RU2483276C1 - Method for detection of sheet irregularities and device for its realisation - Google Patents

Abstract

FIELD: construction.

SUBSTANCE: sheet is moved in a tract through a device with transmitting electrodes on one side of the tract and a receiving electrode on the opposite side oppositely to transmitting electrodes. Transmitting electrodes are divided into groups, besides, one and the same transmitting electrode may belong to more than one group. Each group is divided into two subgroups. An exciting signal is sent to the first subgroup of electrodes, and the signal that is opposite to the exciting signal of the first subgroup is sent to the second subgroup. Correlation is determined between the exciting signal and the signal from the receiving electrode. Correlation is analysed, and on this basis availability of sheet irregularity in the specified area is specified.

EFFECT: improved validity of irregularity detection and reduced complexity of design.

24 cl, 6 dwg

Description

The group of inventions relates to methods for detecting inhomogeneities of sheet material, as well as to devices implementing this method, and can be used in the production of sheet materials, and, in particular, in document processing devices, such as banknotes.

Detection of inhomogeneities of the sheet material is carried out, as a rule, when it passes through the sheet path. The heterogeneity of the sheet can be expressed by the difference in its thickness. This difference in thickness may be inherent in the sheet, as in the case of watermarks on paper, or it may reflect damage and manufacturing defects of the sheet. In addition to differences in sheet thickness, inhomogeneities can occur due to the presence of foreign material in the thickness of the sheet or on its surface.

When processing documents, a device for detecting heterogeneities of sheet material allows you to control the characteristics of the document itself, affecting its identification, conclusions about the authenticity and suitability for use.

Modern banknotes are complex structures with variable thickness, often consisting of many layers. Banknote paper has watermarks, performed as local thickening or thinning of the paper sheet. Polymeric protective tapes are embedded in the thickness of the paper sheet, which leads not only to an increase in the thickness of the banknote at the location of the tape, but also to a change in other physical characteristics. The most noticeable difference in physical characteristics is the change in dielectric constant. Special colorful layers are applied to the banknotes, which can have a sufficiently large thickness and noticeable differences in physical characteristics compared to the original paper sheet. Film applications, such as holograms or polarization films, are also applied to the surface of banknotes, which also increases the local thickness and dielectric constant of the banknote. Thus, monitoring the distribution of heterogeneities on the surface of the banknote provides important information for its identification and authentication. On the other hand, the distribution of heterogeneities over the surface of the banknote makes it possible to judge its suitability for further use. Control of heterogeneity allows you to determine the degree of wear of the banknote, expressed in thinning along the fold line. In addition, there are banknotes in circulation that were torn and then glued with adhesive tape. Such banknotes must be withdrawn from circulation. The layer of adhesive tape is defined by the device as an unacceptable heterogeneity of the sheet, as a result of which the banknote is sent for recycling.

State of the art

The applied methods and devices for detecting sheet inhomogeneities are divided into non-contact and contact. Contact devices make a direct measurement of the thickness of a banknote by clamping it between two surfaces, for example, the surfaces of two rollers located in a sheet duct. Contactless devices are based on the measurement of a certain physical indicator of the sheet material, which makes it possible to detect inhomogeneities of the sheet, for example, optical density or attenuation coefficient of an ultrasonic pulse. Used densitometric, capacitive or ultrasonic type meters.

The claimed group of inventions relates to a capacitive type. In capacitive-type devices to control the heterogeneity of the sheet, in the zone of its location, one or more measuring electric capacitors is used. Each capacitor is positioned so that part of the sheet is in the dielectric gap of the capacitor. The inhomogeneities of the sheet lead to a difference in the electrical capacitance of the measuring capacitor. The change in capacity is recorded as a measure of sheet heterogeneity.

Various principles of the mutual arrangement of the sheet and the electrodes of the capacitor are known. In the application EP 0097570 (publ. 01.04.1984, IPC G01B 7/06), one capacitor electrode is located on one side of the sheet, and the second electrode on the other side of the sheet. Thus, the field lines in the capacitor mainly intersect the sheet in the perpendicular direction. The advantages of such an arrangement include the small dependence of the capacitance on the location of the sheet in the gap between the electrodes, and, by itself, the gap can be made quite wide - within a few times of the sheet thickness. Therefore, such a device can be used when the sheet moves in the gap at high speed, and its constant contact with the electrodes is undesirable for mechanical reasons. The disadvantage is the high sensitivity to instability of the gap between the electrodes. This requires special mechanical measures to stabilize the gap and suppress vibration of the electrodes.

In the application SE 355428 (publ. 04/16/1973, IPC G07D 7/00) a measuring capacitor is used with electrodes located on opposite sides of the document being monitored. One of the electrodes has a special shape corresponding to the shape of the watermark on a particular type of security document. A change in the capacitance of the capacitor manifests itself most strongly with the exact combination of the watermark on this document with the electrode. The change in capacitance is proportional to the spatial correlation between the shape of the electrodes, on the one hand, and the shape of the sheet inhomogeneity, on the other. The confirmation of the authenticity of the document is based on this phenomenon. The disadvantage of this device is its suitability for detecting inhomogeneities of one particular geometric shape, as well as high sensitivity to instability of the gap between the electrodes.

US Pat. No. 5,122,754 (published June 16, 1992, IPC G01R 27/26) describes the use of a pair of measuring capacitors. In each capacitor, one of the electrodes is located on one side of the sheet, and the second on the other side of the sheet. Both capacitors are located next to each other and also have a special shape of electrodes corresponding to the shape of the watermark on a certain type of sheets of security paper. One of the capacitors in the shape of the electrode corresponds to the portion of the sheet with thickening paper, and the second capacitor corresponds to the portion of the sheet with thinning paper. When a watermark hits exactly the location of a pair of capacitors, the capacitance of one of the capacitors drops significantly, while the other increases. The difference in the change in their capacitance is proportional to the spatial correlation between the shape of the electrodes, on the one hand, and the shape of the sheet heterogeneity, on the other. It is considered as a criterion for confirming the authenticity of the watermark on the sheet. Like the previous one, this device is not universal, since it is designed to determine the presence of watermarks of a particular type.

Another solution is described in US patent 5,899,313 (publ. 05/04/1999, IPC G07D 7/00). According to this patent, the capacitor electrodes are located close to each other, on the same side of the sheet. The field lines in the capacitor pass through the sheet material along an arc. The main disadvantage of this solution is the extremely high dependence of the capacitance on the gap between the electrodes and the sheet. Therefore, as described in the known patent, it is optimal to press the sheet directly to the surface of the electrodes. This limits its use to devices in which the sheet is stationary or moves very slowly. The advantages of the described arrangement of electrodes include the compactness of the device, since all electrodes, even for many capacitors, can be placed on a common dielectric plate. On the reverse side of the same plate, by printed wiring, electronic circuits can be made to measure the electrical capacitance.

There is also a solution combining the positive aspects of both previously described known solutions. In the patent US 6,229,317 (publ. 08.05.2001, IPC G01R 27/28) three electrodes are used, two of which are located next to each other on one side of the sheet, and the third is on its other side directly opposite the area of the first and second electrodes. In this case, only the first and second electrodes are included in the measuring circuit, while the third electrode is completely insulated. The function of the third, passive electrode is to direct the lines of force perpendicular to the sheet. The measuring capacitor, in fact, consists of two series-connected capacitors, and the sheet is in the dielectric gap of each of them. In this solution, compactness is achieved due to the possibility of arranging almost all elements of the device, except the passive electrode, on one side of the sheet. But this solution also has a high sensitivity to the instability of the gap between the first and second electrodes, on the one hand, and the passive electrode, on the other.

