WO2007010325A1

WO2007010325A1 - Detecting foreign substances in a textile material

Info

Publication number: WO2007010325A1
Application number: PCT/IB2005/003167
Authority: WO
Inventors: Shekaripuram Narayanaswamy Ramachandran; Narayanaswamy Kavitha; Balasubramaniam Shamuga Sundaram
Original assignee: Premier Evolvics Pvt. Ltd.
Priority date: 2005-07-22
Filing date: 2005-10-13
Publication date: 2007-01-25
Also published as: EP1907830A1; CN101228434A; EP1907830B1; JP2009503276A; CN101228434B; JP5058161B2

Abstract

An apparatus for detecting foreign substances in a strand-like textile material comprises a first light source (2) for emitting visible light, a second (3) light source for emitting radiation in the infrared spectrum, a first detector (4) sensitive to visible light reflected from a the textile material (1, 11) and configured to generate a corresponding first detector signal, and a second detector (5) sensitive to radiation in the near-visible spectrum reflected from the textile material (1, 11) and configured to generate a corresponding second detector signal. The light sources (2, 3) and detectors (4, 5) are arranged to illuminate and view the same section of the yam (1), and a signal processing section (7) is configured to determine the ratio and a weighted difference of the two detector signals and to output a signal indicative of the presence of a contaminant whenever said difference exceeds a threshold value.

Description

DETECTING FOREIGN SUBSTANCES IN A TEXTILE MATERIAL

FIELD OF THE INVENTION

The invention relates to an apparatus and a method for detecting foreign substances in a textile material, in particular for detecting the presence of contaminants in a strand-like textile material such as a yarn, as described in the preamble of the corresponding independent claims.

BACKGROUND OF THE INVENTION

US 6,175,408 describes a measurement apparatus emitting white light and measuring the intensity of light reflected from a travelling yarn in the red, green and blue frequency bands. For each of the three possible pairs of intensity signals a difference signal is formed. A contamination is deemed to be detected when at least one of the difference signals exceeds a given threshold. The focus of the invention is on the use of a frequency transformer for generating white light over a wider frequency spectrum, avoiding the use of narrow-band, single colour LED's. EP 0 399 945 Bl also shows an apparatus with a polychromatic light source in which the difference between the received intensities for two colours is used to detect a contaminant, hi an alternative embodiment, light is emitted at two wavelengths and detected by a broadband detector, as in the following reference:.

WO 03/008950 shows a device emitting light at two different wavelengths, and having a single light receiver. Since the dependency of reflection on wavelength differs according to material, the amount of total reflected light depends on the presence and the kind of contaminant material. In this manner, it should be possible to distinguish between different contaminants.

US 5,915,279 shows a measurement setup for detecting parameters of an object in general terms. A polychromatic light source emitting light at at least two different wavelengths, and a plurality of detectors for different colour bands are used. Methods for the statistical analysis of a plurality of parameters are described, e.g. light intensities detected at different wavelengths, and including time histories of intensities in order to determine deviations. However, the resulting algorithms are complicated and too general to be immediately applicable.

EP 0 652 432 Al discloses the use of sensors for two wavelengths, with the optical arrangement ensuring that the light received by the sensors comes from exactly the same location of the yarn. The light received may lie in the infrared or near-infrared spectrum. From the detector signals corresponding to received intensities, a signal corresponding to a quotient (i.e. a difference of logarithms) is formed, in order to eliminate the influence of yarn diameter, structure etc. DESCRIPTION OF THE INVENTION

It is an object of the invention to create an apparatus and a method for detecting foreign substances in a textile material of the type mentioned initially, which is simple to implement and improves the detection quality.

These objects are achieved by an apparatus and a method for detecting foreign substances in a textile material according to the corresponding independent claims.

