CH649066A5 - METHOD AND APPARATUS FOR DETECTING DEFECTS IN FIBER TAPES. - Google Patents

Description

The present invention relates to a method for detecting defects in fiber ribbons for the detection and classification of foreign fibers in ribbons, for example colored fibers or black hair in ribbons of combed wool.

This detection, which is particularly necessary for combed wool ribbons, has so far usually been done by slowly passing the spread ribbon in the form of a fine veil over a light tank; a human operator then observes the ribbon in order to discern the defective fibers. This method, which has nothing automatic about it, does not use other media for immersing the tape than air.

Attempts, which seem unsuccessful, have however been made to detect colored fibers by other means, such as infrared photomicrography.

However, an automatic inspection method of fiber ribbons for detecting defects, in particular colored fibers, has been described (BP No. 1211463); according to this description, the observation is made by transverse scanning of the ribbon which is held in the air.

On the other hand, an inspection method has also been described (EP No. 0013036A1), which allows the detection of faults of objects immersed in a liquid. However, these are not defects in fibrous materials, but rather surface defects in solid objects, such as metal parts. In addition, according to this description, the refractive index of the object immersion liquid must be such that it allows precise optical observation of the surface of the objects, contrary to the present invention, described in the following.

The method of detecting defects in a fiber ribbon according to the present invention comprises observing the ribbon immersed in a liquid having a refractive index practically equal to that of the fibers.

The method may include moving the ribbon in the liquid, illuminating the ribbon immersed in the liquid, a photodetector being positioned to receive light from the fiber ribbon and to detect the presence or characteristics of the defects from the signals. produced by the detector during the passage of a fault.

Defects can be distinguished from other fibers by optical means. So-called normal wool fibers are made of a relatively transparent material; but, when one observes without particular precautions a bundle of such fibers in a medium with a clearly different index of refraction, such as air, refraction and the scattering of light on the surface of the fibers make them appear clearly at l naked eye. With an elementary observation method of this type, even using an extremely small sample, only the most pronounced defects can be distinguished from other fibers.

By placing the fibers in a transparent liquid whose refractive index is equal to that of the fibers, all the reflections and refractions of the light at the liquid / fiber interface are suppressed, so that a mass of fibers in the liquid will only produce a homogeneous attenuation of the light, which depends on the material of which they are made up and not on their geometry.

A defect such as a colored fiber will obviously keep its opacity and will strongly contrast with its environment, even if it is included in a relatively thick sample.

The image of a cross section of the ribbon is preferably focused on a row of detectors by means of a lens, and the detected signals are then processed.

An apparatus for implementing the method according to the invention comprises a container for the liquid, means for transporting a fibrous ribbon within the liquid in the container, means for illuminating the ribbon immersed in the liquid, a photodetector arranged to so as to receive the light transmitted or reflected by the strip, and means for analyzing the signals produced by the detector.

Using an appropriate scanning circuit, a time dependent analog voltage can be obtained from the array of detectors. The amplitude of this signal at a given time will be proportional to the light intensity reaching the corresponding detector, thus producing a video signal.

The means for analyzing the signals, for example a mini-computer with its associated electronics, allow inter alia the recognition of faults in the ribbon, the rejection of detected events which should not be considered as faults, and preferably the taking in consideration of several faults appearing simultaneously under the detector.

The necessary data on which this analysis is performed is extracted from the video signal.

The amount of raw information contained in such a signal far exceeds the processing capabilities of a reasonable price calculation system. Fortunately, some of this information is superfluous and should be discarded as soon as possible in processing. In addition, if there are periods during which a lot of data must be processed as quickly as possible (which can be the case when several faults pass simultaneously under the detector), there are also intervals during which no useful information must not be extracted (for example when there is nothing in the sample that can be interpreted as a defect). Rhythm

5

10

15

20

25

30

35

40

45

50

55

60

65

3

649,066

to which the relevant data is produced by the detection device is therefore extremely variable. In an exemplary embodiment, a device is provided for extracting from the video signal the information strictly necessary for the subsequent processing of recognition, identification and counting of faults. This device provides useful information digitized in the form of a chain of words which are temporarily stored in an electronic memory of the FIFO type which can be filled at a high rate for short periods and emptied at a rate suitable for the subsequent processing which is performed at a slower, but more uniform rate.