.For all devices in which the electrodes are placed on opposite sides of the sheet, the general dependence of the capacitance C on the thickness of the sheet h and the gap width L is characteristic:

.

.A similar formula can be easily obtained from an equivalent series circuit of two flat capacitors and is known, for example, from SE 55428. Here, ε ₀ denotes the dielectric constant of vacuum (and is almost equal to the permeability of air, ε denotes the relative dielectric constant of the sheet material, and S is the area The change in capacitance of a capacitor reflects both a change in sheet thickness and a change in its dielectric constant due to the introduction of foreign material or its imposition on Since the change in capacity cannot determine which of these factors caused this change, it is convenient to talk about a change in the equivalent sheet thickness h _e , assuming the dielectric constant ε _{p to} be unchanged. In this interpretation, formula (1) is transformed to:

.

A change in the equivalent sheet thickness reflects both the actual change in thickness and the appearance of foreign material on the surface or in the thickness of the sheet.

If the gap between the electrodes of the capacitor in which the sheet passes exceeds the thickness of the sheet by several times, then the inhomogeneities of the sheet lead to a very small change in the capacitance of the capacitor. For example, if the thickness of the paper sheet is 0.1 mm and the gap is 0.5 mm, then reducing the sheet thickness by 2 times will change the capacity by only about 5%. Such a small amount of change makes high demands on the capacitance measurement scheme. On the contrary, the relative change in the width of the gap practically coincides in absolute value with the relative change in capacitance caused by it and is opposite in sign. So, in the above example, a 0.1 mm gap change causes a 20% change in capacitance. This shows that the stability of the gap should be maintained at the level of 10-20 microns, which is a serious design problem, especially if the path should be periodically opened for maintenance.

The performance of a capacitive sensor is limited by the duration of the capacitance measurement process. The measuring capacitor itself changes its capacitance almost instantly and does not make an additional contribution to the duration of the measurement process.

Different methods are used to measure the electrical capacitance of the measuring capacitor. They differ in the physical principle and the duration of the measurement process, which, in turn, determines the total time spent measuring the inhomogeneities of the entire sheet. Historically, the first were solutions in which a measuring capacitor was included in the oscillatory circuit of a self-oscillating circuit, such as described in US 3,764,899 (published on 10/09/1973, IPC G01R 7/26). A change in the capacitance of the capacitor leads to a change in the frequency of self-generation, which is recorded by the frequency meter. The relative change in frequency, for small quantities, is half the relative change in capacitance. The disadvantage of such schemes is the low speed, since reliable measurement of small frequency changes requires a lot of time. In later solutions, to which, in particular, US Pat. No. 5,122,754 relates, the measuring capacitor is included in the single-shot circuit and sets the response time. The measured parameter in this case is the pulse duration, which allows you to quickly get the result.

A device for checking the presence of a polymer metallized protective strip in a sheet of banknote paper is described in US Pat. No. 5,394,969. This device employs two transmitting electrodes in the form of flat plates placed on the same side of the banknote, a receiving electrode in the form of a wire placed on the same side of the banknote between the flat plates, and a protective electrode in the form of a half cylinder surrounding the receiving electrode. The method of checking sheets implemented by this device allows you to detect a metallized protective tape and provides the following sequence of actions. The sheet is moved along the transmitting electrodes, while the phase-shifting electrical signals are applied to the transmitting electrodes, and the resulting signal is removed from the receiving electrode. In the absence of a protective strip near the receiving electrode, there is no signal at the receiving electrode, since capacitive currents flowing into the receiving electrode from the first and second transmitting electrodes are mutually subtracted. When the protective strip is located directly near the receiving electrode, but is shifted towards one of the transmitting electrodes, an unbalance signal appears on the receiving electrode. This signal is further amplified and detected by a synchronous detector. The deviation of the voltage at the detector output from the neutral level indicates the presence of a protective strip near the receiving electrode. The protective electrode has a neutral potential and limits the penetration of an electric field onto the receiving electrode through banknote paper in the absence of a protective strip. The metallization on the protective strip acts as a bridge over the protective electrode and facilitates the penetration of the electric field to the receiving electrode. Thus, this method allows you to determine the presence of only a metallized protective tape. Moreover, sheet inhomogeneities having a different origin and located over the entire surface of the sheet cannot be detected. This verification method is selected as the closest analogue for the claimed method for detecting heterogeneity of sheets.

Progress in the field of high-frequency electronics has recently made it possible to switch to direct measurement of the current flowing through a measuring capacitor and proportional to its capacitance. The high-frequency voltage from the generator, with a frequency of tens or hundreds of megahertz, is applied to the first electrode of the measuring capacitor, called the transmitting electrode. The measurement object is the return current flowing through the second capacitor electrode back to the generator. This electrode is called the receiving electrode. To measure current, a circuit consisting of an amplifier and a detector is used; the detector output is connected to an analog-to-digital converter. This method gives the highest speed and relatively low sensitivity to electrical noise. It is used in the device described in US patent 6,229,317, as well as in US patent 8,028,990 (publ. 04.10.2011, IPC B65H 7/02), selected as a prototype for the claimed device.

The positive side of the prototype device is the use of many measuring capacitors located along a straight line perpendicular to the direction of travel of the sheet. Due to this, the heterogeneity of the sheet along the whole strip intersecting the sheet is simultaneously controlled. This allows you to fully examine the entire sheet at a high speed of its movement in the tract. Since the prototype uses measuring capacitors, the electrodes of which are located on opposite sides of the sheet, it is characterized by a high sensitivity to the instability of the gap between the electrodes. In the prototype of the device, a common generator and a common transmitting electrode are used for all measuring capacitors in the strip. For each capacitor, a separate receiving electrode and separate current measurement circuits are used. Due to the previously mentioned small relative value, changes in the capacitance of the current measurement circuit are quite complex. Their repetition in the number of used measuring capacitors leads to a significant increase in the complexity of the entire device.

The task in the development of the claimed group of inventions was the development of an effective method for detecting heterogeneity of sheets and the creation of a simple and fast device for this, capable of reliable operation with great instability of the gap between its parts.

The technical result of the group of inventions is to increase the reliability of detection of heterogeneity of sheets and reduce the complexity of the design.

Disclosure of invention

The claimed technical result is achieved by the fact that in accordance with the method for detecting sheet inhomogeneity, the sheet is moved in the path through a device containing transmitting electrodes on one side of the path, and a receiving electrode located on the opposite side of the path opposite the location of the transmitting electrodes divides the transmitting electrodes into groups, moreover, the same transmitting electrode may belong to more than one group, each group is divided into two subgroups, each of which contains at least at least one electrode, an analysis is made of the presence of sheet inhomogeneities in the location of the transmitting electrodes of each group, for which an excitation signal is supplied to the first subgroup of electrodes in the group, and an antiphase excitation signal to the first subgroup is supplied to the second subgroup of electrodes, and the correlation is determined between the exciting signal and the signal from the receiving electrode, analyze the correlation and it is judged on the presence of heterogeneity of the sheet in the specified area, and based on the judgments obtained for each g uppy, judged on the presence of heterogeneity in the sheet electrode arrangement.

Unlike the prototype, in the claimed method, a sheet is passed between several groups of electrodes, each of which is simultaneously supplied with an exciting signal and antiphase to it, while each electrode can belong to more than one group.

In the inventive method, when analyzing the presence of sheet inhomogeneities in the area of the location of the transmitting electrodes of each group, an additional measurement can be carried out, in which the signals applied to the two subgroups of the group electrodes are interchanged and the resulting correlation between the exciting signal and the signal from the receiving electrode is subtracted from the original the resulting correlation.

