EP3400584A1

EP3400584A1 - Completeness check of a value document

Info

Publication number: EP3400584A1
Application number: EP16819440.5A
Authority: EP
Inventors: Wolfgang Rauscher; Erich KERST; Thomas Happ
Original assignee: Giesecke and Devrient Currency Technology GmbH
Current assignee: Giesecke and Devrient Currency Technology GmbH
Priority date: 2016-01-05
Filing date: 2016-12-21
Publication date: 2018-11-14
Anticipated expiration: 2036-12-21
Also published as: RU2018127765A3; RU2018127765A; DE102016000011A1; WO2017118467A1; EP3400584B1; US11823522B2; RU2724173C2; US20200273279A1

Abstract

The invention relates to a method, a sensor, a sensor unit, and a banknote processing machine for checking the completeness and/or authenticity of value documents. A value document comprises at least one machine-readable distinguishing substance at at least two locations. According to the method, the value document is excited (S2) at least locally at measurement locations. Furthermore, a distinguishing intensity of the machine-readable distinguishing substance is detected (S3a) at multiple different locations of the value document in a spatially resolved manner. The location-related distinguishing intensities are classified (S5) in a location-related manner using a threshold. Furthermore, location-related limits of an expected location distribution of the machine-readable distinguishing substance are determined (S6). Lastly, a location-related distribution of the classified distinguishing intensities is evaluated (ST).

Description

V o 1 1 s t i n d i g e s t i n g e n t i o n s w e r t d o k m e n t s

The present invention relates to a method and a corresponding device for checking value documents for completeness and authenticity. Forgeries of value documents could be composed of a plurality of sub-documents for which e.g. Sections of true value documents were combined with sections of copies. According to the invention, it is possible to identify such counterfeits safely and to check value dummies for their completeness or authenticity.

DE 197 14 519 A1 teaches scanning a document to be tested over the entire surface or along a defined measurement track with a sensor suitable for detecting the marker substance. The distribution of the marker signal is determined and compared with the expected signal curve given by the pattern of the marker printed on the marker substance. In particular, the general presence of the feature substance, jumps in its distribution as well as areas deviating from the expected reference distribution are examined. DE 10 346 636 A1 describes a sensor-based authenticity check of value documents with a luminescence marker, which takes place integrally along a track across the value document. While the addition of the luminescence signal along the measuring track is well suited for the detection of small, noisy spectral signals, this just prevents a small-scale and therefore precise evaluation of the completeness.

WO 2011/037750 A2 describes the authenticity recognition of bank notes by detecting a homogeneously distributed IR luminescent substance Measuring tracks and adjustment of the measured modulation of the luminescence intensity by overprinting or applied holograms, stripes, etc. with expected target profiles. In this case, areas with high statistical fluctuation, such as eg security thread or hologram stripes, are excluded from the evaluation and an authenticity decision is made if, for example,> 51% of the measured profile coincides with one of the four position-dependent authenticity references.

Thus, under ideal conditions, a completeness statement along the measured track is possible. When measuring on fast-running banknote processing machines under real conditions

However, track variations in machine banknotes, as well as aging or stains on the banknotes, are classified as false due to this method. Conversely, in the described banknotes wrongly classified as wrongly described, weak authenticity criteria of e.g. Only> 51% consistency along a track recognizes a large number of snippet counterfeits as genuine, so that this method does not solve the problem satisfactorily. US 6,393,140 B1 describes another method of bill validation in which a signal, e.g. the color or magnetism is measured at a plurality of defined locations of the banknote and in each case the relative distances of the measured values to a reference value are determined and then normalized. Although this method allows a local authenticity assessment, it does not allow a reliable completeness check.

Especially when measuring on high-speed banknote processing machines with processing speeds up to 12 m / s, however, there are additional challenges beyond a slow ticket inspection which have to be addressed by special methods and algorithms. Only then will it be possible, even under such difficult conditions, to enable a reliably functioning completeness check. For example, it is not always ensured on a banknote processing machine that the denomination or position of the value document and thus the reference distribution to be compared is known at the time of the evaluation.

Furthermore, the achievable spatial resolution of the feature signal can be dramatically reduced in comparison to conventional resolutions of optical image sensors in the visible: the spatial resolution can be determined both by the detector technology used and by intrinsic time characteristics of the feature substance, such as e.g. be limited to the onset of a luminescent. Especially with track-bound sensors, the pixel size can be in the range of a few mm or even a few cm. In order to be able to derive the completeness of the value document as reliably as possible in such situations from the feature measurement, the information of each individual measuring pixel must be adequately evaluated and used for the completeness check.

The present invention aims to provide a reliable completeness measurement or completeness check of modern value documents for the detection of so-called snippet counterfeits under the conditions of fast-running banknote processing machines (ie measurement at relative speeds of, for example, 1-13 m / s, preferably 6-12 m / s between banknote and sensor). The combination of the most diverse security features on the value document and their interaction with the sensor-based measurement of the machine-readable feature also occurs in the case of homogeneous machine-readable features. Often malstoff a complex modulation pattern of the measured intensity of the feature signal on. This makes the direct completeness assessment considerably more difficult. In this case, it is assumed, in particular, from the case frequently encountered in reality that there is no denomination or position information at the time of the measurement or the completeness assessment.

It is thus an object of the present invention to provide a method and apparatus which perform a reliable evaluation of the completeness of a document.

This object is achieved by a method and a device for checking the completeness of value documents with the features of the independent claims. In the dependent claims advantageous embodiments and developments of the invention are given.

Accordingly, a method for checking completeness and / or authenticity of value documents is proposed. According to the invention, at least one value document has at least one machine-readable feature at at least two different locations. In one step, the value document is at least locally stimulated. This can be done for example by means of electromagnetic radiation, for example light having a wavelength in the visible spectral range. In addition or alternatively, a magnetic loading of the value document can take place.

Furthermore, according to the invention, a feature intensity with respect to the machine-readable feature substance is detected at a plurality of different locations of the value document. For this purpose, an optical and / or magnetic detection unit can be used. The registration unit forms Add a characteristic value to each measuring location. From the feature value, the feature intensity is determined location-related to the measurement locations of the value document. The localized surface element of the location-related feature intensity and / or the location-related remission value can subsequently be understood as a pixel. In this case, a feature value or a plurality of, in particular contiguous, feature values can be used. Furthermore, a partial aspect of the feature value, for example a specific wavelength when detecting a spectral range, can be used to generate or capture the feature intensity.

