EP1989139A1

EP1989139A1 - Method and device for rating the quality of a bobbin

Info

Publication number: EP1989139A1
Application number: EP06723104A
Authority: EP
Inventors: Marc Schaad
Original assignee: SSM Schaerer Schweiter Mettler AG
Current assignee: SSM Schaerer Schweiter Mettler AG
Priority date: 2006-02-24
Filing date: 2006-02-24
Publication date: 2008-11-12
Also published as: WO2007098776A1; CN101384500A

Abstract

Disclosed is a method for automatically rating the quality of a bobbin (3), especially for detecting mirrors of the bobbin (3), by means of an arithmetic module (25). Said method comprises the following steps: at least one image of the bobbin (3) is optically detected with the aid of a camera (21); pixels of a digital version of said image are converted into data values using a transformation algorithm; and a quality index is calculated from the data values. Also disclosed is a device for carrying out the inventive method.

Description

Method and device for classifying the quality of a yarn package

The invention relates to a method and a device for automatic classification of the quality of a yarn package, in particular for detecting mirrors of the yarn package. The inventive method is suitable for use in Umspulverfahren of yarns, or in an apparatus for performing the Umspulverfahrens (rewinding), with a Umspulgeschwindigkeit, with the steps of unwinding a yarn from a supply spool and winding the yarn on a rotating with a take-up yarn, wherein the yarn is guided over a yarn path from the supply spool to the yarn carrier. Rewinding of yarns is often done in yarn making. The yarns are to be wound on spools in such a way that the spools in subsequent processes can be used optimally as feed bobbins. In particular, the type of winding and its density should be optimized. Thus, in some yarn processing processes, it is not tolerable if the yarn is unrolled randomly on the supply spool, that is to say has a so-called "wild winding." Further, during processing of yarns, eg during dyeing, the yarn to be dyed is rewound or it is a coil is wound with a yarn having a predetermined yarn length, ie a predetermined length of run.

The rewinding should be carried out with the highest possible winding speed (rewinding speed). However, as the winding speed increases, so does the yarn tension of the yarn to be rewound during the rewinding process, and yarn tension variations are more pronounced. The latter leads to a deterioration in the quality of the winding on the coil created, since thread tension fluctuations during winding lead to Garndichteschwankungen of the windings. The latter can in turn worsen the unwinding of these coils in subsequent processes. Further, the thread tension variations can easily lead to thread breaks during rewinding. The latter occurs especially in areas of integer turns ratio, i. the ratio of coil revolution to Doppelhubzahl, so-called mirror zones or mirrors, the supply spools heaped up. In these areas, the thread layers are close to each other without being fixed by crossing points, which can lead to entrainment of adjacent threads and thus thread tension peaks or thread breaks in the overhead deduction. These thread tension peaks severely limit the rewinding speed.

Usually, the yarn is withdrawn from the supply spool by the rotating movement of a take-off roll (deduction feed). The take-off roll includes usually a motor-driven main reel and a passive separation reel that sets the geometry of wraps of yarn around the main reel. In order to avoid yarn tension variations in the unwinding region, it is known to exert a pressing force on the main roll by means of a "nip-roller", that is to say a further roller. This traps the yarn between the main roll and the other roll, causing the yarn to be pulled off at the tangential speed of the main roll. Further, various biasing systems disposed between the pre-bobbin and a delivery mechanism are known to increase and stabilize the yarn tension of the yarn as it is withdrawn from the supply bobbin.

The yarn is often pulled off during rewinding overhead. When pulling off overhead, the feed bobbin does not rotate itself, so working with an immediately usable spare bobbin is possible when pulling off overhead.

In addition to the yarn tension peaks mentioned due to mirrors of the supply bobbin, the yarn tension of the yarn in the yarn path during rewinding varies for various other reasons:

- The fact that the diameter of the supply spool decreases with time by unwinding the yarn, the geometry of the "balloon", which differs from the

Yarn movements along the coil axis is formed, accordingly, whereby the thread tension is affected.

- By increasing the speed of the balloon with decreasing diameter of the withdrawn coil, the forces acting on the yarn in the "balloon" centrifugal forces increase, which leads to a Fadenspanπerhöerhöhung.

It has been found that the thread tension fluctuations in known rewinding can be reduced only insufficiently while maintaining a high rewinding speed, in particular stress peaks occur in mirror zones. From DE 40 30 892 Al it is known to measure the thread tension during rewinding in the supply reel by means of a yarn tension sensor and to regulate the rewinding speed as a function of the yarn tension in order to avoid yarn breaks in eg supply coil quality-related voltage spikes. The thread tension is measured by frictional heat generated by a thread on the yarn tension sensor. As a result, the quality of the feed bobbin is automatically determined indirectly. A disadvantage of such a thread tension measurement is that an additional load of the thread is generated by the yarn tension sensor, which may favor thread breaks.

