WO2010010767A1

WO2010010767A1 - Thread measuring apparatus, measuring program, and measuring method

Info

Publication number: WO2010010767A1
Application number: PCT/JP2009/061175
Authority: WO
Inventors: 浩孝藤崎; 紘規奥野
Original assignee: 株式会社島精機製作所
Priority date: 2008-07-23
Filing date: 2009-06-19
Publication date: 2010-01-28

Abstract

Disclosed is a thread measuring apparatus, which controls a back light source (102) so that a color difference between a thread (90) and an illumination may be at or higher than a predetermined value. With the illumination color of the back light source (102) being optimized, a digital image of the thread (90) is taken by a camera (6), and a two-dimensional Fourier-transformed image is formed along the width direction and the longitudinal direction of the thread. A low-frequency component corresponding to the thread body is extracted from the two-dimensional Fourier-transformed image, and is subjected to a two-dimensional reverse Fourier-transformation thereby reproducing the image of the thread body. Thus, it is possible to determine the thread diameter precisely and the color components of the thread.

Description

Yarn measuring device, measuring program and measuring method

This invention relates to yarn measurement, and more particularly to measurement of yarn thickness and fluff amount.

One of the inventors has proposed that the yarn image is converted into a Fourier transform or an autocorrelation function, and the background, the yarn body, and the fluff are separated from each other to obtain the yarn diameter and the fluff amount. In the process of improving this technique, the inventor noticed the following problem. For example, when capturing an image of a thread including a bright part and a dark part, it is difficult to obtain a dark part of the thread from Fourier transform data etc. with a dark background color, and conversely, it is difficult to obtain a bright part of the thread with a bright background color. . In other words, it is difficult to obtain data of the portion close to the background color in the yarn from Fourier transform data or the like.

Japanese Patent Application 2008-68841

An object of the present invention is to be able to accurately determine the thread diameter using a normal imaging device such as a digital camera or a scanner, and to accurately measure the thread properties regardless of the thread color. There is to be able to do it.

The yarn measuring device according to the present invention is a device for taking a digital image of a yarn by an imaging means and obtaining a yarn diameter or fluff amount of the yarn.
A backlight light source that projects a background light having a variable light source color onto a thread during imaging;
Color extraction means for obtaining the color component of the thread from the digital image imaged by the imaging means;
Arithmetic means for obtaining the light source color so that a color difference of a predetermined value or more is obtained between the extracted color component and the light source color of the backlight light source;
Control means for controlling the backlight source according to the obtained light source color;
Conversion means for obtaining Fourier transform data or autocorrelation function data by digital signal processing for the digital image of the yarn imaged by the imaging means using the obtained light source color as a background color;
Yarn property calculating means for obtaining the yarn diameter or fluff amount of the yarn from the data obtained by the converting means.

The yarn measurement program according to the present invention includes an image pickup device for picking up a digital image of a yarn, an image pickup device including a backlight source that projects background light having a variable light source color onto the yarn, and an image pickup device. In the program for the yarn measuring device, comprising a signal processing of the digital image captured by the computer and a computer for controlling the backlight light source,
By the computer
Obtaining a color component of the thread from the digital image captured by the imaging means;
Obtaining the light source color so that a color difference of a predetermined value or more is obtained between the extracted color component and the light source color of the backlight light source;
Controlling the backlight source according to the determined light source color;
From the obtained Fourier transform data or autocorrelation function data, the step of obtaining Fourier transform data or autocorrelation function data by digital signal processing for the digital image of the yarn imaged by the imaging means with the obtained light source color as the background color, And determining the yarn diameter or fluff amount of the yarn.