In a further step, the location-related feature intensities are classified based on a threshold value. On the basis of the threshold value, a location-related classification of the feature intensity can be carried out for each location-related feature intensity, for example in locally true or locally unreal. Furthermore, location-related limits of an expected spatial distribution of the machine-readable feature substance are determined. These limits preferably reflect the longitudinal extent and / or width extent, particularly preferably the area extent, of the value document. Furthermore, based on the limits, errors in the structure, for example falsifications of a value document, in particular when falling below a minimum length, can be determined.

In a further step, the location-related distribution of the classified feature intensities is evaluated. In this case, at least two classified feature intensities are evaluated in relation to one another and / or to the specific location-related limits.

The feature intensities at the individual measurement locations are preferably given with suprathreshold intensity as locally true or with subliminal intensities. classified as locally spurious. These measuring locations are also referred to below as classified pixels.

To evaluate the location-related distribution of the classified feature intensities, it is possible, for example, to determine from the number and spatial distribution of subliminal and / or suprathreshold feature intensities. As an alternative to an optionally local threshold value, a reference feature intensity can be used as comparison value.

Under value documents are sheet-like objects with a front and a back understood that represent, for example, a monetary value or an authorization and therefore should not be arbitrarily produced by unauthorized persons. They therefore have features which are not easy to produce, in particular to be copied, whose existence is an indication of the authenticity, i. the manufacture by an authorized agency. Important examples of such value documents are coupons, vouchers, checks and in particular banknotes.

By using the method according to the invention in the two alternative variants of the measurement of the machine-readable feature with or without measurement of the remission values, a reliable evaluation of the completeness or authenticity succeeds. In one embodiment, at least one location or pixel is assigned a specific location-dependent threshold value. Preferably, a plurality of locations or pixels or a group thereof is assigned a location-dependent threshold value. Based on location-dependent thresholds, a more detailed location-dependent classification of the location-related feature intensities is possible. In particular, special the location-dependent properties are represented and checked by the location-dependent threshold values.

According to one embodiment of the invention, the local authenticity is determined on the basis of a feature intensity. Preferably, a plurality of feature values, for example of a luminescence radiation, are used for the assessment of the local authenticity. This includes a temporal behavior of the feature intensity, such as an attack behavior, a decay behavior, a spectral distribution of the feature intensities, and / or spatial information such as e.g. a lane information or a transport position. In particular, a specific comparison of the determined feature values with the expected feature values of a genuine value document can be carried out and taken into account when determining the feature intensity. For example, despite the presence of significant luminescence intensity, the feature intensity may be e.g. are determined to be zero if the spectral distribution of the luminescent radiation does not match the expected spectrum.

Accordingly, with the present invention, based on a local feature value, a local authenticity of the value document can be determined. In addition, an evaluation of the entire value document for completeness and / or authenticity is possible.

In one embodiment of the invention, remittance values are recorded spatially resolved at a plurality of different locations of the value document. The measurement locations of the remission values preferably correspond essentially to the measurement locations of the feature intensities. The area of the measurement location of a remission value may be larger or smaller than the area of the measurement location of the corresponding feature intensity. Preferably, the measuring locations an equal area on. The measuring location of a remission value can also be offset, preferably designed to overlap the measuring location of a feature intensity. The spatially resolved detection of the remission values is preferably carried out simultaneously with the spatially resolved detection of the feature intensity values.

The acquired location-related remission values may, according to one embodiment, be parameters of a characteristic for location-dependent or location-related threshold values.

According to the invention, in one embodiment, the track integrity is determined. In each case, characteristic values, i. location-related feature intensities and, where appropriate, location-related remission values along a measurement track on the value document are detected and taken into account for determining the track completeness.

For example, the data of this measuring track are evaluated by using the invention as a single-track sensor. Essentially, however, only one-dimensional track completion can then be tested and evaluated. For this purpose, first the individual feature intensities and, if appropriate, remission values of the pixels are individually compared with a minimum value and classified in real or spurious, or corresponding, for example, as subliminal, suprathreshold or average. The length of the value document can be determined from the distance between the two outermost measuring points with a signal intensity above a minimum threshold.

Furthermore, the measured length of the value document can preferably be determined in a transport direction by an edge detection from a feature input. intensity curve are determined. In the feature intensity curve spatially resolved feature intensities are recorded. That is, the feature intensity curve includes value points resulting from the feature intensity and the associated location. In the simplest case, the extreme spatial positions are determined at which an averaged threshold (upper threshold - lower threshold) / 2 of the feature intensities is reached. Preferably, instead of the upper threshold and lower threshold, quantile values are used, e.g. B. 75% (= almost white) and z. B. 5% quantiles (= almost black) corresponding to the maximum feature intensity. The difference between the two location-related feature intensities results in the measured length of the value document, whereby the size of the pixel or location of the location is usually taken into account, in particular added.

In one embodiment, the accuracy of the length determination can be increased by interpolating the feature intensity curve between the measured points, preferably linearly (alternatively spline), and thus determining the length subpixel-exactly.

The determined length is then compared to a known minimum length, e.g. corresponds to the true length of the shortest denomination variant of a banknote series.

If the minimum length is not reached, it is in any case assumed that there is a forgery share. The completeness can be determined quantitatively simply by the ratio of the number of pixels with suprathreshold feature intensity, or real pixel to the number of pixels corresponding to the measured length (assuming a constant transport speed or

Spatial resolution). Preferably, by a light barrier (eg, the banknote processing machine), the measurement of the feature sensor is triggered so that the measurement points of different value documents always different

Records can be assigned. The evaluation for completeness from the length determination thus usually concerns exactly one value document.

Basically, however, within a single (real) value document also areas with significantly reduced feature intensity can occur, up to completely vanishing feature intensity, if for the excitation radiation and / or from the value document outgoing radiation opaque features such. metallized holograms, security strips, or similar Cover the machine-readable feature substance. The maximum possible width of the opaque covers, depending on the denomination, and their position relative to the front and / or rear edge in the transport direction are taken into account for the completeness evaluation.

In one embodiment for calculating or determining the track completeness, a sensor has a plurality of measuring tracks, such as, for example, a sensor. B. 2, 4, 6, 8, 10, 20 or more measuring tracks, so that a two-dimensional distribution of the feature intensities is recorded. For the completeness evaluation, a threshold value classification of the individual location-related feature intensities is initially also carried out. Subsequently, a convex hull is calculated around the determined suprathreshold location-related feature intensities. Furthermore, known or predetermined location-related subliminal feature intensities, for example by a background system, are compared with the determined location-related suprathreshold feature intensities covered by the convex hull, for example a number. If, for example, the number of ascertained thresholds ligen feature intensities smaller than the known number of suprathreshold feature intensities, it can be assumed, for example, that an undesired hole in the value document is missing. This method thus permits the detection of holes or opaque spots within the value document which are not wanted and therefore indicate a forgery.

