DK144605B - APPARATUS FOR COMPARATING THE SPECTRAL REVIEW OF COLORED SURFACES - Google Patents

Description

1 UA 605

The present invention relates to an apparatus for comparing the spectral remission of colored surfaces, in particular for the authenticity of banknotes and other securities, with a measuring apparatus for providing multiple measurement values, each comprising only a narrow spectral range, at the same location on a sample object and an original and with an evaluation apparatus for comparing the sample measurement values with the reference measurement values given from the original.

As is well known, the human eye is unable to ascertain which spectral colors a color stimulus is composed of. It is therefore possible that colors of different spectral composition give exactly the same color sense. The reproduction of a particular color is relatively easy, as long as only one physical impression is required. It is much more difficult to reproduce a color with exactly the same spectral distribution. Therefore, when comparing two colors, for example when verifying banknotes authenticity, a complete spectral analysis constitutes a very hard sample.

Apparatus are known for comparing the spectral remission of colored surfaces. However, they are costly and require the operation and assessment of the measurement results by qualified professionals. Such an apparatus is known from United States Patent No. 3,48o,785.

US Patent No. 3,360,653 also discloses a test apparatus by which a number of transmission measurements can be performed by means of a number of light sources and photocells. To compensate for variations in the brightness of the light sources, a photocell is available. This reference photocell is placed in a control circuit which in turn determines the supply voltage to the photocells. However, with this known apparatus, the ratio of the output voltages of the photocells or the reference photocell is not determined. Also, no quotients are provided and therefore no distribution of quotient values can occur. Instead, each measurement value is tested to determine whether it falls within given limits or not.

An apparatus is also known which automatically performs a spectral analysis of a document and makes a yes / no choice. This simultaneously illuminates the same spot on the sample object and an original with a light source with a narrow spectral range, the light reflected from the sample object and from the original is measured by two light sensors and the two measurement values are compared with each other. In addition, the same measurement is repeated within two further narrow spectral regions. The document is judged to be genuine if the difference between the sample measurement value obtained from the sample object and the reference measurement value received from the original does not exceed a predetermined value for all three comparison measurements.

Since soiled specimens absorb significantly more light than new clean documents, the tolerance limits used for the known device in the known apparatus must be chosen relatively wide, which adversely affects the reliability of the specimen. Furthermore, there is a risk that the measured value may be distorted by aging phenomena at the light sources or the two light sensors.

The object of the invention is to provide an automatically functioning and simple yet reliable comparator where dirt deposits on the sample object cannot lead to false indications.

According to the invention, this is achieved by the fact that the assessment apparatus comprises a divisional coupling for forming the quotients between the individual sample measurement values and the corresponding reference measurement values as well as a coupling for determining the distribution of the quotients.

By forming the quotient between the individual test measurement values and the corresponding reference measurement values, you get the advantage that the presence of dirt on the test object cannot give a false measurement result. Also, aging phenomena at the measuring apparatus, such as blackening of its light source or changing the characteristics of some reading means, can not provide any error indication, as long as these aging phenomena do not cause a divergent composition of the spectral sensitivities of the sample object or the original.

In one embodiment of the apparatus according to the invention, there are two light channels emanating from a light source which lead to a single measurement sensor in such a way that alternately a sample measurement value and an associated reference measurement value are generated. This makes it possible to use a particularly simple coupling to provide the quotients between the individual sample values and associated reference values.

3 144605

The invention will now be explained in more detail with reference to the drawing, in which fig. 1 shows a principle diagram of an embodiment of the apparatus according to the invention for comparing the spectral remission of colored surfaces and FIG. 2 is a coupling diagram for a coupling to form the quotients between individual sample values and associated reference values.

In the embodiment of FIG. 1 is shown a light source 1 whose spectrum comprises the entire wavelength range to be analyzed, which sheds light on a moving diffusion filter (transition filter) 4 through a lens 2 and an aperture 3. The filter 4 is covered by a mask 5 which has several apertures 6. Corresponding to the instantaneous position of the filter, which is moved continuously or incrementally in the direction of the arrow during the measurement, light falls through one of the apertures 6 on one of two light guides 7,8. The light guide 7 directs the light beam to a selected location on an original 9, and the light guide 8 to a corresponding location on a sample object 1o. Two additional light conductors 11,12 direct a portion of the light transmitted from the original 9 and from the sample object lo to a single measuring sensor 13, respectively. The measuring sensor 13 is connected to the input 14 of a divisional coupling 15, whose output 16 is connected to a maximum value bearing 17.

To synchronize the division coupling 15 and to control the storage 17, a light sensor 19 is connected which is connected to a control coupling 18 and which receives synchronization light pulses over a light conductor 20 depending on the movement of the filter 4. An output 21 from the control coupling 18 is connected to a synchronization input 22 on the division coupling 15> and a further output 23 is connected to a control input 24 on the maximum value storage 17 ·

After storage 17, n window comparators 25-29 are connected, each corresponding to a predetermined class of output sizes from the maximum value store and emitting a signal when said output size falls within that class. With 3o-33 are designated (n - 1) pulse counters. Each of these pulse counters is connected to the two outputs of two adjacent class window comparators. The outputs of the pulse counters 30-33 are connected to a common OR gate 34.

The apparatus described works as follows.