In the method for measuring a yarn according to the present invention, a digital image of the yarn is picked up by an image pickup means, and the yarn diameter or the amount of fluff is calculated.
Imaging a digital image of the yarn by the imaging means while projecting background light having a variable light source color from the backlight light source to the yarn;
A step of obtaining a color component of the yarn by the extracting unit from the digital image captured by the imaging unit, and a light source color capable of obtaining a color difference of a predetermined value or more between the extracted color component and the light source color of the backlight source. A step to obtain by
Controlling the backlight light source by control means according to the determined light source color;
For the digital image of the yarn imaged by the imaging means with the determined light source color as a background color, Fourier transform data or autocorrelation function data is obtained by digital signal processing using the conversion means;
And a step of obtaining the yarn diameter or fluff amount of the yarn from the data obtained by the converting means by the yarn property calculating means.

In this specification, since the light source color in the color space such as RGB, Lab, YCbCr is handled, changing the lightness is also included in changing the light source color. The control means controls the light source color of the backlight light source so as to match the obtained light source color, but the light source color of the backlight light source may be brought close to the obtained light source color even if it does not exactly match.

Preferably, the backlight light source is a color light source capable of controlling brightness and hue.
In the control of the light source color, the brightness of the light source color is changed to search for a color difference that is greater than or equal to a predetermined value with the thread color component. When the backlight light source is controlled with the brightness and there is no brightness at which a color difference of a predetermined value or more is obtained, the hue of the light source color capable of obtaining a color difference of the predetermined value or more is searched, and the searched hue is obtained. Control the backlight source.

In this specification, the description related to the yarn measuring device also applies to the yarn measuring program and the yarn measuring method, and the description related to the yarn measuring program and the yarn measuring method also applies to the yarn measuring device.

In this invention, the digital image of the yarn is converted into Fourier transform data or autocorrelation function data. The Fourier transform data of the yarn includes data corresponding to the yarn diameter and the fluff amount, and the yarn diameter can be obtained from this. The correlation width in the autocorrelation function data corresponds to the yarn diameter, and the intensity of the signal having a very short correlation distance corresponds to the fluff amount. Therefore, in the present invention, if a digital image of a yarn is acquired with a digital camera or the like, the yarn diameter and the amount of fluff can be obtained only by digital signal processing, and no special optical system is required. Furthermore, if a digital image can be captured even with a translucent thread, the thread diameter and the amount of fluff can be determined. Also, delicate data such as the boundary between the yarn body and the fluff is not required. *

Further, in the present invention, since the background color is controlled so that a sufficient color difference from the color component of the yarn is obtained, accurate measurement can be performed regardless of the color of the yarn.
Further, the yarn diameter can be accurately obtained even with a translucent yarn, and the amount of fluff can be obtained accurately even with a translucent fluff.

最初 First, by searching for the optimal background color by changing the brightness of the light source color, the color component of the thread can be accurately obtained. Accordingly, an accurate image of the thread cut out from the background can be obtained. If the optimum background color cannot be obtained and the background color is searched for by controlling the hue of the light source color, the thread diameter and the fluff amount can be obtained for almost any thread. *

Block diagram of the reference yarn measuring device The figure which shows the process by the two-dimensional Fourier transform with respect to the thread | yarn image in a reference example The figure which shows the process by the one-dimensional Fourier transform with respect to the thread | yarn image in a reference example Diagram showing filter generation in the reference example The figure which shows the thread image used in the test example The figure which shows the two-dimensional Fourier-transform data of FIG. The figure which shows the thread | yarn main body image obtained by the two-dimensional inverse Fourier transform from the data of FIG. The figure which shows the fluff image obtained by two-dimensional inverse Fourier transform and binarization from the data of FIG. Block diagram of the reference thread measurement program Block diagram showing an example of a reference example applied to yarn diameter management Block diagram of the yarn measuring device of the embodiment The block diagram which shows the mechanical structure of the measuring device of the thread | yarn of an Example. Block diagram of backlight light source control unit in the embodiment The flowchart which shows the determination algorithm of the backlight in an Example The figure which shows the example of determination of the backlight in an Example typically Photo of yarn with backlight off The figure which shows the thread | yarn image created by the reference example based on the image of FIG. Photo of yarn with backlight on The figure which shows the thread | yarn image created by the Example based on the image of FIG. The figure which shows the thread | yarn main body image created by the Example based on the image of FIG. The figure which shows the fluff image created by the Example based on the image of FIG.