The apparatus for detecting foreign substances in a textile material, in particular a strand-like textile material such as a yarn, comprises a first light source for emitting visible (VIS) light, a second light source for emitting radiation in the infrared (IR) spectrum, a first detector sensitive to visible light reflected from the textile material and configured to generate a corresponding first detector signal, and a second detector sensitive to radiation in the IR spectrum reflected from the textile material and configured to generate a corresponding second detector signal. The first and second light sources and first and second detectors are arranged to illuminate and view the same section of the textile material, and a signal processing section is configured to determine a weighted difference of the two detector signals and outputting a signal indicative of the presence of a contaminant whenever said difference exceeds a threshold value. "Weighted difference" means that at least one of the detector signals is multiplied by a substantially constant factor before the difference is determined.

The second light source emits light in the near-visible range, i.e. either as an infrared (IR) light source as presented so far, or alternatively as an ultraviolet (UV) light source.

The invention may be applied to an elongated, strand-like textile material such as a yarn or roving or sliver, but also to chunks of cotton being conveyed or blown past - A -

the detecting arrangement. In general, contaminants as well as the textile material reflect light both in the Visible and the Infrared Spectrum. However, the proportion of reflected light in the two spectra varies between a yarn and a contaminant. For example, with a black yarn as a contaminant, the VIS detectors will have less reflected light in the Visible Spectrum but the Infrared detectors will still observe reflected light roughly the same as for the normal yarn. Variations in the yarn diameter, hairiness, position etc. affect both the spectrums proportionally.

The reflection of the yarn in the visible (VIS) and the infrared (IR) spectrum are recorded from a normal running yarn. The two recorded signals may be combined to form a ratio or spectral ratio for the normal running yarn by a long term averaging process. In a preferred embodiment of the invention, this long term average of this spectral ratio for the normal running yarn is normalized to unity. Variations in yarn diameter and hairiness of a yarn to be tested affect both the reflected visible (VIS) and the infrared (IR) spectrum in a similar manner. Typical contaminants distinctly change the reflection behaviour in the visible (VIS) and infrared (IR) spectrum compared to that of the reference yarn, i.e. uncontaminated yarn material.

In another preferred embodiment of the invention, due to the differences in light intensity, sensor efficiency and reflectivity in the two bands of radiation, the two detected signals are weighted with respect to one another, such that their weighted difference becomes zero for the normal running yarn. A weighting coefficient can be adapted in accordance with long-term averages of the first and second detector signal. In a preferred embodiment of the invention, the weighted difference Dw is normalised with respect to the second detector signal, e.g. by dividing it by the IR detector signal.

In a further preferred embodiment of the invention, the first and second detector are arranged to observe, i.e. to measure radiation from a first angle with respect to the yarn, and the device comprises a further VIS and a further IR detector arranged to observe, from a second angle with respect to the yarn. The first and second angle show their maximum value when projected onto a plane that is perpendicular to the running yarn. The first and second detectors constitute a first detector set, and the further detectors constitute a second detector set. The further detectors are sensitive to the same wavelengths as the first and second detector, respectively.

In yet a further preferred embodiment of invention, a third detector is placed opposite to the light sources, to detect the positional variations of the yarn. The third detector is sensitive in either the VIS or the IR or both frequency ranges. Preferably, the position detectors generate detector signals that are used in classifying situations where the reflected signals alone may not take into account certain disturbances. For example, the spectral ratio or the weighted difference Dw may also change when a position variation of the yarn occurs. Taking the third detector signal into account, a position variation can be distinguished from a contaminant signal.

Preferably, the position detectors generate detector signals that are used in classifying situations where the reflected signals alone may not take into account certain disturbances.

In a further preferred embodiment of the invention, the first and second light source are identical, that is, a single source or a combined source emits both the visible and the IR light used for the measurements.

The method for detecting foreign substances in a strand-like textile material, comprises the steps of

• emitting visible light directed at a section of a yarn,

• emitting radiation in the infrared spectrum directed at the same section of the yarn,

• detecting visible light reflected from the yarn and generating a corresponding first detector signal, • detecting radiation reflected from the yarn in the infrared spectrum and generating a corresponding second detector signal,

• determining a weighted difference of the two detector signals,

• and outputting a signal indicative of the presence of a contaminant whenever said difference exceeds a threshold value.