The invention has for its object to provide a method and apparatus for automatically classifying the quality of a yarn package, which avoid the disadvantages of the prior art and their use in the rewinding of yarns, especially at a high rewinding speed, a load of the thread through the quality determination avoids.

This object is achieved by the method and apparatus of the independent claims. The dependent claims represent preferred embodiments of the invention.

According to the invention, the object with regard to the method for the automatic classification of the quality of a yarn package, in particular for detecting mirrors of the yarn package, is achieved by means of a computer module having the following method steps:

optically capturing at least one image of the yarn package by means of a camera,

- Converting pixels of a digital version of the image by means of a transformation algorithm into data values and - calculating a quality score from the data values. By means of the optical quality rating of the yarn package made by the method according to the invention, a non-contact determination of the yarn package quality is made possible. The method can be used as a basis for controlling a rewinding speed when using a computing module operating at a sufficient speed, whereby eg

Thread tension sensors are dispensable during rewinding. The method can be implemented inexpensively and simply using commercially available modules as a camera and a computer as a computing module.

In a first preferred embodiment of the invention

Method is used as a transformation algorithm, a Fourriertransformation, wherein the data values are calculated as the amplitudes of the thus generated Fourriertransformation in the frequency domain. For this purpose, it is advantageous to make a uniform illumination of the yarn package by means of at least one light source. The pixels of the digital version of the image of the yarn package are interpreted as the basis of the known in numerical discrete two-dimensional Fourriertransformation in space. As a Fourier transformation algorithm, for example, the known fast Fourier transformation (Fast Fourier Transformation FFT), which is available as a software module in commercially available program libraries, is suitable. By applying the Fourier transformation, it is possible reliably to detect coarse structures in the image of the yarn package, as occur in poor windings, as low-frequency components of the Fourier transformation.

In this case, the data values can be advantageous before the calculation of the

Quality score from the data values by values, e.g. weighted by multiplication, with weighting factors, e.g. highlight significant frequency bands for certain quality deficiencies. It can e.g. in the simplest case, only a frequency range corresponding to the low frequencies can be selected. The correspond

Weighting factors for selected frequencies 1 and for non-selected 0. If the weighting factors are chosen to be inversely proportional to the frequency values of the associated amplitudes, then the weighting thus made means that lower frequencies are more heavily weighted than high frequencies. So coarse structures in the image of the yarn package are more in the quality index than fine structures.

As an alternative to the Fourier transformation, it may be advantageous to use a conversion into gray values as the transformation algorithm, the data values being calculated as the gray values. This method variant can be carried out with a very low computation effort and is therefore suitable for use at very high rewinding speeds and / or rapidly changing reel qualities over the course of an unwinding process, i. irregular windings.

To carry out the alternative procedure, the yarn package is preferably illuminated by means of a light source, wherein the light source is positioned such that the beam direction of the light source forms an acute angle with the longitudinal axis of the yarn package. As a result, coarse structures of the coil winding cast particularly pronounced shadows, which are easily recognizable on the image of the coil and thus significantly emerge in the gray value representation of the image.

If normalization of the data values is carried out, with the quality index being calculated from the normalized data values, external parameters, such as the illumination intensity when the camera images the image of the yarn package, which can influence the calculated data values, can be excluded. The absolute values of the quality characteristics of consecutively recorded images of a yarn package or of different yarn packages are thereby independent of the external parameters. The use of quality codes for controlling, for example, a rewinding speed is thereby possible without adaptation of the control to the external parameters. In a simple way, the quality index can be calculated by summing the data values of one image at a time.

An inventive method for rewinding yarns with a rewinding speed has the following method steps:

- Unwinding a yarn from a yarn package, and

- Winding the yarn on a rotating with a take-up yarn carrier, wherein the yarn is fed over a yarn path from the yarn package to the yarn carrier.

In this case, a Umspulgeschwindigkeitskontrolle size of the yarn is determined by the yarn package continuously and regulated the Umspulgeschwindigkeit of a control module in response to the Umspulgeschwindigkeitskontrolle size. According to the invention, the rewinding speed control variable used is the quality characteristic value calculated by the method according to the invention for the automatic classification of the quality of a yarn package. This rewinding method allows the rewinding speed control to be controlled by the actual reel quality changing during rewinding without additional yarn load to measure the thread tension of the yarn. If the rewinding speed is reduced as the quality index increases and the rewinding speed is increased as the quality index decreases, coarse structures, i. a coarse winding of the yarn on the current bobbin surface reduces the rewinding speed, thereby avoiding yarn tension maxima.