Certain embodiments of the invention are now described by way of example with reference to the accompanying drawings in which:

fig. 1 is a side view of an embodiment of the apparatus according to the invention for detecting defects in the combed wool ribbons;

fig. 2 is an enlarged detail of FIG. 1;

fig. 3 is a block diagram of the detector of the apparatus of FIG. 1;

fig. 4, 5 and 6 illustrate the examination of the ribbon and the processing of the detector video signals;

fig. 7 is a block diagram of part of the signal processing circuit;

fig. 8 illustrates the analysis of a typical video signal;

fig. 9 is a block diagram of the signal processing circuit;

fig. 10 shows a typical computer bulletin, and FIG. 11 shows a modified version of the device.

In the apparatus shown in fig. 1 and 2 for the detection of colored fibers in combed wool ribbons, a drawing unit 20 is used to prepare the material to be analyzed.

The ribbon 1 is fed by the stretching system 20 comprising a double deck 22 and input rollers 21. The use of the decks reduces the adjustments necessary to adapt the analysis to different types of materials. The ribbon then passes from the stretching system 20 into the detection zone 10 where, preferably, it travels at constant speed, with only a small additional stretch. At this point, the material can typically have a linear density of 1.5 g / m (1.5 Ktex) with a width of about 13 mm.

In the detection zone 10, the ribbon passes into a container or cuvette 11 containing a liquid 12 having a refractive index practically equal to that of the fibers of the ribbon. In the case of wool, liquids such as o-dichlorobenzene, methyl salicylate or benzyl alcohol can be used. Among these, benzyl alcohol has been found to be particularly satisfactory. Of course, care must be taken so that the materials in contact with the liquid are not altered. For example, the bowl 11 can be made of stainless steel or glass plates assembled using a silicone adhesive.

In the detection zone 10, the tape passes between on the one hand a lower plate 13, made of a white reflective material, immersed in the bowl 11 with its upper surface located approximately 4 mm below the surface of the liquid, and on the other hand a glass plate 14, about 10 mm thick, which is situated horizontally above the stretched ribbon and then lowered so as to cause the immersion of the ribbon in the liquid. In its final position, the glass plate 14 rests over the white plate 13 with a spacing of 2 mm between them through which the ribbon passes. The lower surface of the glass plate is immersed, while its upper surface remains dry. The edges of the glass plate are lapped so as to facilitate the sliding of the sample under the plate.

After leaving the bowl 11, the ribbon 1 is driven by additional rollers 24 of the drawing unit, and the liquid is expressed from the ribbon between two such rollers. The spin rollers 24 must be made or covered with a suitable material, resistant to the liquid used, and positioned in such a way that the liquid expressed from the tape is collected in the bowl. A damping plate 19 keeps the surface of the liquid without stirring despite the drops falling from the spin rollers.

After passing between these rollers, the ribbon is conveyed into a container 26 using an additional roller 25. The level of the liquid relative to the glass plate 14 is not critical,

provided that the lower surface thereof remains submerged. However, an addition of liquid can be carried out from time to time, using an appropriate reservoir 17, so as to compensate for the losses of liquid carried away by the ribbon or by evaporation.

The apparatus also includes an optical detection system which consists of a light source and an observation device. The lighting device comprises an intense source of white light 30 and a condenser 31, so as to produce a uniform illumination zone of a few square centimeters on the immersed sample. A slide projector lamp (150 W) with its reflector and condenser is ideal for this use.

The angle between the incident light beam and the plane of the sample is preferably around 45 °. In addition, a cylindrical lens 32, consisting for example of a lucite rod, can be placed on the glass plate 14, so as to produce on the sample a more intense illumination zone corresponding to the region observed.

The observation device comprises a converging lens 33, for example a photographic objective provided with a diaphragm, and a row of photodetectors 34 on which is formed the image of a transverse band of the sample in the illuminated area.

In a certain embodiment of the invention, the detector consists of a row of 1024 adjacent photodiodes whose dimensions are 25 x 25 fim. By adjusting the optics to obtain an enlargement factor of 2, the band observed on the ribbon is approximately 12.8 mm x 12.5 µm. The angle of inclination of the optical axis of the observation device must be different from that of the illumination device, so as to avoid problems of specular reflection. The spacing between the glass plate 14 and the white plate 13, which determines the thickness of the sample, must be sufficiently small (typically, as mentioned, about 2 mm) to avoid problems of depth of field.

