perpendicular to the orientation direction of the yarn of the defect candidate fine line” because yarn slack can be distinguished accurately and easily.

  In addition, for yarns that have been determined to be defective at the determination stage, the position of the defect is calculated and compared with the yarn position recorded in advance, so that even when multiple yarns are simultaneously inspected, It is possible to specify whether or not a defect has occurred.

  Moreover, it is preferable to have a recording step of recording an image of the defective portion or measurement value data of the defective portion for those determined to be defective in the determination step. As a result, the tendency and characteristics of the generated defect can be easily grasped, and a quick and accurate defect source countermeasure can be taken. In addition, by recording defect data for each yarn over time, for example, by comparing the recorded defect data of the recorded yarn with the criteria for determining the quality of the yarn, the yarn and the yarn were wound. The quality of the yarn package can be judged and the quality can be controlled. A package in which a yarn having such an inspection result is wound has not only the presence / absence of a defect but also the number and position thereof from the inspection result. In addition, in the processing process of the textile product using the package, it is possible to remove the defective part or use the inspection result as a quality guarantee to the customer.

The yarn defect detection apparatus of the present invention also includes an illumination unit that illuminates the yarn or background to cause a brightness difference or a hue difference between the yarn and the background, and the yarn together with the background under illumination by the illumination unit. An image processing unit that captures the image and an image processing unit that processes the obtained image. The image processing unit includes image processing and determination in the following stages (A) to (D): It is characterized by.
(A) A binarization step for processing a captured image based on a brightness difference or a hue difference to obtain a binarized image;
(B) A thinning step for thinning the binarized image;
(C) a separation step of removing branch points of the fine lines on the thinned image and separating the yarn-corresponding fine lines and the defect candidate fine lines;
(D) A determination step of comparing the feature quantity of the defect candidate fine line with a threshold value of a defect feature quantity set in advance in a storage unit and determining whether the defect is present and / or the type of the defect.

  Note that the image processing means at the stages (A) to (D) are the same as (a) to (d) of the defect detection method, and thus description thereof is omitted.

  In the yarn defect detection apparatus of the present invention described above, the illumination means is on the side facing the image pickup means across the yarn and at a position where direct light from the illumination means does not enter the image pickup means. It is preferable that it is installed.

  The yarn package of the present invention is formed by winding a plurality of yarns inspected by the yarn defect detection method or defect detection apparatus of the present invention as a package.

  Moreover, it is preferable that the yarn package of this invention attaches the result inspected by the defect detection method or defect detection apparatus of the said invention to the said package.

  The attached inspection results are information such as the presence / absence of defects in the package, the number of defects, the position of the defects, the size of the defects, the type of defects, etc. In addition to removing defects when unwinding the package in a later process, it can also be used for quality assurance of the package.

[Example 1]
In Example 1, it implemented with the structure of FIG. In addition, the schematic which looked at the structure of FIG. 2 from diagonally upward is shown in FIG. As shown in FIG. 11, a plurality of yarns were run in the camera field of view and illumination range, and the inspection was performed. The equipment configuration is as follows.

Illumination means: White bar LED (width 900mm, 91W)
Distance α between illumination means and yarn: 100 mm
Imaging means: Line sensor camera NED SU74 (7450 pixels)
Line camera shutter speed: 0.9 ms
Distance β between imaging means and yarn: 1200 mm
Angle between imaging means and yarn: 90 °
Angle between illumination light and yarn: 70 °
With the above configuration, the yarns with twist were run in parallel with 50 yarns in the line sensor camera field of view at a speed of about 15 m / min to detect defects. In the detection of defects, thread slack was detected and distinguished with fluff and fluff.

  In order to detect defects, as shown in FIGS. 2 and 11, the LED illumination is arranged on the opposite side of the image pickup means across the yarn to illuminate the yarn, and the illumination light does not directly enter the image pickup means. In addition, the scattered light and diffracted light from the yarn and the defect are arranged at a position where they can enter the imaging means. Therefore, the image obtained by the imaging means becomes dark because light does not directly enter the background portion between the yarns. On the other hand, the yarn portion is scattered by the edge portion of the yarn, the gap between the fibers in the yarn, or transmitted through the yarn, and the light is incident on the imaging means. A brighter image was obtained compared to the background. Further, the obtained image was an 8-bit monochrome image, and the brightness value of each part in the image could be quantitatively measured in the range of 0 to 255 with commercially available image processing software. .