Instead of a convex hull to suprathreshold location-related feature intensities, the evaluation can be carried out at subliminal location-related feature intensities. Furthermore, an evaluation with respect to possibly detected and known or predetermined remission values is possible.

In one embodiment, the convex hull is calculated separately for each row, i. in this case, the interval between the front and back of the row. For each interval, subliminal location-related feature intensities are marked as spurious or corresponding.

In one embodiment, a two-dimensional convex hull is calculated over the locations of all the suprathreshold pixels, e.g. with the Graham algorithm. For example, all subliminal pixels with locations within the convex hull are marked as spurious or equivalent.

In one embodiment, subliminal location-related feature intensities within the convex hull may be rejected if, for example, their measurement locations fall within occurring patterns, such as e.g. a transparent window or a metal LEAD strip lie.

Furthermore, the two-dimensional distribution of the suprathreshold pixels is preferably analyzed and evaluated. In particular, the Occurrence of larger holes tested. In particular, location-related feature intensities classified as incorrect or correspondingly are determined and identified, marked and counted in two-dimensional contiguous areas. Lying z. If, for example, more than 2, 3, 5,... (Depending on the resolution) are associated with location-related feature intensities that are associated with each other as spurious or correspondingly, then a potentially missing area is recognized. Subsequently, the position and geometric extent of the regions classified as spurious or correspondingly are analyzed and compared with patterns that occur in a known manner, such as, for example, a transparent window or a metallic LEAD strip. In particular, the shape, maximum width and relative position to the edges or corners of the value document, is checked for plausibility and classified in case of deviations as "incomplete" or similar. In an embodiment with greatly different resolving power of the measurement in the (x) track direction and the y direction (number of tracks), it is possible to selectively count as pseudo or correspondingly classified neighboring pixels in the row direction and to evaluate multiple pixels in this direction.

Subsequently, the individual tracks are evaluated analogously to the single-track sensor described above, which provides several measurements for the track completeness. In one embodiment, the authenticity and / or completeness of the value document is recognized if there is at least one specific relationship between the number and a spatial distribution of the classified pixels, feature intensities and / or remission values. In one embodiment, the measured length can be determined from the maximum value of the individual determined track lengths and, in the case of deviations of the other track lengths, preferably closed taking into account a tolerance value for the presence of missing completeness.

Analogous to the track completeness, a complete column can be determined, in particular by taking into account a plurality of measuring tracks.

In an embodiment, the spatially resolved detection of the feature intensities may comprise the measurement of a spectral luminescence intensity of a luminescent substance. Accordingly, the value document can

Presence or non-presence of a luminescent tested and tested for authenticity and completeness according to its local allocation or distribution. Furthermore, the spatially resolved detection can comprise a spectral measurement of a Raman band and / or a so-called surface-enhanced Raman spectroscopy (SERS). In addition, the detection may comprise a measurement of an absorption band with respect to a specific spectral range, for example infrared and / or the measurement according to magnetic properties.

The detection of feature intensities may be preceded by the detection of feature values. Feature values may include measurement results, for example with respect to a spectrum. In particular, the feature values are specifically processed to provide feature intensities. The characteristic values can be used with a filter, for example for

Evaluation of a spectral range, in particular a wavelength occupied. Furthermore, the characteristic values can be assigned an algorithm. The feature values may be sensor values from which feature intensities are eventually determined. The characteristic values each may comprise a plurality of different feature intensities.

The detection of feature intensities and / or, if appropriate, remission values can take place both on a front side and on a back side of the value document. In particular, feature intensities and / or, if appropriate, remission values can be detected on the same and / or opposite side, in particular with respect to the measurement location. Preferably, the feature intensities and / or, if appropriate, remission values are recorded at the same, opposite locations of the front and back sides.

In one embodiment, location-dependent threshold values can be determined by a characteristic which depends on the feature intensities determined at the respective location on the opposite side of the value document

In one embodiment, transmission values are detected spatially resolved, preferably by time-shifted illumination in the context of remission measurements on the front and back and / or by time-shifted illumination in the context of the measurement of feature values or the detection of feature intensities on the front and back of the value document. In one embodiment, in each case a combined classification taking into account data tuples assigned to the measurement locations can be undertaken at a plurality of measurement locations. The data tuples comprise at least one feature intensity and at least one of the following components: a further feature intensity, a remission value, and / or a transmission value.

Furthermore, the above-mentioned object is achieved by a sensor or a sensor unit and / or a bank-note processing machine, which are designed to carry out a method set forth here.

The sensor may be part of the sensor unit and / or the bank-note processing machine.

The local excitation of the value document, in particular of the feature substance, preferably takes place with the aid of an excitation radiation. The feature substance preferably has a luminescent substance or a Raman-active substance or a substance which can be detected by surface-enhanced Raman spectroscopy (SERS). Furthermore, the feature substance may have magnetic properties. In addition or as an alternative to listing, however, every feature substance with machine-testable properties is conceivable. In the present case, the feature substance can also be regarded as a marker.

In one embodiment, the excitation radiation can be spectrally narrowband, broadband, or a superimposition of different narrowband and / or broadband radiation components.

In one embodiment, the value document is illuminated with a test radiation for checking the presence of a document substrate at the respective measuring point, for size measurement of the value document and / or for measuring a remission value. In order to detect feature intensities and / or remission values, the excitation radiation and / or the test radiation are measured in a spatially resolved manner in accordance with an embodiment of the invention.

The value documents to be checked for completeness in the context of this invention are equipped with at least one machine-readable feature substance which has been introduced or applied at least along a track in the direction of movement of the value document. The machine-readable feature preferably comprises at least one luminescent marker (luminescent substance), more preferably inorganic phosphors based on rare earth or transition metal ion-doped host lattices.

In this case, the machine-readable feature substance is preferably distributed homogeneously over the surface of the value document or homogeneously into the volume of the document

Value document (paper or polymer) introduced. Alternatively, it may be printed over the entire surface or in subregions of the value document, but at least along a track over the length, or in the case of a transverse transport over the width of the document. In the case of a luminescent feature substance, this can either be used at shorter wavelength (anti-Stokes wavelength).

Luminescence or Upconverter) and / or at a longer wavelength than the excitation wavelength (Stokes luminescence) emit. Anti-Stokes emitters are not preferred since they typically have a significantly lower brightness.