The observation is made by reflection on the white plate 13. In the absence of colored fibers, the fibrous structure of the ribbon immersed in the liquid of identical refractive index is not perceptible. The light scattered by the white plate and passing through the ribbon produces a uniform illumination on the detector. Colored fibers, foreign fibers or other types of defects strongly contrast on this background. When such a fault passes into the illuminated zone, its image is formed on the detector, locally producing zones of lower illumination, that is to say detectable events. Due to the method of ribbon preparation (stretching), the fibers are aligned in the direction of travel. Therefore, when the ribbon moves in the observation region with a known speed, the time during which an event is detected can be related to the length of the corresponding colored fiber. In addition, the width of the detected fiber is given by the number of photodiodes undergoing a change in illumination. For example, in the apparatus described above, a fiber of 50 pim in diameter will give rise to an event extending over 4 photodiodes.

In another embodiment of the invention (not shown), the detection can be carried out in transmission. In this case, the observation device will be directed directly towards the light source, the immersed sample being placed between the two.

It should be noted that the device described is capable of separating faults which pass simultaneously into the observation area.

The manner in which the signals from detector 34 are processed is now described.

Detector 34 is a monolithic row of self-scanning photodiodes and comprises (fig. 3) an alignment of silicon photodiodes DO, D1 ... D1023 each with a storage capacity

5

10

15

20

25

30

35

40

45

50

55

60

65

649,066

4

associated CO, Cl ... Cl023 and a multiplexed switch (TO transistors, TI ... T1023) controlled by a shift register SR. The shift register SR is controlled by a clock; each scan is initialized by a start pulse. The starting pulse loads a bit which is moved synchronously in the register, successively opening and closing the switches (TO, TI, etc.) and thus connecting each photodiode (DO, Dl, etc.) in turn to the video line. Each time access to a photodiode is established, the corresponding capacitance is charged to the potential of the video line and is left in an open circuit until the next scan. During the interval between successive scans, the capacity is discharged by an amount equal to the instantaneous photocurrents in the diode, integrated over the scanning time of the entire row. Each time a diode is sampled, this integrated pressure drop must be compensated across the video line. The resulting video signal is a train of pulses, each proportional to the intensity of the light reaching the corresponding photodiode.

The detector can be of the Reticon RL 1024 G type, encapsulated in a housing with 22 dual-in-line legs, which contains 1024 photodiodes 25 µm wide and 26 µm wide, protected by a transparent window.

When the ribbon 1 moves in the Y direction (fig. 4) under the glass plate, successive bands of the ribbon are examined by the detector, as illustrated in fig. 5. The scanning direction (across the sample) is called the X direction.

Fig. 5 also shows the shape of the resulting video signal and the corresponding coordinates X (diodes No. 0, 1, 2 ... 1023) and Y (lines No. n, n + 1, n + 2 ... 1023). The video signal from the detector 34 is transmitted to the front unit 100 which amplifies the signal and also provides the start and clock pulses for the shift register SR. The pulses are so related to the drawing speed that a scan takes place every Vs mm on the material.

Suppose that among the fibers of the sample a defect (dark fiber 35) crosses the illuminated area 36 whose image is formed on the row of photodiodes 34 by means of optics 33. During a scan, the voltage VI ( t) at the output of the front unit will successively take values proportional to the light intensity striking each photodiode in turn, from 0 to 1023. The video signal VI (t) is shown diagrammatically in FIG. 4 where it will be noted that the general envelope of the signal is curved, given that the average light intensity striking the detector is greater in the center than near the edges, this being due in particular to the lack of homogeneity of illumination of the ribbon: the distance from the light source, and therefore the illumination, is not the same for each point (Lc <Lb in fig. 6).

A similar effect exists for light reflected back to the detector. The presence of defect 35 produces a local decrease in light intensity which results in a dip in the shape of the voltage Vl (t) observed.

A correction circuit 101 produces, by sampling and averaging the video signals corresponding to successive scans, a reference voltage which is then subtracted from the instantaneous video signal Vl (t) to provide a normalized signal V2 (t) having a baseline horizontal. As illustrated in fig. 7, the correction device comprises a circuit 102 in which the curvature of the reference signal is extracted from the video signal itself which is sampled. born and maintained in a set of abilities during the scan. A special circuit generates an adjustment signal joining the memorized levels to each other. The video signal is subtracted from the reference signal in a differential amplifier 103 to obtain the standardized video signal on which it is easy to use fixed voltages as threshold levels of the analog / digital conversion. The reference signal can also be used to automatically regulate the intensity of the light source.