In the inventive method, at least two groups of electrodes can be simultaneously excited, and the excitation signal for each of such groups is selected so that it does not correlate with excitation signals used for other simultaneously excited groups.

In the inventive method, as the sheet advances in the path, it is possible to additionally at least once repeat the detection of sheet inhomogeneities.

In the claimed method, subgroups of transmitting electrodes included in one group can be located in close proximity to each other, and the electrodes in each of these subgroups are also located in close proximity to each other, and the threshold for the absolute value of correlation is set in order to detect heterogeneity, which is calculated for each group, and when the threshold is exceeded, a decision is made on the presence of sheet inhomogeneity in the area of the group's transmitting electrodes.

In the claimed method, based on the obtained correlation values, the approximate distribution of the equivalent sheet thickness over its surface can be restored, and then the named distribution can be analyzed to detect heterogeneity, while the groups of electrodes can be spatially arranged so that the correlation measurement result for each group corresponds to a specific component decomposition of the distribution of equivalent thickness over the surface of the sheet on a specific functional basis.

As a functional basis, the Haar wavelet system or the Walsh-Hadamard function system can be chosen. For a variant of the method in which groups of electrodes can be spatially arranged so that the correlation measurement result for each group corresponds to a certain component of the expansion of the distribution of equivalent thickness over the sheet surface over a specific functional basis, subgroups of transmitting electrodes included in one group can be located in immediate proximity to each other, and the electrodes in each of these subgroups are also located in close proximity to each other, and The values of correlation values calculated for each group, considered as a value proportional to the derivative equivalent plate thickness in the direction between the centers of gravity of the first and second subgroups of transmitting electrodes.

The claimed technical result is achieved in that the device for detecting inhomogeneity of the sheet comprises: transmitting electrodes placed on one side of the sheet, divided into groups, each of which consists of two subgroups, each of the subgroups contains at least one electrode, while at least one transmitting electrode may belong to more than one group; a receiving electrode placed on the opposite side of the sheet and covering the area of location of the transmitting electrodes; the transmitting electrode control unit and a signal processing unit from the receiving electrode synchronized with it, while the transmitting electrode control unit is configured to excite the electrodes of each group by supplying an excitation signal to the first subgroup of electrodes and simultaneously supplying an antiphase excitation signal to the second subgroup of electrodes, and the signal processing unit from the receiving electrode comprises a device for determining the correlation value of the exciting signal.

Unlike the prototype, the claimed device uses groups of transmitting electrodes and one receiving electrode, while the groups of transmitting electrodes are further divided into subgroups, and each electrode can belong to more than one group.

In the claimed device, the transmitting electrodes can be made in the form of non-overlapping surface areas, the receiving electrode can be made in the form of a surface area and equidistant from the transmitting electrodes, while the control unit of the transmitting electrodes can be configured to provide a ratio of the amplitudes of the exciting and antiphase signals supplied to it subgroups of electrodes proportional to the ratio of the total electrode areas of the second and first subgroups.

In the claimed device, the signal processing unit from the receiving electrode may include a signal amplifier from the receiving electrode, while for such a device, the signal processing unit may be equipped with at least one additional device for determining the correlation of the exciting signal and the signal from the receiving electrode so that the signal the excitation of each of the simultaneously excited groups is fed to the input of a separate device to determine the correlation

In the claimed device, including for the variant with transmitting electrodes in the form of non-overlapping surface sections, the device for determining the correlation value can be a synchronous detector.

The use of two subgroups of electrodes that are excited by antiphase voltages can significantly reduce the harmful effect of the gap on the detection of sheet inhomogeneities. In fact, the electrodes of the first subgroup and the second subgroup of transmitting electrodes form, together with the receiving electrode, two independent measuring capacitors. The current flowing through the common receiving electrode contains the sum of the current from the first subgroup and the current from the second subgroup. The magnitude of each of these currents is almost inversely proportional to the magnitude of the gap. Due to the supply of antiphase stresses to these subgroups, in the absence of sheet heterogeneity, there is a mutual weakening of these currents, which weakens the harmful effect of gap instability. At the same time, the appearance of sheet inhomogeneity in the region covered by the electrodes of the group leads to an increase in current from the subset of electrodes that covers the area with increased equivalent thickness, and to a decrease in current from the subgroup that covers the area with reduced equivalent thickness. Consequently, when adding antiphase currents, an increase in the recorded change in the current of the receiving electrode occurs, in comparison with changes in currents from each of the subgroups separately. Thus, the first current component of the receiving electrode, which flows even in the absence of sheet inhomogeneities and depends on the instability of the gap, decreases, and the second component characterizing the sheet inhomogeneity increases. This allows you to use a simpler mechanical design of the device, allowing increased instability of the gap, without loss of accuracy in detecting inhomogeneities of the sheet.

In a preferred embodiment, the electrodes of the first and second subgroups are made in the form of non-overlapping sections of the surface, and the receiving electrode is made in the form of a section of the surface equidistant from the transmitting electrodes. In this case, the ratio of the amplitudes of the exciting signal and the signal out of phase with it, applied to the subgroups of the electrodes of the group, is selected proportional to the ratio of the total areas of the electrodes of the second subgroup and the first subgroup. This configuration of the electrodes ensures the equality of currents flowing into the receiving electrode from the first subgroup of electrodes and from the second subgroup of electrodes, provided that there is no inhomogeneity of the sheet. Since these currents are in antiphase, the malicious first component of the current of the receiving electrode described above is reduced to zero. The second, useful component characterizing the presence of heterogeneity, nevertheless, remains dependent on the size of the gap L. However, the negative effect of this dependence is manifested to a significant degree only with large relative changes in the gap and distorts only the measured value of the difference in the effective sheet thickness. The very fact of the presence of a defined difference and its sign are not distorted.

. Если же в одном из измерительных конденсаторов появляется дополнительная неоднородность эффективной толщины Δh_e, то его емкость становится равной

При приложении напряжения U частотой f к каждому из конденсаторов ток будет пропорционален емкости: I=2πfCU. С учетом противофазной подачи напряжения на электроды подгрупп результирующий ток приемного электрода равен

Let us explain the above reasoning by analyzing the current through the electrodes. We assume that h is much less than L. The capacitance between each of the transmitting electrodes and the receiving electrode, in the absence of sheet inhomogeneities, is

. If, in one of the measuring capacitors, an additional heterogeneity of the effective thickness Δh _e appears, then its capacity becomes equal

When voltage U is applied with frequency f to each of the capacitors, the current will be proportional to the capacitance: I = 2πfCU. Given the antiphase voltage supply to the electrodes of the subgroups, the resulting current of the receiving electrode is

Thus, the current of the receiving electrode is proportional to Δh _e and, in the absence of heterogeneity, is equal to zero. The proportionality coefficient is determined by the voltage, frequency, the gap between the receiving and transmitting electrodes and the dielectric constant of the sheet material.

It is important that the mutual subtraction of currents occurs directly in the receiving electrode, that is, to find the difference in capacitance of the measuring capacitors, additional electronic circuits are not required. This distinguishes the claimed invention from circuitry for measuring the difference in capacitance, known from the prior art, where the capacitance is first measured individually by electronic circuits and only then subtracted from one another. Such a scheme is implemented, for example, in US 5,122,754. The claimed invention allows to avoid distortions introduced by electronic circuits when working with a large level of current passing through a separate measuring capacitor, against which small changes due to inhomogeneities are required to be detected.