It is preferably a document of value which has a Stokes luminescent substance introduced into the paper volume in homogeneous distribution during papermaking via the pulp. In a preferred variant, at least two independently measurable feature substances are present in the value document, which are either spatially or equally distributed. This can be, for example, two independent feature substances introduced into the substrate of the value document (polymer or paper). Alternatively, a feature substance may be present in the substrate and a second feature substance may be printed.

The general structure of the sensor is described as follows. To carry out the method according to the invention, a suitable sensor for the machine-readable feature is required. In the case of a luminescent feature or a SERS feature, it is typically designed for the spectrally resolved detection of the feature substance. The feature sensor is preferably installed in a machine for automated checking or sorting of value documents, in particular a bank note processing machine. This transports the value documents to be tested linearly through the detection range of the sensor device in a predetermined transport direction.

The feature sensor may include a luminescence sensor. The luminescence sensor is preferably used as a detection device for the spectrally resolved detection of the luminescence radiation in at least one

formed spectral detection range and supplies detection signals which reproduce at least one, in particular spectral, property of the detected luminescence radiation. The spectral resolution can be determined either by dispersive elements, such. B. diffraction gratings in reflection or transmission or by suitable filters in front of the respective detector elements can be achieved. The spectral resolution of the detector has at least two wavelength channels, preferably> 4, more preferably> 8 different wavelength channels. In order to excite the luminescence radiation emanating from the value document, the sensor illuminates it in a detection area with an excitation radiation. This is tuned to the luminescent substance used to mark the value document and is in the optical range, ie in the UV, VIS or IR spectral range. The excitation radiation can be spectrally narrowband, broadband or a superposition of different narrowband and / or broadband radiation components. The luminescence sensor is preferably additionally equipped with a reflectance sensor. Here it illuminates the document of value in addition to the excitation radiation with a test radiation. This serves to check for the presence of the document substrate at the momentarily illuminated location or the size measurement of the value document and / or the measurement of the remission. In a variant, the test radiation preferably has a spectral distribution which coincides with the spectral detection range of the

Detection device at least partially or completely overlapped. In this case, the remission of the value document can be determined directly without the need for a separate detector.

In an alternative variant, in addition to the illumination device for the test radiation and a separate Prüfstrahlungsdetektor, together with possibly required illumination, collimation and / or imaging optics, available, with the luminescence in addition to spatially resolved remission is measured and the geometric imaging properties of the two Detection channels each associated with the associated measurement locations of the luminescence. The illumination surfaces of the excitation radiation and of the test radiation preferably overlap spatially in the detection range of the sensor or are largely identical, so that the spatial assignment of the measured values can take place directly.

Furthermore, the sensor has a control and evaluation device which controls the emission of excitation radiation or test radiation and receives the signals of the detection device (s), processes them and evaluates them with regard to authenticity or completeness.

The test radiation and the excitation radiation are irradiated with suitable light sources, e.g. Incandescent lamps, flash lamps, LEDs or laser diodes, in particular edge emitter or VCSEL generated. Possibly. In addition, filters or phosphor converters are required to produce the desired spectra. The remission is typically determined in the visible spectral range either in a broad or alternatively narrowly narrowed wavelength range. Alternatively, the remission may also be in a non-visible spectral region such as e.g. be determined in UV or NIR.

In a first step, the luminescence signal obtained during each measurement cycle can be evaluated locally for each individual measurement point. This can include the evaluation of a spectral distribution, for example, after an offset correction or background correction, whereby any signal contributions introduced by scattered light or by the amplifier or evaluation electronics are eliminated. The necessary correction parameters can either be preset or determined dynamically using suitable dark measurements. These can be carried out, for example, if no value document is currently in the detection range of the sensor and / or one (or more) measuring point is "sacrificed" on the value document itself and, instead, a dark measurement is carried out without excitation and without test illumination. Optionally, the measured spectra can be normalized with pre-set or separately measured illumination intensities or remission values etc. measured on special calibration substrates.

Furthermore, the local authenticity of the value document is checked on the basis of the measured luminescence signal. This can be done on the basis of the spectral distribution or additionally evaluate the arrival and / or decay behavior. In this case, at least one intensity value is calculated which represents a measure of the local luminescence intensity and together with the measuring location, i. h., z. B. the x-y coordinate formed from tracking and transport position is stored.

Likewise, the remission value is determined in the case of narrow-band test illumination, or the remission values of several spectral channels in the case of spectrally resolved remission measurement. The determined remission value, together with the measuring location, i. the transport position, saved.

In general, the further evaluation is carried out in two stages: First, the feature values are classified into locations with suprathreshold feature intensity (real or analog) and locations with subliminal feature intensity (spurious or analog). Subsequently, the completeness is determined on the basis of the number and distribution of spatially resolved feature intensities classified as spurious. Case 1 describes an evaluation without denomination information and without remission measurement. In this most difficult case, the sensor merely measures the machine-readable feature without having any further information about the present value document or about its true or apparent size. Thus, only the measurement data distribution of the machine-readable feature is available for the completeness evaluation. Nevertheless, on the basis of this limited information, a well-founded statement of completeness can be made. In reality, counterfeit or incomplete value documents are relatively common in which narrow vertical structures or stripes have been cut out. To be able to recognize this efficiently is one

Rating regarding a column integrity useful. Here, the number of subliminal pixels is determined column by column and compared with a threshold value. If this threshold (of eg 2 or 3) is exceeded in a column, the value document is rejected as incomplete. As a result, these types of counterfeiting are detected particularly effectively with vertically extended manipulations. Preferably, a different evaluation takes place between the edge tracks and middle tracks. This makes it possible to detect missing measuring ranges, which occur due to a tilting of the value document in the transport, and to reduce the frequency of value documents erroneously classified as incomplete. In one embodiment, for example, track completeness in the evaluation can generally be ignored. Alternatively, the edge track in the shortened form can be evaluated within the extent recognized by the remission measurement. In a preferred case, several feature substances which can be measured independently of one another are present in the value document. Of these, separate feature values are advantageously recorded and evaluated or evaluated. If a feature value is present at a measuring location, it follows immediately that the second feature substance must also be measurable at this location, taking into account the spatial distribution of the second feature substance.

In general, in this case, a combination of the convex hulls of the distributions of the two feature material measured values can be used as a measure of the geometric extent of the value document.

Thus, the case suspected of being counterfeit that an outboard track provides a seemingly longer banknote length than a track further inboard can be correctly identified. This means, in particular, that if an outside edge track n is present with a valid first measured value, then the track n, but at least the further inward adjacent track (n-1) is also fully evaluated for the second feature with respect to track completeness. The evaluation is of course always taking into account the expected target distribution of the respective feature substance. This procedure is used analogously for the top and bottom tracks.