The standardized video signal V2 (t) then passes through an analog / digital converter or threshold detector 104 in which the instantaneous value of the signal is compared with 5 reference values or thresholds

(TH1 to TH5 in fig. 8). Levels TH1 and TH5 are adjustable using potentiometers 105 and 106, while levels TH2, TH3 and TH4 are also distributed between levels 1 and 5. When the signal is below the first threshold level, no information is not delivered by the circuits. A detected event begins when the signal exceeds the first threshold (for X = 13 in the example in fig. 8) and ends when the signal drops below this level (for X = 24). There can be more than one such event in a single scan. The detector 104 produces, for each X coordinate, a value from 0 to 5 depending on the level of the signal.

After the conversion, the detection logic 108 stores the information in a FIFO (first in, first out) memory 115 (described below) each time a threshold level is crossed. This digital information, called information X, consists of a 16-bit word containing the level reached and the number of the element of the detector (photodiodes) for which this change in amplitude has taken place. This number is equivalent to the transverse coordinate of the event with respect to the sample (for example, in fig. 8, level 4 for X = 14). This information is distributed as follows:

- 10 bits for the position of the photodiode for which the threshold level has been crossed (from 0 to 1023),

- 3 bits for the threshold level which has been crossed,

- 1 bit to identify X information (the remaining bits are unused).

In this way, information is only stored in the FIFO memory when a fault is actually observed by the detector. The X coordinate is obtained from a counter 111, which is part of the control logic 109, which counts from 0 to 1023 to represent the position of the photodiodes in the video signal. The counter 111 is reset by the start pulse and incremented by one at each clock pulse.

Similarly, a counter Y 110 is incremented by one at each start pulse so as to count the number of scans (fig. 5 and 7). The control logic 109 also includes units for end-of-line information 112 and for control 113 of the transmission of information via the data bus 114 to the FIFO memory.

During the scanning of a line, an event begins when the normalized video signal crosses threshold level 1 and ends when the normalized video signal drops below threshold level 1 (fig. 8). Several events, as defined above, can be simultaneously present on the same scanned line. It is up to the computer program to possibly determine whether these events are part of the same natural defect or actually reflect the presence of several natural faults.

During the duration of an event, a digital integrator 108 calculates the area located under the envelope of the signal corresponding to the event and above the first threshold level (FIG. 8). The number representing this area can be evaluated as being the number of clock pulses for which the signal is above level 1, plus the number of clock pulses for which the signal is above level level 2, and so on up to level 5. For example, for the signal shown in fig. 8, the total is given by:

Level:

1

2

3

11

8 8

4

7

5

4 37

This information is transmitted to the FIFO memory at the end of the detected event with a special marker, forming a word called SIGMA. The information X and X for the signal of fig. 8 are therefore:

5

10

15

20

25

30

35

40

45

50

55

60

65

5

649,066

Threshold

Address

Information X

2

13

start of fault

4

14

5

15

4

19

1

21

0

24

end of fault

landmark

37

opacity of the defect

At the end of each scan during which at least one event has been recorded, an information word Y is generated, which consists of a digital word giving the length of the sample examined so far (more precisely, the number of scans performed since the start of the analysis). This is equivalent to the longitudinal coordinate of the event relative to the sample. Information Y is also a 16-bit word, containing the following information:

- 12 bits for the number of the line scanned (counter Y),

- 3 bits for end of analysis (END), Y counter overflow (OVERFLOW) and interval indication (GAP) marks,

- 1 bit to identify information Y.

Information Y can also be sent alone (without any information X during the corresponding line), simply in order to transmit a reference.

- The line number is transmitted only when an event has appeared in the line.

- The end mark is transmitted after receipt of an appropriate command from the microprocessor to stop the measurement, thus specifying that this information is the last to be transmitted. The end command can be decided by the operator (keyboard command) or automatically by the microprocessor (FIFO FULL).

- The overshoot mark is transmitted when the Y counter reaches its maximum value, in order to inform the rest of the system that the subsequent Y information will be reset. This fact should be taken into consideration for the various logic counters in the microprocessor and the computer.

- The interval marker is transmitted when no event has been encountered during N scan lines; this information is necessary for the fault recognition program to analyze the data.