The following is an optimum embodiment for carrying out the present invention. First, with reference to FIGS. 1 to 10, measurement of yarns in an apparatus without a backlight light source will be described, and then the backlight light source and its control will be described with reference to FIGS.

FIGS. 1 to 10 show a yarn measuring device 2 and a yarn measuring program 80 of a reference example. In FIG. 1, 4 is a bus, 6 is a color digital camera, and may be a color scanner, 8 is a color monitor, 10 is a keyboard, 12 is a color printer, and 14 is a mouse. A trackball, joystick, stylus or the like may be used instead of the mouse 14. Reference numeral 16 denotes a network interface, which inputs and outputs various programs and data, and 18 to 22 are image memories. Among these, the image memory 18 stores the color image of the yarn imaged by the digital camera 6, the image memory 20 stores the image of the yarn body separated from the yarn image by Fourier transform, and the image memory 22 stores the yarn image. The image of the fluff separated from the image by Fourier transform is stored.

The Fourier transform unit 24 transforms the yarn color image into a two-dimensional Fourier transform image. As pre-processing before conversion, the Fourier transform unit 24 may convert a color image into, for example, a lightness image, and may perform one-dimensional Fourier transform instead of two-dimensional Fourier transform. Further, when the color of the yarn is particularly important like melange yarn, for example, a two-dimensional Fourier transform may be performed for each of RGB components. The Fourier transform image of the yarn is stored in the Fourier transform image storage unit 25. The Fourier transform unit 24 also performs a two-dimensional inverse Fourier transform and performs a two-dimensional inverse Fourier transform on the image processed by the filter to create an image of the yarn body and an image of the fluff. The filter generation unit 26 generates a filter for separating the yarn body and the fluff from the Fourier transform image. The binarization unit 27 binarizes the inverse Fourier transformed image, and in particular binarizes the fluff image. The counter 28 counts the number of pixels for the fluff image or the yarn body image, obtains the yarn diameter, the yarn diameter distribution or variation pattern for the yarn body, the fluff amount for the fluff, etc. Ask for.

The program memory 30 stores a yarn measurement program 80 and the like, and the yarn model creation unit 32 uses the yarn diameter obtained by the counter 28 and the yarn texture stored in the image memory 18 to produce a three-dimensional yarn. Generate a model. The simulation unit 34 uses the generated yarn model to perform simulation so as to express individual loops of the knit garment with respect to the knitting data stored in the knitting data storage unit 36.

FIG. 2 shows an outline of processing in the reference example. The short side direction of the yarn is the x direction and the long side direction is the y direction. The image obtained by the digital camera 6 includes a yarn main body image 40 and a fluff image 42, and includes an image in which the yarn main body and the fluff are mixed between the

images

40 and 42. In the two-dimensional Fourier transform image obtained by performing the two-dimensional Fourier transform on this image, the yarn body area 44 is included in the low frequency region, the fluff area 46 is included in the high frequency region, and the intermediate area 48 is included therebetween. In the following description, ω represents a frequency along the x direction, and φ represents a frequency along the y direction. The frequency range for performing the Fourier transform is determined according to the high frequency end of the fluff area 46. The Fourier transform data has both a real component and an imaginary component, but a power spectrum may be used instead. For example, if the real component is Re (ω) and the imaginary component is I (ω), the power spectrum P (ω) is given by (Re ² (ω) + I ² (ω)) ^1/2 .