An inventive bobbin quality grading device is set up to carry out the method according to the invention for automatically classifying the quality of a yarn bobbin. The Gamspulenqualitäteinstufungsvorrichtung invention has a camera and a

Calculation module on. The camera is for optically detecting at least one Image of the yarn package, and the computing module is adapted to convert pixels of a digital version of the image by means of a transformation algorithm into data values and to calculate a quality score from the data values. The yarn package quality grading apparatus of the present invention provides the advantages of the inventive method for automatically grading the quality of a yarn package.

A rewinding device according to the invention for rewinding yarns with a rewinding speed is set up for carrying out the rewinding method according to the invention. The rewinding device according to the invention comprises:

a supply spool receptacle adapted to receive a yarn package with a yarn to be spooled, and

a yarn carrier receptacle rotatable with a take-up speed, adapted for receiving a yarn carrier on which the yarn can be wound up, wherein the yarn can be guided over a yarn path from the yarn package to the yarn carrier.

According to the invention, a yarn package quality grading device according to the invention is provided. The yarn package quality grading device is set up as

Umspulgeschwindigkeitskontrolle size of the yarn from the yarn package continuously to determine the quality index, wherein a control module is provided, configured to control a Umspulaggregat for regulating the Umspulgeschwindigkeit in response to the Umspulgeschwindigkeitskontrolle size. The rewinding device according to the invention provides the advantages of the rewinding method according to the invention and can be advantageously used in yarn processing machines.

The invention will be explained in more detail by means of embodiments with reference to the drawings. FIG. 1 shows a rewinding device according to the invention.

Figures 2a and 2b each show an arrangement of light sources for

Lighting the yarn package.

FIGS. 3a and 3b respectively show an image of a yarn package and a screen photograph of a representation of the data values of the conversion of the yarn package

Pixels by means of a Fourriertransformation.

FIGS. 4a and 4b respectively show an image of a yarn package and a yarn package

Representation of the data values of the conversion of the pixels into gray values.

Figures 5a and 5b each show a representation of quality characteristics of 11 bobbins, the quality characteristics in Figure 5a by

Summing up data values obtained by means of a Fourier transformation were determined and the quality parameters in FIG. 5b

Sum up data values obtained by converting to gray scale values.

FIG. 1 shows a rewinding device 1 according to the invention for rewinding yarns. The yarn is unwound from a yarn package 3 used as a supply spool located in a supply spool receptacle 4. A yarn carrier 7, on which the yarn is wound, is in one

Garnträgeraufnahme 8 arranged. A rewinding unit 9 designed as a friction roller, which is designed as a winding unit for driving the yarn carrier receptacle 8, drives the yarn carrier receptacle 8 such that it rotates at a winding speed. Generally, the rewinding unit 9 is an electric motor engaged with the yarn such that it can at least affect the rewinding speed. The yarn is guided while wrapping in a yarn path IO of the supply reel 3 to the yarn carrier 7. In the yarn path 10, a yarn processing module 12 is further arranged. This may be, for example, an oiling device. In order to regulate the rewinding speed, that is to say in the case of the take-up speed, as a function of a To allow Umspulgeschwindigkeitskontrolle size, a Garnspulenqualitufungsvorrichtung invention is provided. The yarn bobbin quality grading device continuously determines a quality index, or values of quality codes, as the yarn rewinding speed control. By means of a control module 14, the rewinding unit 9 for regulating the rewinding speed in dependence on the

Umspulgeschwindigkeitskontrolle size, ie the values of the quality index, controlled. For this purpose, the control module 14 is connected to the rewind unit 9 via a cable. The yarn package quality grading apparatus comprises a light source 20 adapted to illuminate the yarn package 3, and a camera 21 arranged to optically detect at least one image of the yarn package 3. A yarn package quality grading apparatus computing module 25 is for converting pixels of a digital version of the image Determining, from pixels of digital versions of continuously acquired images, by means of a transformation algorithm into data values and to calculate the quality metrics from the data values. The latter is shown in the figure as a flowchart in the stylized representation of the computing module 25. In order to take over the images from the camera 21, the computing module 25 is connected to the camera 21 via a cable.