As the filter 4 moves in the direction of the arrow, the first of the aperture apertures 4 first reaches the light channel 7 leading to the original 9. This is illuminated by a light beam of medium wavelength Λ In addition, through the same aperture aperture, light on the sample object lo. The same sequence is repeated with the wavelengths X2, \ ^ ··· · At the measurement sensor 13, the signal sequence Ir (X -, ^, I (X- ^, Ιρ (X2) a Ip (X2) · * · * where ^ means ~ where a The reference measurement value given from the original 9 and the sample measured value given by the sample l. These measured values correspond to the remission from the original and the sample object in the spectral range concerned.

The division coupling 15 forms, of each measurement value pair, the quotient, V * i> = k · ϊ ^ λίτ where k is a constant, and represents the current mean wavelength of the corresponding spectral range. In the quotient formation, aging phenomena at the light source 1 and the measurement sensor 13 cannot affect the measurement result.

When the original and sample object exactly correspond to each other over the entire wavelength range included in the measurement, for q the constant value q (X) = k is obtained.

The dirt on a banknote is usually a gray filter, ie. it does not absorb selectively. Therefore, the test of a genuine but soiled banknote with the described device is given q (λ) = k · &, where «- means the absorbency of the dirt.

Thus, as a criterion of authenticity, the quotient value itself does not apply, but the constant of the successive quotients q (\.). Since real sample objects in practice also show differences in relation to the original used, and the soiling does not always act as an ideal gray filter, real sample objects also give a certain range of the quotients provided. With the aid of the apparatus switched on after the division coupling 15, it is tested whether this spreading area is within predetermined limits.

The output signals from the division coupling 15 are fed to the maximum value memory 17. The maximum values of these due to the movement of the filter 4's non-square signals are stored in the maximum value memory 17 until the immediately following signal appears and is fed to the window comparators 25-29. Each of these comparators emits an impulse when its input signal corresponding to the quotient q falls within that of this class. The classes are chosen e.g. as follows: 5 144605

Yindu comparator q 25 0.28 ... 0.37 26 0.37 ... 0.50 27 0.50 ... 0.67 28 0.67 ... 0.90 29 0.90 ... 1 , 20

The width of a class is preferably proportional to the mean of that class.

The pulses delivered from the window comparators 25-29 are counted in the pulse counters 30 ~ 33 in such a way that each pulse counter counts the number of quotients falling in two adjacent classes.

For a sample object not similar to the original, all the q values fall within the class given by the window comparator 29, so that after completion of measurement in pulse counter 33, one pulse number corresponding to the total number of spectral single measurements is stored. A true slightly soiled sample object yields smaller q values whose number is counted by one of the pulse counters 3o-32. The test applies as passed when, for example, 9o $ of all quotients falls within the same or within two neighboring classes, ie. when one of the pulse counters 3o ~ 33 after completion of the measurement has reached a predicted pulse number. In this case, at the output of the pulse counter in question, a signal opens which opens the OR gate 34, whereby the yes selection is displayed in a manner not shown in detail.

For a sample object whose spectral remission does not match the spectral remission of the original, a corresponding distribution of the q values is obtained so that none of the pulse counters 300-33 reach the counter count required for the yes selection. Finally, a very heavily soiled sample object, which gives constant, but below the given classes, q values, is also unable to provide a yes choice. This also applies to a sample object with excessive q values.

The apparatus described can be varied in several ways. As a filter device with several filter regions with different narrow spectral range, instead of a masked diffusion filter, a rotating system with discrete single filters can be used. Preferably, a drum is suitable on whose outer surface the filters are arranged and in whose interior a light source is arranged. Instead of light conductors, a mirror system can also be used which alternately sheds light on the original and on the sample object over the filter device and deflects it from these remitted lights to the measuring sensor.

6 144605

Since a sample measurement value and an associated reference measurement value are not contemporaneous, but appear consecutively over time, the use of a particularly simple division method is made possible. For example, the so-called dual integration method can be used, in that a capacitor is charged for a constant time with a current proportional to the first measurement size and discharged again with a current proportional to the second measurement size. The discharge time then corresponds to the quotient between the two measurement values. A further advantageous possibility is apparent from FIG. 2nd

The input 14 of the FIG. 2, across a resistor 35 is connected to a first input 36 of a differential amplifier 37 / car output 16 of the coupling device and to a grounded controllable resistor 38. A second input 39 of the differential amplifier 37 is connected to a reference voltage UR. From the output of the differential amplifier 37, a control current circuit across a diode 4o leads to the control electrode 41 in the controllable resistor 38. A switch 42, whose control electrode forms the synchronization input 22 for the division coupling, is parallel to a storage capacitor 43 between the control electrode 41 and ground.

Resistors 35 and 38 form a regulated voltage divider.

With discharged capacitor 43, the resistance 38 is very high. When a signal Ir appears on the input 14, the capacitor 43 is charged until the voltage Ug on the input 36 of the differential amplifier 37 reaches the value UR. The resistance 38 thereby adjusts so that Ug = UR.

When signal Ip disappears again, diode 4o prevents a discharge of storage capacitor 43, so that the pitch ratio of voltage divider 35, 38, which has adjusted depending on the size of signal 1 ^, is maintained. When the signal I occurs, a voltage is obtained at the output 16, P lp corresponding to the quotient q = j = -. Subsequently, the control coupler 18 (Fig. 1) emits a pulse on the synchronization input 22. This causes the switch 42 to be short-circuit conducting and the storage capacitor 43 is discharged again.