When a frequency component corresponding to the yarn main body is extracted from the two-dimensional Fourier transform image by the filter 50 that passes only the yarn main body area 44, a two-dimensional inverse Fourier transform is performed to obtain a yarn main body image 41. Similarly, when a filter 51 that passes only the fluff area 46 is applied and two-dimensional inverse Fourier transform is performed,

fluff images

43 and 43 are obtained. For example, the filter 50 passes only low frequency components along the ω direction, and the filter 51 passes only high frequency components along the ω direction. The narrower the width of the frequency passed through the filter 50, the smoother the thread main body image 41 obtained. Conversely, when the width of the component passed through the filter 50 is increased, subtle irregularities and twists of the thread main body are removed. 41 can be reflected. In the reference example, a gap is provided between both ends of the frequency passed by the filter 50 and both ends of the frequency cut by the filter 51 to remove the frequency component of the intermediate area 48. However, the intermediate area 48 may not be removed, and all the frequencies outside the frequency cut by the filter 50 may be passed by the filter 51.

Next, if the width is counted along the x direction with respect to the yarn body image 41, the yarn diameter can be obtained. Alternatively, the yarn diameter can be obtained by counting the total number of pixels corresponding to the yarn body and dividing by the number of pixels in the length direction of the yarn body. Further, if the yarn diameter is obtained at various positions and the average value, dispersion, abnormal value, and the like are obtained, the degree of fluctuation of the yarn diameter due to twisting or knurling can be obtained. Since the fluff image 43 is generally a weak image, the amount of fluff can be obtained by separating the fluff from the background by binarization and counting the number of pixels corresponding to the fluff. In order to obtain the fluff amount, the total number of pixels corresponding to the fluff may be counted, or the number of pixels corresponding to the fluff may be counted on a plurality of lines along the x direction. Also, the amount of fluff can be obtained by converting the Fourier transform component on the high frequency side corresponding to the fluff into a fluff amount for each frequency according to a reference table (not shown) and adding it to the frequency.

In the reference example, two-dimensional Fourier transform is used, but one-dimensional Fourier transform may be used. In the case of one-dimensional Fourier transform, it is easy to understand because there is no processing in the y direction in FIG. 3 is a yarn main body image 52, and there is a fluff image 54 around it. When this is subjected to a one-dimensional Fourier transform along the width direction of the yarn, a yarn body area 56 and a fluff area 58 are obtained. Next, the filter 59 cuts out a frequency component corresponding to the yarn body, and the filter 60 cuts out a high frequency component corresponding to the fluff. When the inverse Fourier transform is performed, a yarn main body image 53 and a fluff image 55 are obtained. Although a rectangular filter is used in FIG. 3, a filter whose transmittance changes smoothly may be used.

FIG. 4 shows filter generation. The image of the yarn body has a peak near the frequency 0, and the width of the peak corresponds to the yarn diameter. The bands of the

filters

59 and 60 are determined from the peak shape of the yarn body area 56 in the Fourier transform. For example, the half width α of the peak of the yarn main body area 56 is obtained, and this is multiplied by a first coefficient to obtain a bandwidth β for cutting out the yarn main body area 56. Similarly, the second coefficient is multiplied to obtain a second bandwidth γ for cutting out the fluff area 58. Bandwidth β is used as a low-pass filter, and bandwidth γ is used as a high-pass filter. In addition to this, by using a zero crossing point between the yarn main body area 56 and the fluff area 58, or by extrapolating a peak line in the yarn main body area 56 and using a point crossing zero, the full width at half maximum α An alternative value can be obtained. When the digital image of the yarn body is represented by a rectangle of width D, the Fourier transform is almost sinc (f · D), where sinc (x) is
It is given by sin (x) / x. If the width of the peak of the yarn main body is known in FIG. 4, D, that is, the yarn diameter can be obtained, and the yarn diameter can be obtained without inverse Fourier transform.