In a preferred case, the completeness check is carried out by a machine-dependent evaluation, in which the actual geometry proportions are taken into account with regard to the transport of value documents. Depending on the machine model, the orientation of the transported value documents can be either along the lower edge or center-centered, for example. This has the consequence that when processing different denominations with different sizes (especially widths) depending on the machine different tracks may expect feature signals. Since these transport properties always remain constant, they are advantageously taken into account for the assessment of completeness and parameterized during installation of the sensor. In particular, it is defined which tracks should always be completely present (center track (s) vs. lowest track or second lowest track for consideration of a skew).

For the completeness evaluation of the value document, both the trace completeness and the area completeness are preferably evaluated and finally combined into a measure for the completeness. In this case, an identified lack of trace completeness can lead to the entire value document being recognized as incomplete, even if the area completeness is perhaps still within an accepted tolerance threshold.

A particularly reliable evaluation of the completeness is carried out under examination at the level of the pixels (pixel completeness), at the level of the measuring tracks (trace completeness) as well as at evaluation of the two-dimensional distribution of the measured values obtained (surface completeness or two-dimensional completeness).

Other features and advantages of the invention will become apparent from the following description of embodiments according to the invention and other alternative embodiments in connection with the following

Drawings that show:

Figure 1: A schematic representation of an embodiment of a method according to the invention; FIG. 2a shows a first diagram according to an embodiment for classifying at the pixel level;

FIG. 2b shows another diagram according to an embodiment for classifying at the pixel level;

FIG. 3 is a schematic representation of a characteristic curve for threshold values for classification at the pixel level; Figure 4: A schematic representation of the timing of the illumination for remission or;

FIG. 5a: a schematic representation for the classification on the pixel level with bilateral feature measurement;

FIG. 5b shows a schematic representation of a further characteristic for the classification on pixel level with double-sided feature measurement;

FIG. 6: a curve of feature intensity, remission value and a dynamically determined threshold value for classification at the pixel level;

FIG. 7: a schematic representation of remission values of a banknote to be checked;

FIG. 8 shows a schematic representation of feature intensities of a banknote to be checked; FIG. 9 shows a schematic representation of feature intensities of an incomplete banknote to be checked;

FIG. 10 a: a schematic representation of location-related distribution of classified feature intensities;

FIG. 10b: a schematic representation of location-related distribution of classified feature intensities of an incomplete banknote; FIG. 11: a representation of transmission values of a banknote;

FIG. 12: another schematic representation of a pixel-by-pixel classification; and Figure 13 is a schematic representation of a combined classification of feature values.

1 shows schematically a method sequence for testing a value document according to the invention.

In a first step Sl, a value document is provided. The value document comprises at least one machine-readable feature substance. The feature substance is arranged at at least two different locations, preferably arranged over a substantial area of the value document. Preferably, the machine-readable feature substance extends partially over the entire areal extent of the value document. In a step S2, the value document is at least locally excited, preferably with electromagnetic radiation. The stimulation can be done by irradiating the entire value document. Preferably, an area-wise, particularly preferably a punctual irradiation of the value document takes place. By means of a sensor unit is spatially resolved a feature value, in particular a feature intensity, with respect to

Machine-readable feature substance detected at several different locations of the value document (S3a). The detection usually relates to the surface portion of the document of value which has been excited by means of electromagnetic radiation, wherein preferably the excited portion has an area equal to or greater than the area or point detected.

Preferably, at substantially the same time as step 3a, a remission value is detected spatially resolved with respect to the feature values acquired in step 3a (S3b), it also being possible to detect a plurality of remission values, which relate, for example, to different wavelengths.

In a step S4, the feature values and the preferably detected remission value are evaluated spatially resolved in accordance with the steps S2, S3a and optionally S3b. In this case, the feature values are compared with expected reference signals, and a feature intensity is determined in each case for the feature values acquired spatially resolved. Preferably, a normalization of the location-related feature intensities takes place.

Based on the evaluation from step S4, a classification of the location-related feature intensities takes place in step S5. The classification is based on a lower threshold value of the feature intensities (see Fig. 2a) or a combined use of a lower and upper threshold value of the feature intensities (see Fig. 2b) or a use different threshold values of the feature intensities, in particular as a function of one or different remission values (FIG. 3).

The evaluation of a feature value and the classification of a feature intensity can be carried out independently of the time of the acquisition of further feature values. Thus, for a feature intensity, step S4 may preferably take place immediately after step S3a, and / or for one or more feature intensities, step S4 may be performed after the detection of the plurality of feature intensities according to S3a. Analogously, for a feature intensity, step S5 can preferably take place immediately after step S4 and / or step S5 can be carried out for one or more feature intensities after the evaluation of the plurality of feature intensities according to S. In step S6, based on the evaluation from step S4 or alternatively based on the classification of the feature intensities from step S5, a location-related distribution of the feature intensities is determined. From the location-related distribution, expected location-related limits of the distribution of the feature substance are derived. These location-related limits are determined either from the distribution of the classified location-related feature intensities, for example by calculating the convex hull of the suprathreshold feature intensities, or by taking into account further measured values, in particular the remission values. Subsequently, in step S7, the location-related distribution of the classified feature intensities obtained in step S5 is evaluated. The evaluation takes place, in particular, with regard to the relative position of the pixels classified above or below the threshold, and with regard to the relative position of the subliminally classified pixels relative to those determined in S6 Limits of the expected spatial distribution of the machine-readable feature substance.

Finally, based on the evaluation from step S7, a completeness measure for the entire value document is determined, which is used for authenticity assessment or, for example, for sorting decisions in one

Bank note processing machine can be used.

In the diagrams now referred to, the colors yellow are used with reference character "ge", green with reference symbol "g", black with reference symbol "s", red with reference symbol "r" and blue with reference symbol "b". All listed color specifications are only examples and are for illustrative purposes only. Of course, values or other names can be used instead of the color information.

FIGS. 2a and 2b respectively show an intensity field for a sampled pixel, wherein the threshold values for feature or remission signals used for classifying at the pixel level are entered by way of example according to one aspect of the present invention.

The classification of the pixels in real / non-genuine takes place by way of example with reference to FIG. 1 as follows. To evaluate a value document for authenticity and / or completeness, a pixel-based classification is performed.