In a further preferred variant of the invention, given the two detector signals corresponding to reflected light in the VIS and IR spectrum, a combination of at least two of the following quantities or factors is used: - a deviation of the ratio of the two detector signals

- a deviation of the visible spectrum detector signals

- a deviation of the IR spectrum detector signals

- a ratio of deviations of the two detector signals

The respective deviations are preferably determined with respect to constant reference values or long term reference values.

In particular, this may done according to the following method:

• Determine the ratio of the two spectral signals. Compare the instantaneous signal value with a minimum threshold limit which is a small percentage above than of the yarn average. If the signal exceeds this limit for a predetermined length, then qualify this as being a contaminant.

• Compute the deviation in the spectral signal in the visible spectrum with respect to the reference material i.e. yarn. This deviation allows to identify a first class of contaminants such as Hair, Seed Coats etc. • Compute the deviation in the spectral signal in the IR spectrum with respect to the reference material. This deviation allows to identify a second class of contaminants such as Polypropylene (PP).

• Compute the ratio of the two deviations and use this ratio to further categorise the contaminant as being one of specific materials like Hair, Seed Coats, PP, Jute, Coloured Yarn etc. Further preferred embodiments are evident from the dependent patent claims. Features of the method claims may be combined with features of the device claims and vice versa.

BRIEF DESCRIPTION OF THE DRAWINGS

The subject matter of the invention will be explained in more detail in the following text with reference to preferred exemplary embodiments, which are illustrated in the attached drawings, in which:

Figure 1 schematically shows a first testing arrangement;

Figure 2 schematically shows an arrangement of sources and detectors; Figure 3 schematically shows another arrangement of sources and detectors;

Figure 4 schematically shows a testing arrangement using a position sensor;

Figure 5 shows an arrangement for observing a flow of cotton chunks;

Figure 6 shows an exemplary trajectory of signals representing reflected visible and IR light; Figure 7 shows an exemplary trajectory of a weighted difference signal; and

Figure 8 illustrates a classification method according to a preferred embodiment of the invention.

The reference symbols used in the drawings, and their meanings, are listed in summary form in the list of reference symbols. In principle, identical parts are provided with the same reference symbols in the figures. DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

Figure 1 schematically shows a first testing arrangement with a yarn 1 seen head-on, a first source 2 and second source 3 arranged to illuminate the same section of the yarn 1 and a first detector 4 and second detector 5 arranged to receive light reflected from said section of the yarn 1. The first source 2 and second source 3 are powered by a driving circuit 6. The signals from the first detector 4 and second detector 5 are processed by a signal processing section 7 by means of analog or digital signal processing elements. The first detector 4 and second detector 5 constitute a first detector pair 10 arranged to observe a first side of the yarn 1.

Figure 1 is highly schematic in that it only shows the functional relation between the elements, but not the true geometric relations. Figure 2 shows an arrangement of sources and detectors for observing the yarn 1. The sources and detectors are arranged in two rows along the length of the yarn 1, with the source for each part of the spectrum (IR, VIS) observed lying in one row, and two corresponding detectors for that part of the spectrum (IR, VIS) lying in the other row. For example, the first source 2, emitting in the visible spectrum, lies in the right row, and the visible light (VIS) detector 4 and a further VIS detector 4' lie in the left row. The second source 3 emitting in the infrared (IR) spectrum and the IR detector 5 and a further IR detector 5' are arranged vice versa, hi a preferred embodiment of the invention, the signals of the VIS detector 4 and the further VIS detector 4' are added to provide the VIS detector signal for further processing, and the IR detector 5 signal and further IR detector 5' are added to give the IR detector signal for further processing, hi another preferred embodiment the various signals of the VIS and the IR Detectors 4, 4', 5, 5' could be processed independently.

For practical reasons, since VIS detectors usually are sensitive to IR light as well, the VIS detectors used here include an IR blocking filter. The arrangement according to Figure 2 will observe a major part of one side of the yarn 1, e.g. ca. 135° of the yarn's circumference, which is sufficient to detect contaminants wound around the yarn 1.