The information X, Y and £ for all the events detected are the only information required for the subsequent calculation of recognition and classification of faults. The speed at which this information follows each other is much slower than the speed at which it would be necessary to process the raw video signal.

Given the difference between the instantaneous speed at which the detector provides information and the speed at which the computer can process it, all 16-bit information words are loaded into RAM used as a microprocessor-controlled FIFO. FIFO memory 115 is a temporary storage device for digital data: the information written first in memory will be the first to be extracted from it. The speeds at which data is entered into FIFO, on the one hand, and extracted from it, on the other hand, can be entirely independent, provided that the average speed of entry does not exceed the average speed of exit, in order to avoid any overflow. The event detection logic, located in block 107, has the task of writing data, X, Y and 2 to the FIFO memory in a timely manner.

The FIFO is emptied by a control device at the request of a calculation system which uses the information provided, in order to recognize a defect from successive events, to classify it according to its size and its opacity, and to establish a statistic from accumulated defects. The FIFO memory comprises a RAM vector of 2 K-words of 16 bits and the microprocessor is used to control the FIFO. The data is loaded into the FIFO at the rate of the detector using the DMA principle, using a write counter 116 as an address pointer. The write counter is incremented for each word loaded in the FIFO. The data is emptied from the FIFO by the microprocessor with normal addressing, using a software reading counter which is incremented with each word read in the FIFO, via the data and address buses 126, 127 of the microprocessor and a multiplexer of address 117. The saturation condition of the FIFO occurs when the detector attempts to load the FIFO at a location not yet emptied by the microprocessor. This situation is detected by an appropriate circuit which commands the microprocessor to stop the measurement. The FIFO data is emptied by the microprocessor at the request of the computer to which it is transmitted in the form of 2 bytes of 8 bits using a serial line of RS 232 type (transmission speed: 9600 baud) .

The control device is a general-purpose microcomputer, containing a microprocessor, a keyboard and a display, the necessary connections with FIFO, on the one hand, and with a computer, on the other hand, and the programs necessary for perform the following tasks:

- ensure logical control of the entire measurement system;

- inform the operator of the state of the system;

- allow the operator to start, interrupt, or end the measurement;

- produce all synchronization or reference signals for the detector, counters, etc .;

- read the data from the FIFO memory and transmit it to the computer, and

- intervene in the event of abnormal conditions such as saturation of the FIFO.

The calculation system is a mini-computer associated with a device for printing the results and programmed to perform the following tasks:

- link the events produced during the scanning of a fault, in accordance with certain rules;

- reduce each recorded fault to a set of characteristic parameters;

- according to classification rules applied to these parameters, decide, for each object detected on the sample, either to reject it as false information, or to keep it in memory as a real defect, and

- perform, at the request of the control system, all the above-mentioned tasks and print the statistical data concerning all recognized faults, including for example size, shape and opacity information.

The general layout of the interface logic is shown in fig. 9. As noted previously, the FIFO is connected via the multiplexer 117 and the buffers 118 to the buses 126, 127 of a microprocessor system which comprises a microprocessor 119, as well as a program memory 120 and a working memory 121.

The microprocessor

- controls the transfer of data between FIFO and the computer;

- displays messages intended to inform the operator of the progress of the measurement, and

- produces the signals necessary to control the interface system (for example start and clock pulses for the front unit).

The display and the keyboard available on the microprocessor are used by the operator before the start of the measurement in order to specify the material to be analyzed and some other household information. This information is also transmitted to the computer and appears in a printed bulletin at the end of the analysis.

This information is:

- batch definition (LOT),

- sample number (SAMPLE),

- customer name (CUSTOMER),

- name of the operator (OPERATOR),

- type of material (TYPE),

5

10

15

20

25

30

35

40

45

50

55

60

65

649,066

6

- stretching (DRAFT),

• - title of the ribbon to be analyzed (TEX).

During the analysis, the display is used to announce messages such as:

WAIT OPERATOR: the operator is invited to start the analysis as soon as it is ready,

WAIT MOTOR: the operator is invited to start the stretching machine,

READY: the operator is informed that the interface system is ready to accept an order, ACTIVE: this message appears while the measurement is actually in progress, accompanied by an indication in meters of the length of the tape analyzed until there, END: the operator is informed that the measurement is finished; he can ask for a new one,

ENDING: the operator is informed that the measurement is about to end, but must wait until the FIFO is completely empty,

ERROR: the operator is informed that an abnormal situation has appeared, such as the saturation of the FIFO or an error in transmission between the computer and the microprocessor.