5 to 8 show the processing results of the thread image. FIG. 5 is a yarn image captured by a digital camera, and FIG. 6 is an image obtained by two-dimensional Fourier transform. In FIG. 6, the directions of the axes are indicated by ω and φ. FIG. 7 shows an image obtained by processing the image of FIG. 6 with a filter and cutting out only the low-frequency side yarn body area and performing two-dimensional inverse Fourier transform on this, which is an image obtained by smoothing the surface of the yarn body. This image is sufficiently clear and the yarn diameter can be calculated. Note that the yarn diameter may be calculated after binarizing the image of FIG. The image of FIG. 8 is a fluff image obtained by applying a filter to the image of FIG. 6 to extract the frequency component of the fluff area, binarizing it after two-dimensional inverse Fourier transform. Note that the fluff image is darker than the main body image of the yarn, and the binarization threshold is set lower for the fluff than the main body of the yarn. If the number of white pixels in FIG. 8 is counted, the amount of fluff can be obtained.

Once the thread body diameter is determined, a thread model can be created. The thread model can be represented by a pillar such as a quadrangular prism or a hexagonal prism, and can be formed into a loop by dividing its surface into polygons and bending the thread model along the sides of the polygons. If the diameter of the yarn is determined, the diameter of the quadrangular prism or hexagonal prism is determined, so that a polygon on the surface of the thread is obtained. Since the texture of the thread has already been imaged with a digital camera, this is texture mapped.

FIG. 9 shows a yarn measurement program 80. A two-dimensional Fourier transform command 81 performs a two-dimensional Fourier transform on the digital image of the yarn. The filtering instruction 82 generates a filter from the Fourier-transformed image, and generates a filter for cutting out the yarn body and a filter for cutting out the fluff. The two-dimensional inverse Fourier transform instruction 83 performs inverse Fourier transform on the filtered Fourier transform data and separates it into a yarn body image and a fluff image. The binarization command 84 may binarize these images, and may omit binarization of the yarn body image as described above. The count instruction 85 counts the yarn diameter and the amount of fluff, for example, by counting the number of pixels corresponding to the yarn in the yarn body image and the fluff image. Further, since the Fourier transform and the inverse Fourier transform have the same processing contents, these processes can be largely shared.

Fig. 10 shows an example of application to yarn quality control. Reference numeral 90 denotes a thread, and

reference numerals

91 and 92 denote thread feeding rollers. The digital camera 6 captures an image of the thread. In order to obtain an accurate image, it is preferable to stop the

rollers

91 and 92 to capture a still image during imaging. For example, the measuring device 2 obtains the yarn diameter of the captured image and stores the yarn diameter in the storage device 94. Then, the quality of the yarn 90 is controlled based on the obtained dispersion of the yarn diameter, the presence or absence of an abnormal value, and the like. The amount of fluff may be measured by the measuring device 2 and stored in the storage device 94 to similarly control the amount of fluff.

In the reference example, the following effects can be obtained.
(1) Only digital camera 6 is required as hardware before signal processing. Since it is sufficient that the image of the thread can be captured by the digital camera 6, no accurate adjustment or the like is required.
(2) A filter for cutting out the yarn body and a filter for cutting out the fluff can be automatically generated from the peak shape on the low frequency side corresponding to the yarn body. For this reason, both thin and thick threads can be easily processed.
(3) Since diffracted light is not used, the thread diameter and the amount of fluff can be measured even with a translucent thread.

In the reference example, Fourier transform is shown, but as is well known, Fourier transform and autocorrelation function have similar properties. That is, the same processing can be performed by using the correlation distance in the autocorrelation function instead of the frequency in the reference example. For example, the yarn diameter D is the width of the correlation in the autocorrelation function. For example, the longest correlation peak value in the one-dimensional autocorrelation function is the yarn diameter. In addition, the intensity of a component having a correlation distance of 0 and an extremely short component in the autocorrelation function represents the amount of fluff. *