The light source 20 illuminating the yarn package 3 is positioned so that the beam direction 27 of the light source 20 forms an acute angle α with the longitudinal axis 29 of the yarn package 3 so that bumps, e.g. caused by a poor winding, cast as much shadow on the Garnspulenoberfläche as possible. As a result, the illustrated arrangement is particularly suitable for the application of a transformation algorithm in which a conversion of the digital pixels into gray values is carried out, wherein the data values from which the quality index is determined are calculated as the gray values.

FIGS. 2a and 2b each show an arrangement of light sources 20 for illuminating a yarn package 3 in an application of the invention Shown method. The camera 21 is in each case with its take-up direction 30 perpendicular to the longitudinal axis 29 of the yarn package 3. The yarn package 3 is arranged in each case an original pair intake 4. In FIG. 2 a, two light sources 20 are arranged symmetrically with respect to the receiving direction of the camera 21 at intervals of approximately 30 cm to the yarn package 3, which leads to a uniform illumination of the surface of the yarn package 3. A uniform illumination is advantageous in the case of an application of a Fourier transformation as a transformation algorithm, the data values being calculated as the amplitudes of the Fourration transformation thus produced in the frequency domain.

In Figure 2b, light source 20 is positioned as in Figure 1 such that the beam direction of light source 20 forms an acute angle with the longitudinal axis 29 of yarn package 3 so that bumps, e.g. caused by a poor winding, cast as much shadow on the Garnspulenoberfläche as possible. As a result, the arrangement shown in FIG. 2b is particularly suitable for the application of a transformation algorithm in which a conversion of the digital pixels into gray values is carried out.

FIGS. 3a and 3b each show an image of a yarn package, image section 7.5 cm by 10 cm, and an illustration of the data values of the conversion of the pixels by means of a Fourier transformation. In the upper illustration of FIG. 3a, an image of the surface of a relatively coarse, ie poor, wound yarn package with pronounced mirror zones is shown. An arrow marked Λ is drawn in the image to symbolize that the image can be interpreted as a two-dimensional (2D) representation of waves in space. It can already be seen that the surface structure generated by the coarse winding is repeated with a long wavelength. The lower graph shows a visualization of the amplitude of a discrete 2D Fourier transform of the digitized image in frequency space, i. a representation of the amplitudes of the image underlying

Frequencies in a two-dimensional (x- and y-frequency) frequency domain, wherein the frequencies are normalized in both coordinate directions each to a maximum value of 100.

FIG. 3b shows a representation corresponding to FIG. 3a of an image in the positional space of a uniformly wound yarn package and the discrete 2D Fourier transformation of this image.

When comparing the representations of the Fourier transforms, it can be seen that in the poorly wound yarn package of FIG. 3a relative to the uniformly wound yarn package of FIG. 3b relatively large amplitude values occur at low frequencies, ie pixels in the representation of the Fourier transformation in the region of the origin of coordinates. In order to obtain values of quality indices independent of the illumination, the amplitudes of the Fourier transformation were normalized with the total intensity of the original image. The latter can be seen directly from the spatial representations, since there the coarse structures of a bad winding lead to long-wave image components. Summing up the amplitude values therefore leads to a higher quality index in the case of the poorly wound coil than in the case of the uniformly wound coil. So here is the quality index inversely proportional to the coil quality. The regulation of the rewinding speed must therefore be carried out in such a way that the rewinding speed decreases as the quality index increases.

FIGS. 4a and 4b each show an image of a yarn package and a representation of the data values of the conversion of the associated pixels into gray values. FIG. 4a shows a yarn package with a uniform winding, and FIG. 4b shows a yarn package with a coarse winding. Correspondingly, in the images converted into gray values, in each case to the right of the original images of the yarn packages, substantially more pronounced shadows can be seen in FIG. 4 b than in FIG. 4 a. That is, in FIG. 4b substantially more pixels with high gray values, ie dark pixels, are present than in FIG. 4a. A summation of the gray values therefore leads to a in FIG. 4a lower quality index than in Figure 4b. The quality parameters thus behave here relative to the coil quality as in the case of the described in Figure 3a and 3b Fourriertransformation.

FIGS. 5a and 5b each show a representation of quality characteristics of 11 bobbins. In this case, the quality indices in FIG. 5a were determined by summing up data values obtained by means of a Fourier transformation, and the quality indices in FIG. 5b were determined by summing up data values obtained by converting to gray values.