Figure 3 shows another preferred arrangement of sources and detectors for observing the yarn 1. The arrangement is shown in a view perpendicular to the yarn 1 and a roughly planar arrangement of sources and detectors (top), and in a view in the direction of the yarn 1 (bottom). The sources and detectors are arranged in three rows along the length of the yarn 1, the first row comprising a VIS detector 4 and a further IR detector 5', the second row comprising the first source 2 and the second source 3, and the third row comprising an IR detector 5 and a further VIS detector 4'. Having two detector pairs 4, 5'; 5, 4' lying essentially in a plane along the yarn 1 and observing the yarn 1 from different sides gives a good coverage of the test material under consideration. The signals from the second pair are independently processed by means of analog or digital processing elements.

Li the signal analysis for the two pairs of sensors, the signals of each pair of IR/VIS sensors may be processed as shown below, and the output signals detecting the presence of a contaminant may be AND / OR combined. Alternatively, the signals of the two visible light detectors 4, 4' are summed and the signals of the two IR detectors 5, 5' are summed, and the summed signals are processed as shown below.

Figure 4 schematically shows a second testing arrangement, further comprising a third detector 24 operating as a position detector, and whose detector signals are e.g. processed by the signal processing section 7 as well. The position detector is arranged, with respect to the yarn 1, in opposition to the respective source, such that the light emitted by the source casts a shadow of the yarn 1 on the detector. The position detector 24 is sensitive to the visible and/or IR light.

Depending on the position of the yarn in the detecting apparatus, the position sensor signal will vary. Similar variations may also be observed due to changes in thickness or hairiness of the yarn (and can be used to determine thickness and/or hairiness). Therefore, the actual state and position of the yarn is inferred from a combination of the position signal and the reflected signals in the visible and IR spectrum. This inference or discrimination of different states is based on the following principles:

• With no yarn 1 present, the position detector 24 receives the maximum light intensity, denoted in the following with the position signal value 100.

• When the yarn 1 is placed inside the detector slot in its nominal position, then the yarn 1 obstructs the light falling on the detector 24 and hence less light is received. The position sensor value lies around 100-k, with (k>0). The reflected signals in the visible and IR spectrum show their nominal or reference values, denoted b and c, respectively.

• A thicker yarn 1 results in even less light received, and a thinner yarn in more light received by the position detector 24, corresponding to position signal values of 100-k' and 100-k", respectively, where e.g. k'=2k and k"=k/2. The reflected signals increase for a thicker and decrease for a thinner yarn, e.g. to double or half their nominal value. Depending on actual circumstances, another factor than 2 may be chosen to define a thick or thin yarn.

• When yarn 1 moves closer to the position detector 24, the shadow cast becomes thinner, e.g. with a position signal value 100-j, with (j< k). The reflected signals decrease, e.g. to values b' and c', respectively, with (b'<b) and (c'<c).

• When then yarn 1 moves towards the source, away from the position detector 24, more light is blocked, and the position signal becomes e.g. 100-1, with (l>k). The reflected signals increase, e.g. to values b" and c", respectively, with (b">b) and (c">c).

• When the yarn moves outside the slot sideways from its intended position, it blocks light of lesser intensity. Hence the position signal becomes e.g. 100-m with (m < k). Due to the asymmetric arrangement of the visible and IR detectors, the reflected signal of one of them shall increase, and the other one decrease, e.g. to values around 2b for the visible and c/2 for the IR detector. Again, another factor than 2 may be appropriate. Eventually, when the yarn 1 leaves the detection zone, the position signal becomes 100, corresponding to a no yarn situation.

The same holds for the position signal when the yarn 1 moves sideways into the slot, with a reference value 100-n with (n<k). However, the increase and decrease of the visible and IR reflected signals is the inverse from the above case. That is, the visible signal is around b/2 and the IR signal is around 2c. Again, another factor than 2 may be appropriate.

A contamination spirally wound around the yarn will not change the position signal. The reflectance in the visible range shall be around a value x and in the IR range around a value y.