All measurement points or pixels which have remission values above a certain threshold Ri in the remission channel must also provide a sufficient feature intensity in the feature signal in order to be recognized as a true part of the value document. The feature intensity must therefore higher than a lower threshold of feature intensity Mmin. This classification of all pixels using fixed threshold values can be illustrated graphically with reference to the 4-field table according to FIG. 2a. Figure 2b shows threshold values for feature level or reflectance values for pixel level classification in a modified 4 quadrant scheme using a lower threshold Mmin, Ri and an upper threshold M _ma x. All pixels which are sufficiently bright (ie remission value R> remission threshold Ri) and deliver a sufficiently intense feature signal (feature intensity M> minimum feature intensity Mmin (lower threshold of feature intensity)) are rated as "green". Too dark pixels (R <Ri), as z. B. through holes in

Value documents are classified "black" while existing areas of the value document (R> Ri), i. h., a sufficiently high remission value is detected, classified without adequate feature signal as suspected of being forged, in particular as snippet forgery, "red". If there are areas with inadequate remission but with sufficient feature intensity, these are classified as a "feature excess" "yellow". This can be z. For example, in the case of heavy soiling (with special spectral behavior of the illuminated surfaces) or in window areas with an invisible feature.

Furthermore, according to FIG. 2b, an upper threshold for the expected feature intensity Mmax is used. Here then all areas can be classified with an excess of feature signal "yellow". The combined evaluation of remission and feature intensity at the pixel level allows in any case a simple consideration of otherwise problematic situations, such. As a run-up (ie y-offset) or skew one Value document in the processing machine as a result of a transport fault.

In a further embodiment, the pixel-level remission signal is used to normalize the feature signal (only in the linear region) for the purpose of fouling or overprint correction. Likewise, edge effects are taken into account if the value document edge only partially overlaps with the measurement pixels and therefore reduced feature and remission intensities are recognized.

Alternatively, advantageously, the threshold for the feature intensity required for authentication can be adapted pixel-by-pixel on the basis of the measured remission signal. Here, a characteristic curve or a characteristic map for the authenticity detection is defined, as shown in FIG.

FIG. 3 shows a characteristic curve for the threshold values for classification on the pixel level. The presence of a document is detected for remission values R above a reflectance threshold Ri. This threshold can be set uniformly for all tracks, or, preferably, for each track, individually parameterized using reference measured values for white or black samples.

If a very dark area is registered on the document, a reduced threshold for the feature intensity M is also used (Mi> M). If correspondingly lighter areas (Ri <R <R ₂ ) are present, the required feature intensity threshold is preferably increased correspondingly between Mi and M ₄ . At particularly strong reflecting sites (R> R ₂ ) it can be assumed that there is no normal security substrate but a metallic reflector such. A hologram, Safety stripes or similar. Since these are typically opaque to optical radiation, the threshold value for the feature signal is correspondingly reduced to M3, since the covered areas under

Circumstances can only deliver a greatly reduced signal contribution. If the spatial resolution of the feature sensor is not significantly higher than the dimensions of the opaque structures, masking will not be digital but will mostly occur partially. This is due to a gradual reduction of the feature threshold between M ₄ and M3 in the range

R2 <R <R3 taken into account. In the sense of a strong detection of forgeries, a minimum of feature signal M2 can be demanded even with very high remission values R> R3. For a particularly strict evaluation M ₂ = M3 can also be chosen. In these classification variants, a hologram strip is marked in "red". Alternatively, M2 can also be parameterized to very low values, which results in a classification of reflective hologram strips as "green".

At the edge of the value document, only a partial overlap between the document of value and the measuring pixel causes random "red" pixels, which must be treated or tolerated separately in the further evaluation. Alternatively, the formation of these red boundary pixels can be prevented by a suitable parameterization of the threshold characteristic curve for Ri or Mi. In this case, M ₁ (relative to the maximum intensity) is set lower than R ₁ , so that the situation can not occur due to the purely geometric loss of intensity, which affects both remission and feature intensity equally, although R> Ri but already M < Wed is.

If several independently measurable feature substances are present, the characteristic measured values can of course be evaluated both individually and in combination, as described in the case without remission measurement. Pixel completeness:

The first check for completeness is now made on a pixel basis: Within the recognized area of the value document, the number of measuring points or pixels classified as "red" must not exceed a certain threshold. In the strictest interpretation with the threshold 0, this means that not a single measurement location with insufficient feature intensity may be present, so that the value document is recognized as complete. In other variants, individual "red" pixels can be tolerated.

Here again, the ratio of the number of all green pixels relative to the number of all pixels within the extent of the value document can be formed and checked against a minimum threshold. This corresponds to an area percentage or the area-related degree of completeness.

Track completeness:

The trace lengths determined from the reflectance measurements are each used as a yardstick for evaluating the trace completeness. To calculate the metric for the track integrity, the number of "green" classified pixels in that track is divided by the number of all pixels within that track length. Alternatively, one obtains a somewhat sharper test criterion when calculating the measure of the

Trace integrity, the number of "green" classified pixels in that track is divided by the number of pixels corresponding to the maximum length of the value document. Another test criterion is the number of adjacent "red" pixels within the length of the value document and within one track. If this exceeds a defined threshold, the track is counted as incomplete. For the parameterization of this threshold, it makes sense to use the maximum width of true "red" areas of real documents of value, such as: B. the maximum extent of hologram patches o. Ä. Considered.

Analogously to the procedure described above for determining the trace integrity without remission measurement, here too measurement tracks can be evaluated in a peripheral position different from center tracks, although the corresponding position uncertainties are much lower here due to the remission measurement.

Two-dimensional completeness:

In the preferred case that the sensor has a plurality of measuring tracks, the two-dimensional distribution of the feature intensity or the two-dimensional distribution of the classified pixels is also evaluated here as described above.

By means of the convex hull around the pixels with suprathreshold remission, holes or opaque patches within the value document can be located. Here, the occurrence of larger holes is specifically examined. For this purpose, "red" subliminal subpixels are searched for within the extent of the value document determined by the convex hull, and two-dimensional, contiguous areas are counted and identified / marked. Lying z. B. more than 2, 3, 5, ... (resolution-dependent) before contiguous red pixels, so a potentially missing area is detected. Subsequently, the location and geometric extent of the analyzed "red" areas and patterns occurring in a known manner such. B. a transparent window or a metallic hologram strip aligned. In particular, the shape, maximum width and relative position to the edges or corners of the value document is checked for plausibility and classified in case of deviations as "incomplete".

Here, too, an evaluation with regard to column completeness can be made for the efficient recognition of forgeries or incomplete value documents with vertical manipulation structures. Here, the number of red pixels is determined column by column and compared against a threshold value. If this threshold (of eg 2 or 3) is exceeded in a column, the value document is rejected as incomplete.