FIG. 11 to FIG. 21 show examples and results. FIG. 11 shows a configuration of the embodiment. Reference numeral 100 denotes a backlight light source control unit which controls the backlight light source 102, and the backlight light source 102 gives background light to the yarn to be measured. The backlight light source control unit 100 obtains an approximate color component of the thread 90 from the image from the digital camera 6 so that the color of the thread itself and the background color (light source color) have a sufficient distance in the color space. The light source color of the backlight light source 102 is controlled. When the thread 90 has a plurality of color components, for example, a color component of the thread closest to the background color is set to have a sufficient distance from the background color. In this specification, the distance regarding the color is a distance in the color space, and the distance has the same meaning as the color difference. In addition, when there are a plurality of color components, they are represented by the smallest of the distances between the background color and each color component. With respect to the other parts, the embodiment is the same as the reference example of FIGS.

FIG. 12 shows the backlight source 102 and its surroundings. 91 and 92 are the rollers, and 6 is the digital camera. The digital camera 6 and the backlight light source 102 are disposed so as to face each other with the thread 90 interposed therebetween. Preferably, in addition to the backlight light source 102, a standard light source (not shown) is disposed on the front side and the back side of the digital camera 6, for example. The standard light source is arranged so that only the reflected light enters the digital camera 6 from the standard light source.

The backlight source 102 is realized by, for example, an array of three or more LEDs, a liquid crystal panel with a backlight, a white fluorescent lamp, or the like. For example, the color coordinates of the background light are controlled by changing the brightness of the LED for each color. When a white light source is used as the backlight light source 102, for example, an array of white LEDs or a white fluorescent lamp is used as the light source. Further, the backlight source 102 is provided with a translucent plate (not shown) on the output side so that a direct surface light from an LED, a fluorescent lamp, or the like does not leak. In addition, a dark filter may be detachably attached to the output side of the backlight light source 102 in case a darker light source color is required.

FIG. 13 shows the configuration of the backlight light source control unit 100. The color component extraction unit 104 extracts an approximate color component of the yarn from the digital camera 6. The color component is a color coordinate of the color space, and any color space such as RGB or HLS can be used. Here, the RGB space is used. As shown in FIG. 12, since the thread 90 is passed between the

rollers

91 and 92, when pixels on the line connecting the

rollers

91 and 92 are taken out from the digital camera 6, a plurality of pixels of the thread 90 are obtained. . For this pixel, color components, that is, color coordinates are obtained, and when they have a wide distribution, for example, in order to facilitate processing, they are collected into a relatively small number of components. The color component distribution may be used as it is.

The background color calculation unit 106 controls the brightness while keeping the backlight light source 102 in black and white, for example. Here, the brightness of the backlight light source can be referred to as the brightness of the background color. Then, while changing the background color, a search is made for a lightness difference between each color component extracted by the color component extraction unit 104 and a predetermined value or more. In other words, a search is made for a color space in which a distance greater than or equal to a predetermined value is obtained for all color components of the yarn. If there is one that can obtain such a distance, the backlight light source 102 is driven with the lightness. The search may be continued until the maximum distance is obtained or may be terminated when a distance greater than a predetermined value is obtained. Here, changing the background color is not actually controlling the backlight source 102 but changing the background color in calculating the distance in the color space. If the thread is translucent, in order to obtain a more optimal background color, the backlight light source 102 is controlled to actually change the background color, and the thread color component and the background color are measured each time. You may measure the distance.

If the predetermined distance cannot be obtained even when the backlight light source 102 is controlled in the black and white range, the background color calculation unit 108 can change the light of the backlight light source 102 in the RGB space to obtain a distance of a predetermined value or more. Explore things. Here, the lightness that maximizes the distance from each color component of the thread has already been obtained by the background color calculation unit 106. While the lightness is fixed at this value, the hue and saturation of the backlight light source 102 are changed. In other words, for the line segment connecting white and black in the RGB space, the light source color is changed in a plane perpendicular to the line segment from the optimum point obtained by the background color calculation unit 106, and a distance of a predetermined value or more is obtained. Obtain the obtained light source color. The method for obtaining the optimal light source color is arbitrary, and the light source color that is farthest from the color component of the yarn may be obtained, or the search is stopped when a predetermined distance or more is obtained for each component of the yarn. Also good.