The actual values for the above constants b, b', b", c, c¹, c", i, j, k, 1, m, n, x, y are determined experimentally according to the exact geometry of the detector arrangement and the properties of the yarn and contaminants.

The following table gives an overview of the different states that are distinguished by the different combinations of sensor values mentioned above. The second, third and fourth column give reference values for position, visible and IR sensor values corresponding to the respective state.

In a preferred embodiment of the invention, an electronic circuit or computer program distinguishes these cases by checking whether the measured values for each of the sensor are within a predetermined tolerance around the respective reference value. This checking can be done sequentially or in parallel or through a decision tree etc.

In a preferred embodiment of the invention, the first source 2 emits visible light and the second source 3 emits infrared light. In another preferred embodiment of the invention, the second source 3 emits ultraviolet light. Thus, in both cases, the light emitted by the second source 3 is in the near- visible spectrum.

Figure 5 shows another embodiment of the invention wherein the same arrangement of sources 2, 3 and detectors 4, 4', 5, 5' is utilized to detect contaminations 12 in chunks of cotton 11. Such arrangements are normally on the preparatory stages of the spinning mills. The detecting setup views the flow of the material that is blown through a material flow duct 9. The flow is viewed via a window, optionally incorporating a lens 8, which lets pass both the visible and infrared light. An array of the sensor arrangement 4, 4', 5, 5' is placed to cover the entire field of view, i.e. the entire cross section of the material flow duct 9. The signal processing is similar to detecting contaminants in a moving thread, but here serves to discriminate contaminant material 12 from chunks of cotton.

The following figures illustrate the characteristics and the use of detecting visible and infrared light. Figure 6 shows an exemplary trajectory of signals representing reflected visible and IR light along the length l_y of the yarn being inspected. It can be seen that most of the time, or for most of the length of the inspected yarn 1, the two detector signals move in parallel. This movement corresponds to e.g. thickness, density, fluctuation, fibre distribution and hairiness variations. Between length 127 and 136 the detected visible light, with respect to the detected infrared light, drops of significantly.

Figure 7 shows an exemplary trajectory of a weighted difference of the two signals. The weighted difference Dw, also called yarn deviation, is normalised with respect to the second detector signal, and is preferably computed according to Dw = ( (k * Sv) - Si ) / Si where Sv is the value of the first detector signal - for example, from the visible light detector - and Si is the value of the second detector signal - for example, from the IR detector - and k is a substantially constant weighting coefficient.

"Substantially constant" means that the value of k remains essentially the same and changes only slowly with regard to the values of the detector signals, e.g. with a time constant that is more than a hundred or a thousand times larger than the time constants of the changes in the detector signals. The weighting coefficient k can be adapted in accordance with a long-term average LTA of the first and second detector signal, e.g. as k = LTA( Si ) / LTA ( Sv ) .

Where the LTA is the average computed over a time window of e.g. several seconds or minutes or tens of minutes, or over a length of several meters of running yarn.

Alternatively, the yarn deviation may be computed by computing the ratio Rs = Sv/Si and from this the yarn deviation Dw = (Rs - Ra)/Ra, with Ra = 1/k or the long term average of Sv/Si, that is, Ra = LTA(SWSi).

As seen in Figure 7, the weighted difference or yarn deviation remains within a certain band, and for the position (between length 127 and 136) in which a substantial difference between the two signal occurs, the weighted difference exceeds said band, indicating the presence of a contamination. For example, the predetermined threshold for detecting a contamination may be set to a small percentage value (a percent value of k as defined above is obtained by multiplying k with 100).

When illuminating the yarn 1 with the second source 3 (and fourth source 23) being ultraviolet emitters, with wavelengths e.g. from 254nm to 395nm, then certain contaminants such as polypropylene will fluoresce and convert UV light to visible light.