For those counterfeiting classes in which portions of the true value document in the marginal area are replaced by e.g. A photocopy has been replaced by the combined evaluation of the remission and the feature intensity a real qualitative advantage: By accurately determining the actual extent of the value document these fakes can now be reliably detected. In particular, a targeted check for the presence of "red" classified edge columns (which were determined by counting the red pixels in the column direction) can be made. In this case, the outermost two columns are preferably evaluated in order not to over-evaluate or incorrectly evaluate the red edge pixels occurring randomly due to edge effects.

In an embodiment with greatly different resolving power of the measurement in the (x) track direction and the y direction (number of tracks), this is taken into account by specifically selecting "red" neighboring pixels in line direction. tion and multiple pixels in this direction are evaluated particularly seriously. In particular, the maximum occurring width of a hologram strip (or similar security features such as metal paint) may be considered by classifying value documents having a larger number of red pixels in the higher resolution measurement direction than a defined threshold directly as incomplete.

Double-sided Measurement In particularly preferred variants, the authenticity sensor comprises two

Partial sensors that allow a two-sided measurement of the feature intensity on each document of value. In this case, at least on one side-or more preferably on both sides-a remission channel is also available, with which the (track) length as well as the exact position and orientation of the value document are determined.

In one embodiment, the two part sensors are centrally controlled in order to synchronize the timing of the excitation or measured value recording for both partial sensors. Alternatively, two separate front and backside sensors are used, which are synchronized in one master / slave configuration by one of the two sensors ("master"). For example, this master sensor sets the operating mode and outputs to be maintained time delays for the measuring pulses and / or measured value recording after a trigger signal before.

Furthermore, different sensor architectures for the master or slave sensor can preferably be used. Thus, for example, one of the sensors with a more complex measurement technique than the other sensor be equipped and check the feature values with a higher accuracy or a higher spectral resolution.

The two partial measurements of front and back are then evaluated in combination. The measurement data are assigned to the respective measurement locations on the value document, the location-related data tuples are formed and evaluated (remission, feature, feature 2) or (remission 1, remission 2, feature 2, feature 2). Preferably, the position or timing of the two measurements (front, back) is coordinated so that the value document at the same pixel positions on the front and back are measured. Most preferably, the measurement takes place (almost) simultaneously, i. that a measuring point at a location of the value document is detected almost simultaneously from the front and from the back.

In addition to the simpler and clearer evaluation of the measured values obtained in this way, this has the advantage that a mostly unavoidable crosstalk between front and rear side measurement does not lead to artifacts and interference signals, but instead amplifies the feature signal to be measured.

In this case, the illumination of the first partial sensor can advantageously also be exploited for a transmission measurement with the detector part of the second partial sensor if the two illumination light pulses have a small time offset so that the transmission signal can be recorded separately from the remission signal 2. This temporal sequence of the light pulses or detections is shown schematically in FIG. In this case, transmission, remission, remission2 are then available for each measuring location as well as feature, feature2 as database for the completeness evaluation. This allows the complete completeness evaluation even with existing opaque (metallic) or transparent (window) security features, which may otherwise hinder the completeness check of certain parts of the value document.

The illumination for the remission measurement (alternatively: feature measurement) of the front and the back are slightly offset in time, so that detector 2 can determine the transmitted portion of the illumination 1 independently and undisturbed by the illumination 2, as shown in FIG.

In the simplest case, in the evaluation, the sum (or the mean or the maximum) of feature and feature 2 is formed at each measurement location and then classified and evaluated according to the procedures described above.

A more accurate rating is achieved when individual thresholds for feature and feature2 are applied. These can depend both on remission and on the other characteristic value. In place of the characteristic curve described above for the pixel-wise red / green

Evaluation then occurs here a corresponding map. This can be adapted / parameterized precisely to the typical optical effects occurring on real value documents. FIGS. 5a and 5b show a characteristic diagram for the threshold values for classification on the pixel level with bilateral feature measurement. In the figure 5a is a classification based on static thresholds of

Feature value (Wed, min, With the map in Figure 5b is a Classification taking into account change effect effects, such. B. reflection on unilaterally applied metallic surface structures.

If, for example, a banknote (reflective, therefore opaque) metallic strip is applied to a banknote on one side, then it is to be expected that the one-sided feature value is very small, but the expected feature value 2 is due to the reflections that occur with respect to the immediate environment (or opposite the average over the entire banknote) is increased. This can be mapped by the corresponding parameterization of the threshold characteristic field. Conversely, when overprinting on the side Bl with black, spectrally broadband absorbing (soot) color, the remission value and feature value is low, while feature value 2 is at a normal level. The parameterization of the classifier is advantageously dependent on the location, i. H. z. B. relative to the leading edge, relative to the corners, or concrete position within the convex hull, etc. This allows a correct treatment of absorptive and reflective interference depending on the (depending on the lags and stöck- depending) in these areas possibly occurring effects. In both cases, in spite of the insufficient feature intensity on one side, the two-sided feature measurement can reliably evaluate the corresponding range as true.

This allows complete proof of completeness regardless of the BN design, even in difficult situations with (unilaterally occurring) covers / shadowing by opaque elements such as e.g. Aluminum coated hologram stripes. This can also be areas of

Value document reliably checked for completeness / authenticity, which can not be evaluated in one-sided measurement. In a preferred variant, the complete existing data set (transmission, remission, remission, feature intensity, feature intensity) is combined and evaluated. In addition to areas with opaque, absorbent or reflective overlaps, in particular holes or window areas can be reliably identified via the transmission signal and their position and extent compared to the values permissible for genuine value documents checked. Further embodiments will be described below.

Example 1: Here a spectrally resolving Emspur luminescence sensor with remission measurement is used for the completeness test. The sensor is operated on a banknote processing machine at 11 m / s transport speed and used for authenticity and completeness testing of banknotes with a luminescence marker which is matched to the luminescence sensor and inserted into the paper. The banknotes have a reflective hologram strip on the front in the right area.

FIG. 6 shows a feature curve (O), a remission curve (x) and the dynamically calculated feature threshold (dashed line) of a genuine and complete banknote. Both remission and feature intensity are significantly modulated. Nevertheless, the completeness can be correctly determined by applying a reflection-dependent threshold in the classification of the feature intensity.

Example 2: Here, a spectrally resolving 11-track luminescence sensor with reflectance measurement is used for the completeness test. The sensor is placed on a banknote processing machine at 11 m / s trans- operated for speed and authenticity and completeness of banknotes with a luminescence marker introduced into the paper. The banknotes have a reflective hologram strip on the front in the right-hand area and a transparent window in the left-hand area.

FIG. 7 shows a representation of the measured remission values of the banknote. High remission occurs especially in the area of the reflective hologram strip, while very low remission is present in the transparent window.