FIG. 14 shows the configuration of a program for determining the background color. This program is stored in a storage medium such as a CD-ROM or supplied by a carrier wave from the Internet or the like. The program for determining the background color is installed in an appropriate computer, thereby realizing the backlight light source control unit 100. In step 1, color components are extracted from the yarn image captured by the digital camera 6 or the like. Note that the process of collecting color components into a small number may be omitted. In step 2, the background color is changed in the black and white range, the distance in the RGB space from each color component of the yarn is obtained, and a search is made for those that can obtain a distance of a predetermined value or more for all the color components. If a background color that can obtain a distance greater than or equal to the predetermined value is not obtained in step 2, the lightness of the background color that maximizes the distance is used in step 3, and the background color is changed in the color space, so that all the color components of the thread A search is made for a distance that is greater than or equal to a predetermined value.

When a white light source is used as the backlight light source 102, the background color calculation unit 108 in FIG. 13 and step 3 in FIG. 14 are not provided. Further, instead of the procedure of obtaining the optimum background color in black and white first and then obtaining the optimum background color in color, the optimum background color may be obtained in color from the beginning.

Fig. 15 shows two examples of background color determination. For example, when only the white color component 110 and the black color component 111 are present as in the thread on the left side of FIG. 15, for example, there is an optimal background color in the gray area. On the other hand, when the gray color component 112 is included in addition to the white and

black color components

110 and 111 as shown on the right side of FIG. It is difficult to separate them sufficiently. In such a case, the color component of the thread and the background color are sufficiently separated using the background color.

Figures 16 to 21 show the results of the examples. FIG. 16 is an image of a thread imaged with a dark background color with the backlight light source off. FIG. 17 is a yarn image obtained by subjecting the yarn image of FIG. 16 to two-dimensional Fourier transform and two-dimensional inverse Fourier transform. In the following, the two-dimensional Fourier transform and the two-dimensional inverse Fourier transform are performed for each component of the color space such as RGB. The black part of the thread disappears in FIG. This means that the color component of the thread close to the background color disappears. FIG. 18 is a yarn image captured using a black and white background color with optimized brightness. FIG. 19 is a yarn image reproduced by performing two-dimensional inverse Fourier transform and two-dimensional inverse Fourier transform on this. Since the color component of the thread and the background color are sufficiently separated at the stage of FIG. 18, an image that sufficiently reproduces the thread is obtained in FIG. 20 is an image of the yarn main body in the process of creating the image of FIG. 19, and FIG. 21 is an image of the fluff. The yarn main body and the fluff are sufficiently reproduced for the black portion of the yarn. Instead of the two-dimensional Fourier transform, a one-dimensional Fourier transform along the radial direction of the yarn may be used. For example, if there is one piece of one-dimensional Fourier transform data, the yarn diameter data and the fluff amount data can be obtained one by one, but the fluff amount varies greatly depending on the imaging position. It is difficult to find accurately. Next, if there are a plurality of one-dimensional Fourier transform data, both the yarn diameter and the fluff amount can be accurately obtained.

If a black and white background color is used, it is easy to accurately extract the color components of the yarn body and fluff. In particular, the translucent fluff color component can be accurately extracted. On the other hand, with the background color, it is difficult to separate the fluff color from the background color. In the embodiment, even when the yarn has a plurality of color components, each color component can be obtained, and an accurate image of the yarn or fluff can be obtained. Therefore, the yarn image can be used for simulation of garment using the measured yarn. Can be used. In addition, an image in which the yarn body and the fluff are separated can be obtained, and the average value and distribution of the fluff amount, the yarn diameter, the abnormal value, and the like can be obtained.