The inclusion of the same electrode in more than one group can improve the reliability of detecting inhomogeneities. This is because the effective sheet thickness in the region where such an electrode is located is taken into account in the data of more than one measurement, thereby reducing the influence of the error of each individual measurement on the overall result.

The small value of the current of the receiving electrode and its high frequency do not significantly simplify the processing circuit of the signal from the receiving electrode. The simplification of the device is achieved through the use of only one such circuit in the signal processing unit. Compared with the prototype, in which the number of signal processing circuits is equal to the number of measuring capacitors, this significantly reduces the total number of electronic elements.

The claimed device uses many transmitting electrodes, while providing a signal to them can be implemented with simple equipment. In fact, it is required from the control unit of the transmitting electrodes to provide switching of the signal from the generator of the exciting signal in order to transmit it to one or another electrode. This function can be performed either by a multichannel analog switch, or by a programmable logic matrix (classes CPLD or FPGA). Both of these types of electronic components are mass-produced and widely available as a single chip. The complexity of such a control unit of the transmitting electrodes is incomparably less than the complexity of the signal processing unit, consisting of many separate signal processing circuits from the receiving electrodes used in the prototype. Thus, the transition from the circuit "one transmitting and many receiving electrodes" to the scheme "many transmitting electrodes and one receiving" greatly simplifies the device.

In a preferred embodiment, an amplifier is used to process the signal processing unit from the receiving electrode. The magnitude of the current of the receiving electrode is very small, and it decreases with decreasing voltage at the transmitting electrodes and with increasing gap between the transmitting and receiving electrode. A small current of the receiving electrode, when directly fed to the correlator, may be insufficient for its stable operation. The use of an amplifier allows one to reduce the amplitude of the signal supplied to the transmitting electrodes, increase the gap size to facilitate the transport of the sheet through the device, and reduce errors in finding correlations.

The use of a correlator as an AC voltage detector provides high noise immunity and reduces the requirements for shielding the receiving electrode from interference, in comparison with other types of detectors (for example, a peak amplitude detector).

The correlator can be performed in various ways in which multiplication of input signals and integration of the result are realized. One of the simplest and most accurate devices of this type is a synchronous detector, made on the basis of a standard microcircuit of a balanced mixer with an additional RC filter at the output.

A certain problem can be caused by the penetration of the exciting signal to the correlator input through spurious communications in the device. As a result, the signal at the output of the correlator receives an offset, which may be mistaken as a sign of sheet heterogeneity. The internal zero offset, which is almost inevitable when using analog correlators, such as a balanced mixer, leads to the same result. To exclude the influence of the offset of the correlator output signal, regardless of its origin, an additional measurement is carried out, in which the signals applied to two subgroups of the group electrodes are interchanged, and the resulting correlation between the excitation signal and the signal from the receiving electrode is subtracted from the originally obtained correlation . Thus, by subtracting the offset of the output signal is completely neutralized, and the useful result of determining the correlation is doubled. As a result, the reliability of detecting sheet heterogeneity is increased.

Excitation of groups of electrodes can be carried out both sequentially, group by group, and several groups simultaneously. Simultaneous excitation of several groups allows you to quickly get the result from all groups in the device, which increases its speed. Since the receiving electrode is common to all groups, in order to distinguish currents from different groups from each other, one correlator is used for each of the simultaneously excited groups. For example, if two groups are simultaneously excited, then two correlators are required. The excitation signal for each of these groups is chosen so that it does not correlate with excitation signals used for other simultaneously excited groups. For each of the correlators, as a reference signal, an excitation signal of a separate group is supplied. A signal is generated at the output of the correlator, which is proportional only to that component of the current of the receiving electrode that came from the same group. The current from other simultaneously excited groups has no correlation with the reference signal and does not contribute to the output signal of the correlator. Thus, a separate measurement of the current from each of the simultaneously excited groups is provided.

As exciting signals for the simultaneous excitation of different groups, different systems of orthogonal functions known from the theory of radio communications can be used. In the case of two simultaneously excited groups, it is most preferable to use quadrature signals of the same frequency. For more groups, you can optionally use several different frequencies. Both of these approaches are optimally combined with the use of synchronous detectors as correlators. If the device uses correlation detection using digital signal processing, then it may be optimal to simultaneously excite all groups using a set of uncorrelated pseudo-noise signals.

The claimed device allows you to measure the difference in the effective sheet thickness for each group of electrodes, while the difference is measured between the region of the electrodes of the first subgroup and the electrodes of the second subgroup. If the electrodes of the device, at each moment of time, cover only part of the surface of the sheet, then it is possible to take measurements for a significantly larger area of the sheet. To do this, you need to repeat the measurements as the sheet moves and its next sections move under the electrodes of the device. Thus, it is possible to provide an increase in the measured region of the sheet without increasing the number of electrodes. Typical is the arrangement of the transmitting electrodes in the form of a strip crossing the path perpendicular to the direction of movement of the sheet. In this case, the registration of sheet inhomogeneities begins at the moment of its entry into the indicated strip and completely ends at the moment of exit from this strip.

Consider the case when the subgroups of the transmitting electrodes included in one group are located in close proximity to each other, and the electrodes in each of these subgroups are also located in close proximity to each other. We will call this group of electrodes local. Then, the measured difference in the effective sheet thickness reflects the local thickness difference located in the area of the sheet covered by the group electrodes. If you set a threshold for the absolute value of the correlation, which is calculated for each group, then, exceeding this threshold, you can find a local thickness difference located in the region covered by the electrodes of one or another group. This method is especially effective for finding heterogeneities having a sharp boundary, for example, strips of adhesive tape glued to a sheet. Its advantage is the small amount of computation.

To study heterogeneities that do not have a sharp boundary, it is more preferable to restore the approximate distribution of the effective sheet thickness over its surface. To do this, it is necessary to measure the difference in thickness between different areas of the sheet and use the appropriate mathematical algorithm to restore the distribution. In this case, the areas along the edges of the tract that are not covered by a sheet can be used as known boundary conditions.

In one of the possible implementations, measurements are used using a set of local groups of electrodes, which make it possible to measure the difference in effective thickness over the entire area of the sheet. The measurement result using a group of electrodes gives the local value of the derivative of the sheet thickness in the direction specified between the centers of gravity of the subgroups of electrodes. Based on these values, as well as the known boundary conditions, the thickness distribution can be restored. Numerical integration methods applicable for this purpose are well known to those skilled in the art.

In other possible implementations, the groups of electrodes are spatially arranged in such a way that the output signal of the signal processing unit for each group corresponds to a certain component of the expansion of the distribution of equivalent thickness over the sheet surface over a specific functional basis. In this case, the opportunity is used to determine the spatial correlation between the regions where the electrodes of the subgroups are located and the distribution of the effective thickness over the sheet surface, known from the prior art and previously mentioned. It should be noted here that the spatial correlation in question here should not be confused with the temporal correlation of signals, which is determined by the correlator in the signal processing unit.

The most convenient for implementing a system of functions that take two or three discrete values. In one of these options, each group is made up of electrodes in such a way that the area they occupy corresponds to a specific Haar wavelet. The sections of the wavelet with a positive value correspond to the first subgroup of electrodes, and the sections with a negative value correspond to the second subgroup of electrodes. In areas with a zero value, group electrodes are not placed. In another embodiment, the electrodes of one or another group are arranged in accordance with the values of the periodic rectangular function (meander) of one or another frequency. Another possible basis is the Walsh-Hadamard system of functions. The thickness difference values obtained in each group correspond to the decomposition coefficient for that basic function, in accordance with the values of which the electrodes are located in subgroups. It is important that the applied system of functions be complete, otherwise the loss of information about the distribution of the effective sheet thickness is inevitable.

