METHOD FOR CLASSIFYING A BANKNOTE, CLASSIFICATION DEVICE, METHOD FOR PRODUCING A CLASSIFIER AND USE FOR THE CLASSIFICATION OF A BANKNOTE

Abstract

A method for classifying a banknote has at least one luminescent feature, in which the following steps are carried out: a) exciting the luminescent feature; b) recording at least a first intensity value of the excited luminescent feature at a first time and a second intensity value at a second time that is different from the first time; c) determining a decay time of the luminescent feature by approximate integration over the intensity values between the first time and the second time; d) comparing the determined decay time with a reference decay time of the luminescent feature; and e) classifying the banknote based on the comparison.

Claims

1.-15. (canceled)

16. A method for classifying a banknote having at least one luminescent feature, in which the following steps are carried out: a) exciting the luminescent feature; b) recording at least a first intensity value of the excited luminescent feature at a first time and a second intensity value at a second time that is different from the first time; c) determining a decay time of the luminescent feature by approximate integration over the intensity values between the first time and the second time; d) comparing the determined decay time with a reference decay time of the luminescent feature; and e) classifying the banknote based on the comparison.

17. The method according to claim 16, with at least five intensity values being recorded, and the decay time being determined based on the at least five intensity values.

18. The method according to claim 16, with the approximate integration being carried out with the trapezoidal rule.

19. The method according to claim 16, with any noise of the intensity values being reduced prior to the determination of the decay time.

20. The method according to claim 19, with the noise of the recorded intensity values being reduced by a moving average across multiple intensity values.

21. The method according to claim 20, with the moving average being determined across at least five intensity values.

22. The method according to claim 16, with a third intensity value of the excited luminescent feature being recorded at a third time that lies between the first time and the second time.

23. The method according to claim 16, with the intensity values being recorded temporally equidistantly.

24. The method according to claim 16, with the first intensity value being recorded only after a predetermined waiting time after the excitation of the luminescent feature.

25. The method according to claim 16, with the decay time being determined depending on a normalization value that is formed by subtracting the first intensity value from the second intensity value.

26. A classification device for a banknote having at least one luminescent feature that is adapted to carry out the following steps: a) exciting the luminescent feature; b) recording at least a first intensity value of the excited luminescent feature at a first time and a second intensity value at a second time that is different from the first time; c) determining a decay time of the luminescent feature by approximate integration over the intensity values between the first time and the second time; d) comparing the determined decay time with a reference decay time of the luminescent feature; and e) classifying the banknote based on the comparison.

27. A method for producing a classifier for classifying a banknote having at least one luminescent feature, in which the following steps are carried out: a) exciting the luminescent feature; b) recording at least a first intensity value of the excited luminescent feature at a first time and a second intensity value at a second time that is different from the first time; c) determining a model decay curve of the luminescent feature based on the intensity values; d) determining an antiderivative of the model decay curve; e) determining an approximation of the antiderivative; and f) providing the classifier for classifying the banknote, with the approximation of the antiderivative.

28. The method according to claim 27, with a single piecewise integral being calculated approximately in determining the approximation of the antiderivative.

29. A use of an approximate integration for the classification of a banknote having at least one luminescent feature.

30. The use according to claim 29, with the trapezoidal rule being used as the approximate integration.

Description

[0061] Further features of the present invention are evident from the claims, the drawings and the description of the drawings.

[0062] Exemplary embodiments of the present invention are explained in greater detail below by reference to a schematic diagram.

[0063] Here are shown:

[0064] FIG. 1 a schematic diagram of an exemplary embodiment of a classification device according to the present invention; and

[0065] FIG. 2 a schematic diagram of an exemplary embodiment of a model decay curve of a luminescent feature.

[0066] In the drawings, identical or functionally identical elements are labeled with identical reference signs.

[0067] FIG. 1 shows an exemplary embodiment of a classification device 1. The classification device 1 is formed to classify banknotes, and according to the exemplary embodiment, part of a banknote processing device 100.

[0068] The classification device 1 comprises an evaluation unit 2, for example a microprocessor, and a recording unit 3, for example a photodetector.

[0069] Also shown is a banknote 4 having a luminescent feature 5 and multiple further luminescent features 6.

[0070] Further shown is an excitation unit 7, especially an illumination unit or light flash unit, and a transport unit 8 having direction of travel 8a.

[0071] The transport unit 8 is formed to transport the banknote 4 from the excitation unit 7 to the recording unit 3.

[0072] The excitation unit 7 is formed to excite or charge the luminescent feature 5 and/or the further luminescent features 6, especially with light.

[0073] After the excitation, the luminescent feature 5 and/or the further luminescent features 6 gradually emit the charged energy in the form of luminescent light. That is, the luminescent features 5, 6 luminesce, i.e. the luminescent features 5, 6 still glow even when said features are no longer illuminated.

[0074] The luminescent light emitted by the luminescent feature 5 and/or the further luminescent features 6 after excitation is recorded by the recording unit 3. The recording unit 3 records intensity values, especially at least a first intensity value and a second intensity value, of the luminescent light emitted by the luminescent feature 5 and/or the further luminescent features 6. The recording unit 3 preferably records the intensity values at different times. In this way, interpolation values or interpolation points for the decay behavior of the luminescent feature 5 and/or the further luminescent features 6, especially a decay curve, are recorded.

[0075] The intensity values recorded by the recording unit 3 are evaluated in the evaluation unit 2.