2 Measuring Device 4 Bus 6 Digital Camera 8 Color Monitor 10 Keyboard 12 Color Printer 14 Mouse 16 Network Interface 18-22 Image Memory 24 Fourier Transform Unit 25 Fourier Transform Image Storage Unit 26 Filter Generation Unit 27 Binary Unit 28 Counter 30 Program Memory 32 Yarn model creation unit 34 Simulation unit 36 Knitting

data storage unit

40, 41

Yarn body image

42, 43 Fluff image 44 Yarn body area 46 Fluff area 48

Intermediate area

50, 51

Filter

52, 53

Yarn body image

54, 55 Fluff image 56 Yarn body area 58

Fluff area

59, 60 Filter 80 Yarn measurement program 81 Two-dimensional Fourier transform instruction 82 Filtering instruction 83 Two-dimensional inverse Fourier transform instruction 84 Binary instruction 85 Count Decree 90

yarn

91, 92 roller 94 storage device 100 backlight source control unit 102 backlight source 104 color component extracting unit 106 background

color calculation unit

110, 111 color component

Claims

In an apparatus for taking a digital image of a yarn with an imaging means and obtaining the yarn diameter or fluff amount of the yarn,
A backlight light source that projects a background light having a variable light source color onto a thread during imaging;
Color extraction means for obtaining the color component of the thread from the digital image imaged by the imaging means;
Arithmetic means for obtaining the light source color so that a color difference of a predetermined value or more is obtained between the extracted color component and the light source color of the backlight light source;
Control means for controlling the backlight source according to the obtained light source color;
Conversion means for obtaining Fourier transform data or autocorrelation function data by digital signal processing for the digital image of the yarn imaged by the imaging means using the obtained light source color as a background color;
And a yarn property calculating means for obtaining the yarn diameter or fluff amount of the yarn from the data obtained by the converting means.
The backlight light source is a color light source capable of controlling brightness and hue,
The calculation means searches for a color difference of a predetermined value or more with the color component of the yarn by changing the brightness of the light source color, and if there is a brightness that can obtain a color difference of a predetermined value or more, When the backlight unit is controlled with the brightness by the control means and there is no brightness that can obtain a color difference of a predetermined value or more, the hue of the light source color that can obtain a color difference of a predetermined value or more is searched, and the searched hue is obtained. The yarn measuring apparatus according to claim 1, wherein a backlight light source is controlled by the control means.
An image pickup device including an image pickup means for picking up a digital image of the yarn, a backlight light source for projecting background light having a variable light source color to the yarn at the time of image pickup, and a signal processing of the digital image picked up by the image pickup means In addition, in a program for a yarn measuring device comprising a computer for controlling the backlight light source,
By the computer
Obtaining a color component of the thread from the digital image captured by the imaging means;
Obtaining the light source color so that a color difference of a predetermined value or more is obtained between the extracted color component and the light source color of the backlight light source;
Controlling the backlight source according to the determined light source color;
From the obtained Fourier transform data or autocorrelation function data, the step of obtaining Fourier transform data or autocorrelation function data by digital signal processing for the digital image of the yarn imaged by the imaging means with the obtained light source color as the background color, And a step of obtaining a yarn diameter or a fluff amount of the yarn.
In a method of taking a digital image of a yarn with an imaging means and obtaining the yarn diameter or fluff amount of the yarn,
Imaging a digital image of the yarn by the imaging means while projecting background light having a variable light source color from the backlight light source to the yarn;
A step of obtaining a color component of the yarn by the extracting unit from the digital image captured by the imaging unit, and a light source color capable of obtaining a color difference of a predetermined value or more between the extracted color component and the light source color of the backlight source. A step to obtain by
Controlling the backlight light source by control means according to the determined light source color;
For the digital image of the yarn imaged by the imaging means with the determined light source color as a background color, Fourier transform data or autocorrelation function data is obtained by digital signal processing using the conversion means;
And a step of obtaining the yarn diameter or fluff amount of the yarn from the data obtained by the conversion means by means of the yarn property calculating means.