The expansion coefficients for a given functional basis can be used for the inverse transformation in order to obtain an approximate distribution of the effective thickness over the sheet surface. Further, this distribution can be analyzed by known image processing methods. An example of processing can be the search for a contrasting object, the construction and analysis of a histogram for a general assessment of the presence of heterogeneities, as well as the search for a watermark by spatial correlation with a sample. However, the decomposition coefficients can be used to directly analyze inhomogeneities without conversion to the distribution of the effective thickness. Such methods are also known from image processing technology. For example, searching for a watermark correlation on a sheet and on a known sample, using decomposition coefficients, under certain conditions, requires less computation than using a thickness distribution.

Note that the use of only local groups of electrodes has an additional advantage in conditions of low rigidity of the device structure and strong vibrations. Under these conditions, the gap between the receiving and transmitting electrodes is not only unstable, it also varies differently in time for different transmitting electrodes. This introduces especially strong errors in the operation of the device, since it worsens the compensation of the first, harmful component of the current of the receiving electrode. However, with a local arrangement of the electrodes in the group, the indicated instability of the gap for the electrodes within the group is minimized due to the close arrangement of the electrodes.

Of the above methods of electrode placement, local groups are obtained in the case of comparing the local difference with the threshold, determining the derivative in direction, as well as in the case of Haar wavelets. They are preferred for use in conditions of low structural rigidity and strong vibrations.

It should be noted that antiphase signals in engineering practice mean both signals having a relative shift by half the oscillation period and periodic signals having the opposite polarity. For signals in which the positive half-wave is symmetric to the negative half-wave, both of these interpretations of the term “antiphase” are equally valid. For the purpose of describing the claimed invention, any of these interpretations may be applied, since the difference between them does not affect the essence of the invention.

In Fig.1, the device is shown in cross section in a plane parallel to the direction of movement of the sheet.

In Fig.2, the device is shown in cross-section in a plane perpendicular to the direction of movement of the sheet.

Figure 3 shows a simplified circuit diagram of the device.

Figure 4 illustrates the grouping of the transmitting electrodes of the device in accordance with the values of the Haar wavelets.

Figure 5 shows a variant of a simplified circuit diagram of a device in which compensation for the zero offset of a balanced mixer is provided.

6 illustrates an embodiment of a device in which simultaneous excitation of two groups is used to increase speed.

Practical implementation

The practical implementation of the device is intended for use in a banknote counter. It consists of two printed circuit boards 1 and 2, installed on different sides of the path of movement of banknotes 3 on the guides 4, 5 of the path. The printed circuit board 1 contains a control unit 6 of the transmitting electrodes, and the transmitting electrodes 7 themselves, placed in a straight line perpendicular to the direction of movement of the banknotes. The transmitting electrodes 7 are the same metallization rectangles of the printed circuit board, separated by narrow gaps. To reduce the wear rate of the electrodes, a thin layer 8 of a protective wear-resistant polymer is laminated on top of them.

The printed circuit board 2 comprises a signal processing unit 9 from the receiving electrode and the receiving electrode 10 itself. The receiving 10 electrode, similarly to the transmitting 7 electrodes, is made by metallization of the printed circuit board and is also laminated by a layer 8 of a protective wear-resistant polymer. The boards 1 and 2 are connected by a cable 11, through which the supply voltage of the signal processing unit 9 is transmitted, as well as the reference and output signals of the correlator. Board 1 is connected by cable 12 to the banknote counter controller (not shown in FIG. 1 and FIG. 2). The cable 12 from the controller is supplied with the supply voltage of the device and, in the form of a binary parallel code, the number of the activated group. On the same cable from the device to the controller, an output analog signal of the correlator is supplied. The controller contains an analog-to-digital converter and a processor that allow mathematical processing of this signal to detect sheet inhomogeneities.

The receiving 10 electrode and the transmitting 7 electrodes are separated from the external fields by means of 13 shielding electrodes located in the plane of the electrodes and in the inner layer of printed circuit boards 1 and 2. The shielding 13 electrodes do not allow the electric field in the capacitor to go beyond the gap 14 between the printed circuit boards 1 , 2 and protect the device from the penetration of external fields to the receiving 10 electrode. Sixteen transmitting 7 electrodes are designated E0-E15 (Figure 3). The transmitting electrode control unit 6 comprises a crystal oscillator 15 with a frequency of 50 MHz, a counting trigger 16 with direct Q and inverse / Q inputs, and a multiplexer 17. Antiphase rectangular meanders with a frequency of 25 MHz equal to half the frequency of the crystal oscillator 15 are received from the outputs of the trigger 16. Along the lines GN0-GN3 to the multiplexer from the banknote counter controller, the number of the electrode group to be excited at the moment is transmitted. The multiplexer 17 is a combinational logic circuit, which, at each moment of time, outputs the signals Q and / Q to the outputs E0-E15 connected to the transmitting electrodes of the same name. The logic of signal wiring, implemented by the multiplexer, is determined by the principle of combining the electrodes into groups and will be considered later. The multiplexer 17 is implemented in the form of a programmable logic matrix made by CMOS technology. The signals arriving at the transmitting 7 electrodes are in the form of rectangular meanders and have a span of 3.3 B. If the signal is not supplied to the electrode, the logic level 0 is applied to it. The cards 1 and 2 are installed with a constant gap sufficient for free passage of the banknote, in including worn, wrinkled, or stuck with another banknote. The transmitting 7 electrodes E00-E15 form, together with the receiving 10 electrode, a series of measuring capacitors. The capacitance of the parasitic capacitor formed by the receiving 10 and shielding 13 electrodes is many times higher than the capacitance of each of the measuring capacitors. The signal processing unit 9 from the receiving electrode 10 contains an amplifier 18 with a resonant input circuit 19 tuned close to half the frequency of the crystal oscillator 15. From the output of the amplifier 18, the signal is supplied to the balanced mixer 20. An excitation signal from the output Q of the counting trigger is supplied to the reference input of the balanced modulator 16. The balance mixer 20 performs the function of analogously multiplying the signals supplied to its inputs. At the output of the balanced mixer 20, an RC filter 21 is connected, which sets the correlator integration time constant. From the output of the RC filter, the output signal of the device for detecting sheet inhomogeneities is sent to the banknote counter controller 22 for analog-to-digital conversion and subsequent computational processing. The input resonant circuit 19 converts the current of the receiving electrode 10 into the input voltage of the amplifier 18. Since this voltage is many times less than the excitation voltage at the transmitting electrodes, when calculating the current of the receiving electrode through the gap, we can assume that its potential is 0. The resonant nature of the input 19 circuit provides the allocation of the fundamental harmonic of the excitation frequency. Conversely, higher harmonics of the excitation frequency, as well as external RF interference, are suppressed by this circuit. The inductance in the resonant circuit 19 avoids signal loss due to the influence of stray capacitors between the receiving 10 electrode, on the one hand, the shielding 13 electrodes and the inactive transmitting 7 electrodes, on the other hand. The signal on the way from the trigger 16 through the multiplexer 17, the electrodes 7 and 10, the resonant input circuit 19 and the amplifier 18 acquires a phase shift. In order to provide the most efficient detection with a balanced mixer 20, this shift must be 0 or a multiple of 180 degrees. This value can be selected due to the parameters of the elements of the amplifier 18 and the resonant circuit 19. If, on the contrary, this shift approaches 90 degrees, then the useful signal at the output of the mixer will drop to 0. The same will happen with shifts shifted by 180 degrees from this point: 270, 550 degrees, etc. The time constant of the RC filter 21 should be many times greater than the period of the exciting signal, but significantly less than the time it takes to measure the difference in effective thickness. This time is set by the controller in the form of an interval T, during which the number of a specific group of electrodes is supplied to the inputs GN0-GN3. The RC filter suppresses ripple at the output of the balanced mixer and extracts the average value of the product of the input signals. By the end of the measurement interval, the output stabilizes the level proportional to the correlation coefficient between the signal from the receiving electrode and the reference signal.