The three graphs of FIG. 5a formed by the quality characteristics of the eleven yarn packages were obtained by summing different frequency ranges of the Fourier transformation. The quality indexes of the 11 bobbins were calculated by summing 20 by 20, 15 by 15, and 10 by 10 data values in the lower frequency range, i. in the area of

Coordinate origin, calculated. It turns out that even summing relatively low frequency ranges is sufficient to determine a meaningful quality score, since the relations of the quality measures of the different summations are independent of the resolution, so that a relative quality statement is possible even when adding up a low frequency range.

Also in the case of the quality characteristics shown in FIG. 5b, which are obtained by summing up gray values, the relations of the quality characteristics are essentially unchanged, so that the quality ratings of the coils correspond to the corresponding quality ratings from FIG. 5a.

Proposed is a method for automatic classification of the quality of a yarn package 3, in particular for detecting mirrors of the yarn package 3, by means of a computing module 25, with the method steps optical detection at least one image of the yarn package 3 by means of a camera 21, converting pixels of a digital version of the image by means of a transformation algorithm into data values and calculating a quality index from the data values and an apparatus for performing the method.

The invention is not limited to the embodiments given above. Rather, a number of variants are conceivable, which make use of the features of the invention even with fundamentally different type of execution.

Claims

claims

1. A method for automatic classification of the quality of a yarn package (3), in particular for detecting mirrors of the yarn package (3), by means of a computing module (25) with the method steps

- Optical detection of at least one image of the yarn package (3) by means of a camera (21),

- Converting pixels of a digital version of the image by means of a transformation algorithm into data values and

- Calculate a quality score from the data values.

2. The method according to claim 1, characterized in that as a transformation algorithm a Fourriertransformation is applied, wherein the data values are calculated as the amplitudes of the thus generated Fourriertransformation in the frequency domain.

3. Method according to claim 2, characterized in that the data values are weighted from the data values by weighting with weighting factors before the quality code is calculated, preferably wherein the weighting factors are inversely proportional to the frequency values of the associated amplitudes.

4. The method according to claim 1, characterized in that a transformation into gray values is used as a transformation algorithm, wherein the data values are calculated as the gray values.

5. The method according to at least one of claims 3 to 4, characterized in that by means of a light source (20) the yarn package (3) is illuminated, wherein the light source (20) is positioned such that the beam direction (27) of the light source ( 20) forms an acute angle with the longitudinal axis (29) of the yarn package (3).

6. The method according to at least one of claims 1 to 5, characterized in that a normalization of the data values is carried out, wherein the

Quality index is calculated from the normalized data values.

7. The method according to at least one of claims 1 to 6, characterized in that the quality index is calculated by adding up the data values.

8. A method for rewinding yarns at a rewinding speed, comprising the steps of - unwinding a yarn from a yarn package (3), and

- winding the yarn onto a take-up speed rotating yarn carrier (7), the yarn is fed via a yarn path (10) from the yarn package (3) to the yarn carrier (7), wherein a Umspulgeschwindigkeitskontrolle size of the yarn from the yarn package

(3) is determined continuously and the rewinding speed is controlled by a control module (14) as a function of the rewinding speed control variable, characterized in that the rewinding speed control variable used is the quality index number calculated according to at least one of claims 1 to 7, Preferably, with increasing quality index, the rewinding speed is reduced and the rewinding speed increases as the quality index decreases.

9. Garnspulenqualitäteinstufungsvorrichtung, set up for

Implementation of the method according to at least one of claims 1 to 7, with

- A camera (21), arranged for optically detecting at least one image of the yarn package (3), and - a computing module (25), arranged for converting from

Pixels of a digital version of the image by means of a transformation algorithm into data values and to calculate a quality score from the data values.

10. rewinding device (1) for rewinding yarns with a

Umspulgeschwindigkeit, adapted for carrying out the method according to claim 8, with

- A supply spool receptacle (4) adapted to receive a yarn package (3) with a yarn to be abzuspulten, and - a rotatable with a winding speed

A yarn carrier receptacle (8) adapted to receive a yarn carrier (7) on which the yarn can be wound, the yarn being able to be guided via a yarn path (10) from the yarn package (3) to the yarn carrier (7), characterized in that a yarn package quality grading device is provided according to claim 8, arranged as Umspulgeschwindigkeitskontrollgrößen the yarn from the yarn package (3) continuously determine the Qualitätskenπzahl, wherein a control module (14) is provided, furnished, a Umspulaggregat (9) for regulating the Umspulgeschwindigkeit depending on the Umspulgeschwindigkeitskontrolle size Taxes.

11. Yarn processing machine with at least one rewinding device according to claim 10.