[0076] FIG. 2 shows an exemplary embodiment of a schematic model decay curve 9. On the ordinate 11 is plotted, on a relative scale, which luminescence intensity of the luminescent feature 5 is recorded. “1.0” means that the maximum luminescence intensity of the luminescent feature 5 is recorded. “0” means that no luminescence of the luminescent feature 5 is recorded, i.e. the luminescent feature 5 is completely discharged. On the abscissa 10 is plotted the time in the unit Δt. In particular, on the abscissa 10 is plotted the ratio oft/Δt, i.e. the time since the first time or the first measurement time in multiples of the sampling time interval Δt.

[0077] Furthermore, at least a first intensity value 12 of the luminescent feature 5, which is recorded at a first time 13, and a second intensity value 14 of the luminescent feature 5, which is recorded at a second time 15, are marked in the figure.

[0078] Further shown is at least a third intensity value 16 of the luminescent feature 5, which is recorded at a third time 17, and (without a reference sign) further intensity values at further times.

[0079] The third time 17 is especially arranged between the first time 13 and the second time 15.

[0080] The times 13, 17, 15 are preferably equidistant (not depicted).

[0081] The intensity values 12, 14, 16 correspond especially to the luminescence intensity emitted by the luminescent feature 5, which is recorded or measured by the recording unit 3 at the respective time 13, 15, 17.

[0082] In the method for producing a classifier, a model decay curve 9 is determined based on the measured intensity values 12, 14, 16. The model decay curve 9 can be expressed, for example, by y(t)=a.Math.e.sup.−/τ, where y is an intensity value at time t, a is a scaling factor that is proportional to the intensity of the luminescent feature, and τ (tau) is the decay time, for example with the unit s (seconds).

[0083] Upon transitioning to the antiderivative Y(T) of the model decay curve y(t), the factor −τ occurs as a negative decay time:

Y(T):=∫.sub.0.sup.Ty(t)dt=[−τ.Math.a.Math.e.sup.−t/τ].sub.0.sup.T=−τ.Math.(y(T)−y(0))

Thus, for any time T>0:

[00001]$\begin{array}{cc}\tau =-\frac{{\int }_0^{T}y\left(t\right)\mathrm{dt}}{y\left(T\right)-y\left(0\right)}& \left(I\right)\end{array}$

applies. Said formula (I) thus enables a determination of the decay time τ based on the antiderivative ∫.sub.0.sup.T y(t)dt of the model decay curve y(t).

[0084] Now let the decay curve of a real banknote be given by the measured values or intensity values y.sub.i.sup.m=y.sup.m(t.sub.i) that were recorded at discrete times 0=t.sub.0< . . . <t.sub.n-1=T. Here, y.sup.m describes the real measurement data, y the model decay curve. Here, the measurement data y.sup.m can deviate from the model decay curve y due to noise, but also systematically. The model decay curve and the time range of the measurement are preferably chosen such that said deviation is minor, then the method enables a particularly precise determination of the decay time. But also for a significant deviation of the measurement data from the model decay curve, for example for non-monoexponential behavior of the measured luminescent feature, the method can advantageously be used and enables a more reliable classification of the banknote.

[0085] In the method for producing a classifier, the antiderivative ∫.sub.0.sup.T y(t)dt is now, for example, approximately numerically determined from the intensity values y.sup.m(t.sub.i), and said approximation of the antiderivative used for the calculation of the decay time, in the example according to formula (I). The decay time can then be determined with the approximated formula

[00002]$\tau =-\frac{{\int }_0^T{y}^m\left(t\right)\mathrm{dt}}{y_{n-1}^m-y_0^m}.$

[0086] For the special case that the measurement times 13, 17, 15 of the intensity values 12, 16, 14 are temporally equidistant with time interval Δt, according to the trapezoidal rule, for example, approximately

[00003]${\int }_0^T{y}^m\left(t\right)\mathrm{dt}=\Delta t.Math.\left(\frac{1}{2}{y}_0^m+\frac{1}{2}{y}_{n-1}^m+\underset{i=1}{\overset{n-2}{.Math.}}{y}_i^m\right)$

results, and thus the following formula for determining the decay time (τ) from the intensity values:

[00004]$\begin{array}{cc}\tau =-\frac{\Delta t.Math.\left(\frac{1}{2}{y}_0^m+\frac{1}{2}{y}_{n-1}^m+{.Math.}_{i=1}^{n-2}{y}_i^m\right)}{y_{n-1}^m-y_0^m}=\Delta t.Math.\left(\frac{{.Math.}_{i=0}^{n-2}{y_i^m}_i}{y_0^m-y_{n-1}^m}-\frac{1}{2}\right)& \left(\mathrm{II}\right)\end{array}$

[0087] Said formula (II) can be provided and used as a classifier for the classification of banknotes, especially in a method for the classification of banknotes. Alternatively, other model functions or approximations can be used for the integral so as to obtain other classifiers.

[0088] With sufficiently many measurement times, the method for the classification of banknotes can be made even more insensitive to noise by smoothing the intensity values 12, 14, 16 prior to the determination of the integral or prior to the classification, and then applying the above formula (II) to the smoothed decay curve or the smoothed intensity values, respectively.

[0089] Alternatively, the smoothing can be taken into account in the method for producing a classifier, and the classifier produced in such a way that it comprises a smoothing of the intensity values.

[0090] Through the method for the classification of banknotes using a classifier that comprises an approximate integration of the intensity values, the decay time (τ) can be determined more quickly and more precisely. As a result, the banknote 4 can be classified more precisely and more reliably. The counterfeit security of the banknote 4 is increased.