Groups are composed of neighboring electrodes according to the following principle. Group Gi consists of two transmitting electrodes: an electrode E _i constituting the first subgroup, and an electrode E _{i + 1} constituting the second subgroup. The total number of groups is one less than the number of transmitting electrodes and equal to 15. The described principle is implemented as a logical function of multiplexer 17, which provides the wiring of Q and / Q signals along the transmitting electrodes E0-E15. The E0 electrode is located in the extreme area of the path where the sheet never gets when it moves. This electrode is used as a reference point relative to which the effective sheet thickness will be calculated. It is believed that in the region of the location of the electrode E0, the effective sheet thickness is 0. The controller 22 controls the process of recording sheet inhomogeneities. This process begins as soon as the sheet, as it moves, enters the electrode location area. The controller 22 determines this moment based on information from optical sensors located in the path and allowing to track the progress of the sheet (not shown). Inhomogeneities are recorded on the entire surface of the sheet sequentially, strip after strip, as the sheet moves in the path. Registration begins with a strip indicated by number 0. The controller 22 sequentially obtains the result of measuring the difference in effective sheet thickness for each group of electrodes. To do this, he sets the number of the initial group (G0) to the inputs GN0-GN3 of the multiplexer and starts the countdown of the measurement time interval T. At the end of the interval, the controller 22 performs an analog-to-digital voltage conversion at the output of the RC filter 21 and stores its value D _{0 0} in the internal memory. Here, the first index indicates the group number, and the second indicates the number of the strip in the measurement sequence. Next, the controller 22 proceeds to the measurement using the following (G1) group of electrodes, which is produced in a very similar way. This process is repeated until the results of measuring D _{0 0} -D _{14 0} for all fifteen electrode groups (G0-G14) are remembered. At its end, the processor’s memory contains data on the differences in the effective sheet thickness over fifteen pairs of areas related to the strip on the sheet covered by the electrodes of the device. The areas in each of these pairs are zones covered by the first and second subgroups of electrodes of the corresponding group. The width of the strip is equal to the size of the transmitting electrode L in the direction of movement of the sheet along the path. As soon as the sheet in the path advances by an amount approximately equal to the length of the transmitting electrode L, the measurement process is repeated, as a result of which the data on the differences in the effective thickness D _{0 1} -D _{14 1} related to the next (along the sheet) strip are stored in the memory of the controller 22 . These steps are repeated until the sheet leaves the zone of arrangement of the electrodes of the device. At this point, the controller stores data D _{0 0} -D _{14 N} o the differences in the effective sheet thickness related to its entire surface. Here, the letter N denotes the maximum number of the strip registered on the sheet.

The heterogeneity of the sheet is detected based on the distribution of the effective thickness of the sheet, which is based on the data on the differences in the effective thickness stored by the controller 22. The distribution of the effective thickness, from the point of view of the theory of image processing, is a two-dimensional image with a pixel size equal to the size of the transmitting electrode. The value of the pixel Pij in the distribution of the effective sheet thickness is proportional to the average effective sheet thickness within this pixel.

Note that the values of Dij are proportional to the derivative of the effective sheet thickness in the direction perpendicular to the direction of movement of the sheet. The value of the pixels Pij is restored in strips by the method of numerical integration. Index i denotes the pixel number within the strip, counted similarly to the transmitting electrodes E0-E15. Recovery begins with a zero pixel corresponding to the E0 electrode, in the region of which the sheet never appears in the path. Suppose P _{0 0} = 0. Then calculate the following pixel: P _{1 0} = P _{0 0} + D _{0 0} . The process is repeated sequentially for all pixels in the strip, so that the last pixel in it is calculated as P _{15 0} = P _{14 0} + D _{14 0} . The restoration of the next strip starts in the same way, from the zero pixel P _{0 1} = 0, and ends with the last P _{15 1} = P _{14 1} + D _{14 1} . After repeating for all N + 1 bands, the distribution of the effective sheet thickness is completely restored. Naturally, it is approximate, since it is limited by the discreteness of dividing the sheet surface into pixels.

Further processing of the resulting distribution is determined by what kind of heterogeneity the device should detect. One of the urgent tasks of analyzing the suitability of banknotes for further circulation is the detection of glued banknotes. To detect pieces of adhesive tape on a sheet, it is enough to set a certain threshold for the absolute difference between adjacent pixels, and look for distribution areas where this threshold is exceeded. If this threshold is exceeded in any place of the sheet, except for its edges, then in this place there is a piece of adhesive tape or another foreign object with sharp boundaries.

For a more complex analysis, which includes determining the correctness of the structure of the watermark, it is necessary to apply correlation analysis. Namely, the correlation coefficient between the obtained distribution of the effective thickness and the known distribution of the effective thickness of the reference sheet should be calculated. If the watermark pattern coincides, the correlation coefficient will have a positive value close to unity. If the watermarks are different, then the correlation coefficient will have a negative or a small positive value. This task of image processing is typical in the analysis of the authenticity of banknotes. Similarly, the presence of polymer security elements such as ribbons or appliqués can be analyzed.

It is possible to use a different principle of dividing the electrodes into groups to directly obtain the wavelet transform coefficients as a result of registration. The distribution of the equivalent sheet thickness along the strip is subjected to wavelet transform. The selected complete basis of the 15 Haar wavelets is shown in FIG. 4. Each of the 15 groups of electrodes corresponds to a specific wavelet. Haar wavelet, by definition, takes the values 1, -1 and 0. For each wavelet, we include in the first subgroup those electrodes in the region where it takes the value 1, and in the second one those electrodes on which the wavelet value is -1. Depending on the frequency parameter of the wavelet, each subgroup may include from 1 to 8 electrodes. The combination logic of connecting the Q and / Q signals corresponding to the described splitting of the electrodes into groups is implemented in the multiplexer 17.

The process of registering a set of values D _{0 K} -D _{14 K} for each band does not differ from the previously described. The index K indicates the number of the strip. The values in the resulting set are proportional to the coefficients of the wavelet transform of the equivalent sheet thickness along the strip and with sampling by the size of the transmitting electrode. Indeed, this value of D _{i K is} proportional to the difference between the sheet thickness in the area covered by the electrodes of the first subgroup and the area covered by the electrodes of the second subgroup. But since the first subgroup refers to the zone with the value of wavelet 1, and the second with the value -1, then D _{i K is} proportional to the result of numerical integration (summation) of the product of the distribution of the equivalent sheet thickness in the strip K by the wavelet corresponding to index i.

As is known to specialists, a discrete wavelet transform implements a sequential refinement due to a sequential twofold increase in the wavelet frequency. The groups depicted in FIG. 4 implement sequential refinement of the difference in effective thickness in the strip controlled by the device. The G0 group provides the coarsest measurement of the difference in effective thickness between the two halves of the strip: the first half is covered with E0-E7 electrodes, and the second with E8-E15 electrodes. The difference in effective thickness within these halves is specified by groups G1, G8, which have twice the spatial frequency of the wavelet, which provides a spatial resolution of a quarter band. Namely, the difference in the effective thickness in the sections E0-E3 and E4-E7, as well as E8-E11 and E12-E15, respectively, is determined. Within these sections in the ¼ band, further refinement is provided by groups G2, G5, G9, G12. As a result, a resolution of 1/8 band is realized. And finally, the groups G3, G4, G6, G7, G10, G11, G13, G14 perform the last refinement step, providing a resolution of 1/16 band.