The position sensor signals are used as follows, e.g. in cases in which the weighted difference alone does not discriminate among different situations. The position signals • may be the detector signals themselves, which are approximately proportional to the amount of light received by the detector, • or they may be complementary signals, which are approximately proportional to the size of the shadow of the yarn 1 on the detector surface, on the respective VIS /IR position detector 24. The signal of the position detector 24 alone is used, in a preferred embodiment of the invention, to distinguish contaminations from vibrating yarn conditions. If the weighted difference indicates a contamination, which may also be caused by other disturbances, e.g. by the yarn vibrating, then the position signal is examined. If its change exceeds a predetermined threshold, then the situation is classified as being a vibration of the yarn instead of a contamination. Otherwise, a contamination is indicated and the corresponding section of the yarn 1 is cut out.

For detecting a change in the third detector signal, the position signal is e.g. compared to a substantially constant reference value such as a long term average. The flowchart of Figure 8 illustrates a classification method according to a further preferred variant of the invention. In the initialisation step 100, labelled "start", processing begins. It is assumed that prior to this, a long term average reference value for the type of yarn being examined has been determined both in the visible (Yv) and in the infrared spectrum (Yi). From this, a reference ratio Y = Yv/Yi is computed. Alternatively an average of the ratios Y = LTA (Sv/Si ) is computed.

In the input step 101, labelled "inp", a simultaneous sampling of the outputs of the VIS and IR detectors is performed. The observed signals are Sv and Si.

For the first comparison step 102, labelled "V/I?", the ratio S = Sv/Si is computed and compared to the reference ratio Y. If the absolute value of the deviation Dr = (Y-S)/Y exceeds a predetermined first threshold value, then the presence of a probable contaminant is indicated, and the procedure continues to categorise the contaminants by means of further steps, beginning with the second computation step 103. Otherwise, no contamination is indicated, and a further set of samples is processed in step 101.

For the second computation step 103, labelled "dV", a deviation Dv of Sv from Yv is computed. Preferably, Dv is normalised with respect to Yv, such that Dv = (Yv- Sv)ZYv. If the absolute value of the deviation Dv exceeds a predetermined second threshold value, then the presence of a certain probable types of contaminants, e.g. hair or seed coats, is confirmed, and the procedure continues to verify this by means of further steps, beginning with the third computation step 104.

For the third computation step 104, labelled "dl", a deviation Di of Si from Yi is computed. Preferably, Di is normalised with respect to Yi, such that Di = (Yi-Si)/Yi. If the absolute value of the deviation Di exceeds a predetermined third threshold value, then the presence of another set of contaminants such as Polypropylene is confirmed, and the procedure continues to verify this by means of the fourth computation step 105.

For the fourth computation step 105, labelled "dV/dl", the ratio Dv/Di is computed. This value is compared with predetermined spectral behaviour of different materials, i.e. yarn and contaminant materials, and contaminants are categorised accordingly.

If the Dr values indicating the presence of a specific contaminant exceed a predetermined clearing first threshold limit, the contaminant is to be cut from the yarn. This is done in step 106 labelled "cut", and a further set of samples is processed from step 101. A qualified contaminant is categorised and cut from the yarn.

Threshold values depend on the actual reference yarn and contaminant materials, and also on the type of sensor used and the spectral sensitivity curves of the sensors. For example, the following table gives measured values for the intensity of reflected light in the visible an IR spectrum for different materials.

The values observed in each spectrum in the contaminated yarn may be estimated analytically as being a weighted sum of the values from the contaminant and from the reference, the weight of the contaminant being typically between 15% and 50%.

The other characteristic values or ratios described in the above are computed accordingly. Given these characteristic values and their stochastic distribution, the thresholds for distinguishing between different classes of contaminants and individual contaminants can be determined. While the invention has been described in present preferred embodiments of the invention, it is distinctly understood that the invention is not limited thereto, but may be otherwise variously embodied and practised within the scope of the claims.