  Next, in the stage of binarizing the obtained image, the background is set to black by performing a process of setting a portion where the brightness value in the image is equal to or greater than the threshold value to white and a portion where the brightness value is less than the threshold value to black. And the processing was performed so that the defective portion was white. A specific binarization threshold was set in the range of 20 to 30 in an 8-bit monochrome image.

  Next, in the step of thinning the binarized image, the thinning process is performed on the image shown in FIG. 5 in the image obtained by binarization, and the process of removing white pixels in the image is repeated, A process of trimming pixels was performed until the width became one pixel, and an image as shown in FIG. 6 was obtained.

  Next, in the separation step of removing the branch points of the fine lines on the thinned image and separating the yarn corresponding fine lines and the defect candidate fine lines, the yarn candidate fine lines as indicated by reference numeral 9 in FIG. Described in digital image processing (issued by the Foundation for Image Information Education (CG-ARTS Association)), a branch point where a long line in the yarn running direction is an intersection with a defect candidate fine line, which is branched in the middle by a defective part Extraction was performed by performing raster scan on the thinned image using a 3 × 3 logic filter determined by the above method.

  After the branch point was extracted, the branch point portion was deleted from the thinned image to obtain an image in which the yarn-corresponding fine line and the defect candidate fine line were separated as shown in FIG. In the separated image, by deleting the line having the feature amount “long in the yarn running direction in the image” from the separated image as the yarn-corresponding thin line, it is possible to obtain an image in which only the defect candidate fine line is extracted. did it. Specifically, a line having a feature amount whose length in the yarn running direction exceeds a threshold and whose length in the yarn width direction is less than the threshold is separated as a yarn-corresponding thin line. An arbitrary fixed value in the range of 50 to 150 was used as the threshold value for the length in the yarn running direction, and an arbitrary fixed value in the range of 5 to 20 was used as the threshold value for the length in the yarn width direction.

  However, with this image as it is, the branch point portion that was deleted in the middle of the extracted defect candidate thin line will be missing, so by adding the extracted defect candidate thin line and the deleted branch point together, An image obtained by extracting only defect candidate thin lines as shown in FIG.

  Further, at the determination stage in the image processing, as a means for determining whether or not the defect is present, “the size of the smallest rectangle circumscribing the defect candidate thin line” which is the feature amount of the defect candidate thin line obtained in the separation stage is used. Specifically, the left and right ends of the defect candidate fine line are calculated in the direction parallel to the width direction of the yarn, and the upper and lower ends of the defect candidate fine line are calculated in the direction parallel to the running direction of the yarn, and the calculated four points Estimate the “minimum rectangle circumscribing the defect candidate fine line” parallel to the width direction of the yarn from the value, compare the value of the area of the rectangle calculated by the number of pixels with the preset threshold, and exceeded the threshold In this case, the defect candidate fine line was determined as a defect. As the threshold value, an arbitrary fixed value in the range of 200 to 400 was used.

  In addition, regarding the determination of the type of defect that distinguishes fluff and / or fluff and slack, “width in the direction perpendicular to the orientation direction of the yarn of the defect candidate thin line” was used. Specifically, the length of the axis in the width direction of the rectangular yarn in the “minimum rectangle circumscribing the defect candidate thin line” parallel to the yarn width direction calculated in the determination step is expressed as “the defect candidate thin line It was used as “width in the direction perpendicular to the orientation direction of the yarn” and the value of the length of the shaft was compared with a preset threshold value. As the threshold value, an arbitrary fixed value in the range of 10 to 25 was used.

  In parallel with the above inspection, the same inspection object was visually inspected by one inspector. As a result, it was confirmed that the fluff defect of each yarn checked by the inspector can be detected without leaking by the method of the present invention.

  In addition, using the feature quantity of “the smallest rectangular size circumscribing the defect candidate fine line”, the detected defects can be classified and recorded for each size, and fluff and / or fluff and thread slack Can be distinguished with an accuracy of about 90%.

The inspection results obtained by the present invention described above can be used to determine the quality of the yarn and the yarn package wound with the yarn, and have been confirmed to be effective for quality control. In addition, by using data such as increase / decrease in the number of defects, types of defects that occurred, images of defects that occurred, etc. using defect data, process improvements can be implemented in a timely manner and product yield can be improved. It was confirmed that it was connected to.
[Example 2]
Next, in the device configuration of Example 1, only the shutter speed of the camera was changed to 0.6 ms, and other device configurations and threshold values such as binarization were added under the same conditions as in Example 1. Defects were detected by running the yarns at a speed of approximately 10 m / min in parallel with 50 yarns in the line camera field of view. In addition, since the shutter speed of the camera is set to 2/3 in accordance with the speed of the yarn becoming 2/3, the size of one pixel in the captured image does not change between the first and second embodiments. .