может быть использована для восстановления дискретизированного распределения

эффективной толщины листа по известным алгоритмам обратного вейвлетного преобразования. Отметим, что, без применения дополнительной информации, набор из 15 вейвлетных коэффициентов недостаточен для восстановления дискретизированного распределения по 16 площадкам, соответствующим областям покрытия электродов. В качестве такой дополнительной информации могут быть использованы известные граничные условия, как, например, нулевая толщина листа в области электрода Е0 (куда лист не может попасть из-за конструкции тракта). Либо, в качестве дополнительной информации может использоваться среднее значение эффективной толщины листа по всей площади полосы, которую можно измерить, подав один и тот же сигнал Q на все электроды Е0-E15. В этом, последнем, случае конфигурация электродов и принцип измерения будет соответствовать простому емкостному датчику толщины листа, подобно описанному в EP 0097570. Однако многие из алгоритмов обработки и распознавания изображения работают непосредственно с коэффициентами вейвлетного преобразования. В качестве примера можно назвать алгоритмы фильтрации или обнаружения границ. Для этих алгоритмов матрица результатов регистрации

, содержащая эти коэффициенты, может быть использована непосредственно, в качестве входных данных. Таким образом, исключается дополнительный этап вычислений для прямого вейвлетного преобразования, что повышает быстродействие способа обнаружения неоднородностей листа.After registration for the whole sheet, the results matrix

can be used to restore the discretized distribution

effective sheet thickness according to well-known inverse wavelet transform algorithms. Note that, without the use of additional information, a set of 15 wavelet coefficients is insufficient to reconstruct the discretized distribution over 16 sites corresponding to the electrode coverage areas. As such additional information, known boundary conditions can be used, such as, for example, zero sheet thickness in the region of the E0 electrode (where the sheet cannot get due to the design of the path). Or, as additional information, the average value of the effective sheet thickness over the entire area of the strip, which can be measured by applying the same signal Q to all electrodes E0-E15, can be used. In this last case, the configuration of the electrodes and the measurement principle will correspond to a simple capacitive sheet thickness sensor, similar to that described in EP 0097570. However, many of the image processing and recognition algorithms work directly with wavelet transform coefficients. Examples include filtering or border detection algorithms. For these algorithms, the registration results matrix

containing these coefficients can be used directly as input. Thus, an additional calculation step for direct wavelet transform is excluded, which increases the speed of the method for detecting sheet inhomogeneities.

Depending on the particularities of the problem of detecting inhomogeneities, the decomposition coefficients can be realized in a similar way on a different functional basis. For example, when using the basis of the Walsh-Hadamard functions, the first subgroup includes those electrodes in the region where the Walsh-Hadamard function takes the value 1, and the second includes those electrodes on which its value is 0.

The originally considered registration method using local groups consisting of only two electrodes can also be considered from the perspective of a wavelet transform. Namely, in this case, each group corresponds to a Haar wavelet of the same frequency, and these wavelets differ only in location. To compensate for the zero shift of the balanced mixer, as well as the penetration of its reference signal to the input of the amplifier 18, exclusive-OR logic elements 23 and 24 are included in the device, which are connected between the outputs of the trigger 16 and the inputs of the multiplexer 17, as shown in FIG. 5. The GN4 signal supplied to the inputs of these logic elements from the output of the controller 22 allows, if necessary, to invert both signals Q and / Q. If a low logic level is applied to GN4, inversion does not occur, and the device operates identically to that shown in Figure 3. When GN4 is supplied with a high logical level, the excitation signals supplied to the electrodes of the first and second subgroups in accordance with the group number on the GN0-GN3 lines are interchanged. Accordingly, with this inversion, the polarity of the output signal at the output of the correlator (output of the RC filter 21) changes.

To implement the specified compensation, for each group number on the GN0-GN3 lines, under the control of the controller, the measurement is performed twice in a row - once with a low level on GN4, and a second time with a high one. The second is subtracted from the first obtained measurement result, and this difference is considered the result of measuring the difference in effective thickness between the sections of the sheet, determined by the location of the electrodes of the group whose number was issued on GN0-GN3. In this operation, a constant level of zero shift is completely subtracted, and the useful signal corresponding to the difference in effective thickness is increased by 2 times.

Note that, in this case, a signal can be supplied to the ADC input of the controller through capacitive isolation, which excludes the DC component. Indeed, the signal at the output of the RC filter 21 is an alternating voltage, to which a constant level is superimposed. A positive half-wave corresponds to a low GN4 level, and a negative half corresponds to a high level on this line. When registering, only the magnitude of the alternating voltage has value, while the constant level is still reduced to zero when subtracted. However, a constant level at the input of the ADC can limit the usefulness of its operating range, so it is advisable to exclude a constant level using capacitive isolation. If the useful level at the output of the RC filter 21 has a small amplitude, then to transfer it to the ADC, it is possible to use an additional AC voltage amplifier.

If the device has high performance requirements, then it is possible to speed up its work by almost 2 times. This is achieved due to the simultaneous supply of exciting signals to the electrodes of two groups, moreover, the exciting signals of these groups have a mutual phase shift of 90 degrees. For this, as shown in Fig.6, the trigger 16 is replaced by a shift 4-bit register 25, closed in a ring. The initial loading of the register 25 occurs on the signal RESET, issued when the device is turned on. When loading through the inputs PD0-PD3, the binary number 0011 is set in the register 25. The outputs of the register are supplied to the multiplexer 17, the multiplexer having input signals Q0 and / Q0 to excite the first of two groups, and signals Q1 and / Q1 to excite the second of two groups . The generator 15 generates a frequency of 100 MHz supplied to the clock input of the shift of the contents of the register 25. As a result, at the inputs Q0 and / Q0 of the multiplexer 17 are formed antiphase signals with a frequency of 25 MHz. At the inputs Q1 and / Q1, antiphase signals are also formed with the same frequency, but the signal Q1 is shifted 90 degrees relative to Q0.

The multiplexer 17 has three inputs GN0-GN2 for selecting the excited groups, which corresponds to 8 different states. Its combination logic provides the simultaneous supply of excitation to groups that do not have common electrodes. The groups corresponding to the wavelet decomposition shown in FIG. 4 are used. In the ground state, only the group G0 is excited, since it involves all the transmitting electrodes. In state 1, groups G1 and G8 are excited, in state 2, G2 and G5, 3 - G3 and G4, 4 - G6 and G7, 5 - G9 and G12, 6 - G10 and G11, 7 - G13 and G14. Groups can be related to the state of the multiplexer in a different way. It is important that at the same time two groups are excited that do not have common electrodes, and when the states change from 0 to 7, as a result, all available groups will be excited.

An additional balanced mixer 26 and an additional RC filter 27 are introduced into the signal processing unit 6 from the receiving electrode. An Q1 signal is supplied to the reference input of the mixer 26. Since the signals at the receiving electrode received from two simultaneously excited groups have a 90 degree phase shift (are quadrature), they are independently detected by mixers 20 and 26. After smoothing by RC filters 21 and 27, the detection results are fed to two separate channels of the controller's ADC . Thus, the device registers with 15 groups of electrodes in just 8 separate measurement intervals, compared with 15 intervals in the device shown in Fig.3. This result is achieved due to a slight complication of the transmitting electrode control unit 6 and the signal processing unit 9 from the receiving electrode. Note that the requirements for the electrodes 7 and 10, the input circuit 19 and the amplifier 18 are not changed. Further processing of the obtained results does not differ from that described previously.