LIST OF DESIGNATIONS

1 yarn

2 first source

3 second source

4, 4' first detector

5, 5' second detector

6 driving circuit

7 signal processing section

8 lens

9 material flow duct

10 first detector pair

11 cotton chunks

12 contamination

24 third detector, position detector

Claims

P A T E N T C L A I M S

1. An apparatus for detecting foreign substances in a textile material, in particular a strand-like material such as a yarn (1), or a flow of pieces of textile material (11), comprising a first light source (2) for emitting visible light, a second (3) light source for emitting radiation in the infrared spectrum, a first detector (4) sensitive to visible light reflected from the textile material and configured to generate a corresponding first detector signal, and a second detector (5) sensitive to radiation in the infrared spectrum reflected from the textile material and configured to generate a corresponding second detector signal, the light sources

(2, 3) and detectors (4, 5) being arranged to illuminate and view the same section of the textile material, and a signal processing section (7) configured to determine a weighted difference of the two detector signals and outputting a signal indicative of the presence of a contaminant whenever said difference exceeds a threshold value.

2. The apparatus of claim 1, configured to normalise the weighted difference with respect to the second detector signal.

3. The apparatus of claim 2, wherein the signal processing section (7) is configured to determine the weighted difference Dw as

Dw = ( (k * Sv) - Si ) / Si where Sv is the value of the first detector signal, and Si is the value of the second detector signal, and k is a substantially constant weighting coefficient.

4. The apparatus of one of claims 1 to 3, wherein the second light source (3) is an infrared light source.

5. The apparatus of one of claims 1 to 3, wherein the first and second light source (2, 3) are identical, that is, form a combined source.

6. The apparatus of one of the preceding claims, with the textile material being a yarn (1) and the apparatus further comprising at least a third detector (24) a Ls_ position detector arranged, with respect to the yarn (1), in opposition to the corresponding light source (2, 3), and wherein the signal processing section (7) is configured to use a signal from the third detector (24) in order to indicate the presence of a contaminant.

7. The apparatus of claim 6, wherein the processing section (7) is configured to indicate the presence of a contaminant only if a change in a position signal does not exceed a predetermined threshold.

8. The apparatus of one of the preceding claims, configured to adapt the weighting coefficient k in accordance with a long-term average of the first and second detector signal.

9. A method for detecting foreign substances in a textile material, in particular a strand-like textile material, comprising the steps of • emitting visible light directed at a section of a the textile material,

• emitting radiation in the infrared spectrum directed at the same section of the textile material,

• detecting visible light reflected from the textile material and generating a corresponding first detector signal,, • detecting radiation in reflected from the textile material in the infrared spectrum and generating a corresponding second detector signal,

• determining a weighted difference of the two detector signals,

10. The method of claim 9, comprising the step of

• normalising the weighted difference with respect to the second detector signal, wherein the normalisation is done by determining the weighted difference Dw as Dw = ( (k * Sv) - Si ) / Si where Sv is the value of the first detector signal, and Si is the value of the second detector signal, and k is a substantially constant weighting coefficient.

11. The method of one of the claims 9 to 10, comprising the step of

• using a position signal from a third detector (24) in order to indicate the presence of a contaminant in a yarn (1), with the third detector (24) being arranged, with respect to the yarn (1), in opposition to the light sources (2, 3).

12. The method of one of the claims 9 to 11, comprising the steps of

• determining a sample value Sv of a visible light detector (4) and a sample value Si of an infrared light detector (5), and using a combination of at least two of the following quantities for deciding whether a contaminant is present and should be eliminated: o a deviation of the ratio Sv/Si of the two detector signals from a predetermined reference ratio Y; o a deviation of the visible light value Sv from a predetermined reference value Yv; o a deviation of the IR light value Si from a predetermined reference value Yi; o a ratio of the two preceding deviations.

13. The method of claim 12, comprising the steps of

• determining the ratio Sv/Si of the two spectral signals, comparing the instantaneous ratio value Sv/Si to a predetermined minimum threshold limit, and indicating the presence of a contaminant if the signal exceeds this limit for a predetermined time;

• computing the deviation of the spectral signal Sv in the visible spectrum from a predetermined reference value Yv, and indicating the presence of a first class of contaminants if this deviation exceeds a predetermined limit; • computing the deviation of the spectral signal Si in the IR spectrum from a predetermined reference value Yi, and indicating the presence of a second class of contaminants if this deviation exceeds a predetermined limit; and

• computing the ratio of the two deviations and further categorising the type of contaminant according to this ratio.