As a result, even in the condition of Example 2 where the yarn speed is 10 m / min, an image similar to that of Example 1 can be taken, and by the method of the present invention, fluff defects for each yarn checked by the inspector can be detected without leaking. Was confirmed.
[Example 3]
Next, it implemented with the structure of FIG. 1 which has arrange | positioned illumination on the same surface side as a camera with respect to the yarn. The equipment configuration was the same as that of Example 1 except that the “angle formed by the illumination light and the yarn” of illumination was set to 20 °. In addition, threshold values such as binarization were the same as in the first embodiment. With this configuration, defects were detected by running 10 yarns in parallel in the line camera field of view at a speed of approximately 15 m / min.

  As the configuration shown in FIG. 1, the LED illumination is arranged at a position where the illumination light does not directly enter the imaging unit, and the reflected light or scattered light from the yarn or the defect can enter the imaging unit. As in Example 1, the background portion between the yarns is dark because light does not enter directly, and the yarn portion is reflected from the surface of the yarn and scattered light is incident on the imaging means. As a result, an image in which the yarn was captured brighter than the background was obtained.

As a result, even under the conditions of Example 3, it was confirmed that an image similar to that of Example 1 could be taken and the fluff defects could be detected without leaking.

  According to the method of the present invention, a plurality of yarns are simultaneously inspected online regardless of whether or not the yarn is twisted, the presence or absence of a defect and the type are determined, and information on defects for each yarn. Therefore, quality control of the yarn and the yarn package can be performed appropriately. The method of the present invention can be suitably used in a yarn manufacturing process and a yarn processing / processing step, and enables production of a yarn package in which the presence or absence of a defect or its state is grasped.

1: Yarn 2: Illuminating means 3: Imaging means 4: Arrow indicating the traveling direction of the yarn 5: Arrow indicating the illuminating direction of the illuminating means 6: Reflector 7: Fluff defect 8: Fluff defect 9: Branch point 10 : Thread slack 11: Data processing means

Claims (7)

A method for detecting defects in a yarn by illuminating the yarn or background to cause a brightness difference or a hue difference between the yarn and the background, then imaging the yarn together with the background, and processing the resulting image A yarn defect detection method characterized by performing image processing and determination in the following stages (a) to (d).
(A) a binarization stage for processing a captured image based on a brightness difference or a hue difference to obtain a binarized image; (b) a thinning stage for performing a thinning process on the binarized image;
(C) a separation step of removing branch points of the fine lines on the thinned image and separating the yarn-corresponding fine lines and the defect candidate fine lines;
(D) A determination step of comparing the feature quantity of the defect candidate fine line with a threshold value of a defect feature quantity set in advance in the storage means to determine whether or not the defect is present and / or the type of the defect. In the determination step (d), the width of the defect candidate thin line in the direction orthogonal to the direction of the yarn orientation is measured as a feature amount, and compared with a threshold value to determine whether the defect is fluff or slack in the yarn. The yarn defect detection method according to claim 1. 3. The yarn defect detection method according to claim 1, further comprising the following recording step (e) for the detected defect.
(E) A recording stage for recording an image of a defective portion determined to be a defect or measurement value data of the defective portion. Illumination means for illuminating the yarn or background to cause a difference in brightness or hue between the yarn and the background, imaging means for imaging the yarn with the background under illumination by the illumination means, and the obtained image A yarn defect detection device having image processing means for processing,
A yarn defect detection apparatus, wherein the image processing means performs image processing and determination at the following stages (A) to (D).
(A) A binarization step for processing a captured image based on a brightness difference or a hue difference to obtain a binarized image;
(B) A thinning step for thinning the binarized image;
(C) a separation step of removing branch points of the fine lines on the thinned image and separating the yarn-corresponding fine lines and the defect candidate fine lines;
(D) A determination step of comparing the feature quantity of the defect candidate fine line with a threshold value of a defect feature quantity set in advance in a storage unit and determining whether the defect is present and / or the type of the defect. 5. The lighting device according to claim 4, wherein the illuminating unit is disposed on a side facing the imaging unit across the yarn and at a position where direct light from the illuminating unit does not enter the imaging unit. Yarn defect detection device. The yarn is formed by winding the yarn subjected to the inspection by the yarn defect inspection method according to any one of claims 1 to 3 or by the defect inspection according to claim 4 or 5. Yarn package. The yarn package according to claim 6, wherein information on inspection results is attached.

Images