Imperfections in the mechanical design of the device, expressed in an unstable gap between the electrodes or deviation of their area from the calculated one, can lead to incomplete compensation of the current of the receiving electrode, which occurs in the absence of sheet inhomogeneities. To exclude this current from the registration results, it is necessary, when the device is turned on, to register in the absence of a sheet in the path. The obtained values for each group are stored in the memory of the controller 22 and, subsequently, are always subtracted from the registration results. Thus, the influence of the mentioned imperfections is compensated. Registration in the absence of a sheet in the path can be carried out periodically to take into account mechanical deviations that change during the operation of the device.

The apparatus and method described herein is one of many implementations possible in accordance with the claims.

Claims (24)

1. A method for detecting sheet inhomogeneity, according to which the sheet is moved in the path through a device containing transmitting electrodes on one side of the path, and a receiving electrode placed on the opposite side of the path opposite the location of the transmitting electrodes, the transmitting electrodes are divided into groups, one and the same transmitting electrode may belong to more than one group, each group is divided into two subgroups, each of which contains at least one electrode, an analysis of the presence inhomogeneities of the sheet in the region of the location of the transmitting electrodes of each group, for which a stimulating signal is supplied to the first subgroup of electrodes in the group, and an antiphase signal to the first subgroup is supplied to the second subgroup of electrodes in the group, and the correlation between the exciting signal and the signal from the receiving electrode is determined, analyze the correlation and judge by it the presence of heterogeneity of the sheet in the specified area, and based on the judgments obtained for each group, judge the presence of heterogeneity of the sheet in Asti electrode arrangement. 2. The method according to claim 1, in which when analyzing the presence of inhomogeneities of the sheet in the location of the transmitting electrodes of each group, an additional measurement is carried out, in which the signals applied to the two subgroups of the electrodes of the group are interchanged, and the resulting correlation between the exciting signal and the signal is subtracted from the receiving electrode from the originally obtained correlation. 3. The method according to any one of claims 1 and 2, in which at least two groups of electrodes are simultaneously excited, and the excitation signal for each of these groups is selected so that it does not correlate with the excitation signals used for the others simultaneously excited groups. 4. The method according to any one of claims 1 and 2, in which, as the sheet moves in the path, the detection of sheet inhomogeneities is repeated at least once more. 5. The method according to claim 3, in which, as the sheet moves in the path, the detection of sheet inhomogeneities is repeated at least once more. 6. The method according to any one of claims 1, 2, 5, in which subgroups of transmitting electrodes included in one group are located in close proximity to each other, and the electrodes in each of these subgroups are also located in close proximity to each other, and To detect heterogeneity during the analysis, a threshold for the absolute value of correlation, which is calculated for each group, is set, and when the threshold is exceeded, a decision is made on the presence of sheet inhomogeneity in the area of the group's transmitting electrodes. 7. The method according to claim 3, in which subgroups of transmitting electrodes included in one group are located in close proximity to each other, and the electrodes in each of these subgroups are also located in close proximity to each other, and to detect heterogeneity during analysis set the threshold for the absolute value of the correlation, which is calculated for each group, and when the threshold is exceeded, a decision is made on the presence of sheet inhomogeneity in the region where the group’s transmitting electrodes are located. 8. The method according to claim 4, in which subgroups of transmitting electrodes included in one group are located in close proximity to each other, and the electrodes in each of these subgroups are also located in close proximity to each other, and to detect heterogeneity during analysis set the threshold for the absolute value of the correlation, which is calculated for each group, and when the threshold is exceeded, a decision is made on the presence of sheet inhomogeneity in the region where the group's transmitting electrodes are located. 9. The method according to any one of claims 1, 2, 5, in which, based on the obtained correlation values, the approximate distribution of the equivalent sheet thickness over its surface is restored, and then the named distribution is analyzed to detect heterogeneity. 10. The method according to claim 3, in which, based on the obtained correlation values, the approximate distribution of the equivalent sheet thickness over its surface is restored, and then the named distribution is analyzed to detect heterogeneity. 11. The method according to claim 4, in which, based on the obtained correlation values, the approximate distribution of the equivalent sheet thickness over its surface is restored, and then the named distribution is analyzed to detect heterogeneity. 12. The method according to claim 9, in which the groups of electrodes are spatially arranged so that the correlation measurement result for each group corresponds to a certain component of the decomposition of the distribution of equivalent thickness over the sheet surface according to a specific functional basis. 13. The method according to any one of claims 10 and 11, in which the groups of electrodes are spatially arranged so that the correlation measurement result for each group corresponds to a certain component of the decomposition of the distribution of equivalent thickness over the sheet surface according to a specific functional basis. 14. The method of claim 12, wherein the Haar wavelet system is selected as the functional basis. 15. The method according to item 13, in which the Haar wavelet system is selected as the functional basis. 16. The method according to item 12, in which the system of Walsh-Hadamard functions is selected as the functional basis. 17. The method according to item 13, in which the system of Walsh-Hadamard functions is selected as the functional basis. 18. The method according to claim 9, in which the subgroups of the transmitting electrodes included in one group are located in close proximity to each other, and the electrodes in each of these subgroups are also located in close proximity to each other, and the values of the correlation calculated for of each group are considered as values proportional to the derivative of the equivalent sheet thickness in the direction between the centers of gravity of the first and second subgroups of transmitting electrodes. 19. The method according to any one of claims 10 and 11, in which the subgroups of the transmitting electrodes included in one group are located in close proximity to each other, and the electrodes in each of these subgroups are also located in close proximity to each other, and the values correlations calculated for each group are considered as proportional to the derivative of the equivalent sheet thickness in the direction between the centers of gravity of the first and second subgroups of transmitting electrodes. 20. A device for detecting inhomogeneity of the sheet, containing transmitting electrodes placed on one side of the sheet, divided into groups, each of which consists of two subgroups, each of the subgroups contains at least one electrode, at least one a transmitting electrode may belong to more than one group, a receiving electrode placed on the opposite side of the sheet and covering the area of location of the transmitting electrodes, the control unit of the transmitting electrodes, and the synchronized block about it the signal from the receiving electrode, while the control unit of the transmitting electrodes is configured to excite the electrodes of each group with the supply of the exciting signal to the first subgroup of electrodes and simultaneously supplying a signal in antiphase exciting to the second subgroup of electrodes, and the signal processing unit from the receiving electrode contains a device for determine the correlation value of the exciting signal and the signal from the receiving electrode. 21. The device according to claim 20, in which the transmitting electrodes are made in the form of non-overlapping surface sections, the receiving electrode is made in the form of a surface section and is equidistant from the transmitting electrodes, while the transmitting electrode control unit is configured to provide a ratio of the amplitudes of the exciting signal and its antiphase signal applied to the subgroups of the electrodes of the group, proportional to the ratio of the total areas of the electrodes of the second subgroup and the first subgroup. 22. The device according to claim 20, in which the signal processing unit from the receiving electrode comprises a signal amplifier from the receiving electrode. 23. The device according to any one of paragraphs.20 and 21, in which the device for determining the magnitude of the correlation is a synchronous detector. 24. The device according to item 22, in which the signal processing unit is equipped with at least one additional device for determining the correlation of the excitation signal and the signal from the receiving electrode, so that the excitation signal of each of the simultaneously excited groups is supplied to the input of a separate device for determining the correlation .

Images