FIG. 8 shows a representation of the feature intensity of the banknote. White corresponds to high intensity, while black corresponds to low values. In the area of the window (left) and the hologram strip (right) only very low feature intensity is detectable.

Example 3:

For comparison, appropriately prepared chip counterfeits were measured with about 10% counterfeit share.

FIG. 9 shows a representation of the feature intensity of an incomplete banknote with a diagonally inserted stripe of a copy without a feature.

FIG. 10a shows a pixel-wise classification of the banknote (FIGS. 7-8) with a dynamic threshold. The low feature intensity in the area of the hologram strip could be taken into account by the dynamic threshold, while the missing feature intensity in the window area could be taken into account. gels remission signal is marked in red. (0 = black, 1 = red, 2 = yellow, 3 = green)

FIG. 10b shows a pixel-by-pixel classification of the incomplete bank note (FIG. 9) with a dynamic threshold. The small feature intensity in the area of the hologram strip could be corrected by the dynamic threshold, while the missing feature intensity in the window area is marked in red for lack of remission signal. The missing feature area is recognized correctly and marked in red. (0 = black, 1 = red, 2 = yellow, 3 = green)

Example 4:

The banknote of FIGS. 7-8 was measured again with a sensor structure with double-sided measurement. Characteristics (front), feature2 (rear), remissionl (front), remission2 (rear) and transmission were measured.

FIG. 11 shows transmission data of the banknote For classifying the measurement pixels, the front and rear sides were classified separately with a dynamic feature threshold and then separated according to the following assignment of the class assignments determined on the front (classification) and back (classification) to a total classification for each pixel combined, as shown in Figure 12 is shown.

The window area was subsequently identified on the basis of the high transmission> 85 and accordingly classified as "window" (4). FIG. 12 shows a pixel-by-pixel classification of the measured data of the complete test banknote with dynamic threshold and transmission measurement determined on both sides. (0 = black, 1 = red, 2 = yellow, 3 = green, 4 = light blue) Here, despite the metrologically difficult architecture of the banknote with metallically reflecting and transparent window areas, all areas are reliably checked for local authenticity and the completeness is assessed correctly.

FIG. 13 diagrammatically shows a combination of feature values classified on both sides, according to which evaluation of the value document or of the banknote likewise takes place for authenticity and / or completeness.

Claims

Patent claims

1. A method for checking the completeness and / or authenticity of value documents, wherein at least one value document comprises at least one machine-readable feature substance in at least two locations, with the steps:

at least local stimulation of the value document (S2)

location-resolved detection of a feature intensity (M) with respect to the machine-readable feature substance at a plurality of different locations of the value document (S3a);

Location-based classification of location-related feature intensities using a threshold value (S5);

Determining location-related boundaries of an expected location distribution of the machine-readable feature substance (S6); and

Evaluate a location-based distribution of the classified feature intensities (S7).

2. The method according to claim 1, characterized in that the classification (S5) of the location-related feature intensities (M) takes place on the basis of location-dependent threshold values.

3. The method according to any one of the preceding claims, characterized by the step of the spatially resolved detection of remission values (R) (S3b) at a plurality of different locations of the value document.

4. The method according to claim 3, characterized in that the threshold value is formed as a location-dependent threshold, the one from the on specific location remission value (R) dependent characteristic is determined.

5. The method according to claim 3 or 4, wherein the measurement locations of the remission values (R) overlap with the measurement locations of the feature intensities (M) and are preferably identical.

6. The method according to claim 1, characterized by the step of calculating a track completeness by a comparison of the number of measurement locations with suprathreshold feature intensity (M) with the number of detected measurement locations within a convex envelope of the measurement sites with suprathreshold feature intensity (M) or, if applicable within a convex hull of the sites with suprathreshold remission value (R).

7. Method according to one of the preceding claims, characterized by the step of checking a two-dimensional distribution of the classified feature intensities (M) relative to a convex envelope of the measurement sites with suprathreshold feature intensity (M) or if remission values (R) were detected relative to a two-dimensional one Distribution of the measured values of the remission measurement, wherein preferably the checking of the two-dimensional distribution of the classified measurement locations comprises a calculation of a column completion.

8. The method according to any one of the preceding claims, characterized in that the feature intensities (M) along at least one measuring track are detected on the document of value.

9. The method according to any one of the preceding claims, characterized in that the spatially resolved detection of the feature intensities (M) (S2) with respect to the machine-readable feature substance, the measurement of a spectral luminescence intensity of a luminescent substance and / or the spectral measurement of a Raman band of a Raman-active substance and / or a substance detectable by surface-enhanced Raman spectroscopy (SERS) and / or the spectral measurement of an absorption band of an absorbing substance in the infrared spectral range and / or the measurement of the magnetic properties of a ferromagnetic substance.

10. The method according to any one of the preceding claims, characterized in that based on at least one feature value (M) a local authenticity of the value document is checked.

11. The method according to any one of the preceding claims, characterized in that a number and a spatial distribution of subliminal classified measurement locations is compared with reference values.

12. The method according to any one of the preceding claims, characterized in that the value document during the measurement at a speed of 1-13 m / s, preferably 6-12 m / s is moved.

13. Method according to one of the preceding claims, characterized in that the spatially resolved detection of the feature intensities (M) of the machine-readable feature substance and / or optionally the remission values (R) on the front and back of the document of value, in particular at the same, opposite locations of Front and back, done.

14. The method according to claim 13, characterized in that the location-dependent threshold values are determined by a characteristic curve which depends on the feature intensity (M) of the value document determined at the respective location.

15. Method according to one of the preceding claims, characterized in that transmission values of the value document are recorded with local resolution.

16. The method according to claim 15, characterized in that the transmission measurement is carried out by a time-shifted illumination in the context of remission measurements on the front and back and / or by a time-shifted illumination in the context of the measurement of feature values on the front and back.

17. The method according to any one of the preceding claims, characterized in that at a plurality of measurement locations in each case a combined classification taking into account the data points associated with data tuples is carried out, the data tuples at least one feature intensity (M) and at least one of the following components comprise: a further feature intensity ( M), a remission value (R), and / or a transmission value.

18. Sensor, for detecting a feature intensity (M) and / or a feature value, designed to carry out a method according to one of claims 1 to 17.

19. sensor unit having a sensor, wherein the sensor is designed to detect at least one feature intensity (M), a feature value, a remission value (R) and / or a transmission value, in particular according to claim 18, and / or wherein the sensor unit is embodied, to carry out a method according to one of claims 1 to 17.

20. Banknote processing machine with a sensor unit according to claim 19, a sensor according to claim 18 and / or suitable for carrying out a method according to one of claims 1 to 17.