The computer 122 communicates with the microprocessor via an interface 123 and a serial link 124; it is used to calculate a set of parameters for each defect, which are chosen so as to characterize it completely.

The computer software program does not need to keep in memory all of the data which represents a fault as it is sent by the interface system. Rather, the program retains a set of values characterizing at any time the part of the defect analyzed so far. On each scan line, this set of values is updated using the information sent by the microprocessor.

When the end of the fault is recognized (on receipt of an interval indication, GAP), a final set of parameters is calculated which completely characterizes the fault in size, shape, irregularity, etc., taking into account all the data accumulated regarding this fault.

These parameters are compared with certain reference parameters, in order to determine the class in which the fault is to be recognized.

When the end mark (END) is detected, the computer prints the results of the fault classification calculation in the form of a bulletin (printer 125).

The classification requires the calculation of the following parameters:

- cut,

- shape,

- shades of grey,

- projected length,

- discontinuities.

io fig. 10 is a representation of the fault as it is possible to reconstruct it from information X and Y received by the computer. The printed character is the threshold level crossed. The area corresponding to one character on the paper represents a square of Va mm on the side of the ribbon.

In a typical embodiment of the apparatus, the detector is scanned at a speed such that each photodiode "tracks" contiguous squares of Vbo mm side on a strip 12.8 mm wide. The speed of the sample moving under the detector can exceed 1.5 m / min.

At the option of the operator, the sample can be moved under the detector without an analysis being carried out, so as to allow the drawing machine to achieve a homogeneous and stationary production before start measurements.

Means are provided for automatically transmitting the measurement results (statistics of defects per unit length or weight of the sample) to another computer for archiving and later use.

The detection system described above can also be combined with the dual reflection / transmission defect detector, described in our UK patent filing No. 8110057, as shown in fig. 11 where such a double detector device 60 is placed on the same drawing machine as the immersion detection unit 10. The signals from the three detectors can be processed jointly by the same units: interface (100 to 115), control (119 to 121) and computer accompanied by a printer (122, 125). In this combination, it can be provided that the immersion detection system observes the presence of defective fibers such as black hairs in a combed woolen ribbon, while the dual detector observes other types of defects (including for example foreign matter such as thistles or shards of 40 seeds in combed wool ribbons); in this combined arrangement, the dual detector has a lower resolution than the immersion detector (for example Ve mm compared to Yho mm).

In other cases, the dual detection system can be designed to observe the sample immersed in the liquid.

10 sheets of drawings

Claims (10)

649,066 2 1. A method of detecting defects in a fiber ribbon, comprising observing the ribbon immersed in a liquid having a refractive index practically equal to that of the fibers. 2. Method according to claim 1, comprising moving the ribbon in the liquid, illuminating the ribbon immersed in the liquid, a photodetector being positioned so as to receive the light from the fiber ribbon and to detect the presence or the characteristics of the faults from the signals produced by the detector during the passage of a fault. 3. Method according to one of claims 1 or 2, in which the tape is continuously driven in the liquid. 4. The method of claim 3, in which an image of a portion of the ribbon, transverse to the direction of movement, is focused on the photodetector. 5. Apparatus for implementing the method according to claim 1, comprising a container for the liquid, means for transporting a fibrous ribbon within the liquid in the container, means for illuminating the ribbon immersed in the liquid, a photodetector arranged to receive the light transmitted or reflected by the ribbon, and means for analyzing the signals produced by the detector. 6. Apparatus according to claim 5, wherein the illumination means comprise a cylindrical lens arranged to illuminate a strip of the ribbon, transverse to the direction of movement. 7. Apparatus according to one of claims 5 or 6, comprising focusing means arranged so as to form on the photodetector an image of a strip of the ribbon, transverse to the direction of movement. 8. Apparatus according to claim 7, wherein the photodetector consists of a row of adjacent detectors producing a serial video signal. 9. Apparatus according to one of claims 5 to 8, comprising wringing means for expressing the liquid from the tape after its immersion in the liquid. 10. Apparatus according to one of claims 5 to 9, comprising two additional detectors spaced along the path of the fibrous ribbon and means for illuminating the ribbon arranged so that the light transmitted or reflected by the ribbon falls on one, respectively on the other, of these two detectors.