EP0947964A1 - Method for testing valuable papers - Google Patents

Abstract

A scanning area is designed as a matrix of scanning values. This matrix has a preset number (n) of columns (S1, ..., Sn) and a number (m) of lines (Z1, ..., Zm). The scanning values can correspond to the intensity of a particular color. Other combinations of physical measurements can also be used. A line vector (ZV) is generated and analyzed to match a specimen. An Independent claim is included for a circuit arrangement for carrying out the method.

Description

Technical field

The invention relates to a method and a device for recognizing and / or testing of securities, the security being scanned in a predetermined grid is identified and checked for authenticity by a pattern comparison.

State of the art

The automatic detection of banknotes and other securities is both in the Bank area as well as, for example, an important point of sale Issue.

A simple device for checking banknotes is e.g. from US 5,607,040 known. The banknote is drawn lengthways between two scanner units.

From DE 39 18 798 A1 a bank note checking device is known, which the Banknotes can also be recognized in an inclined position. Depending on the position of the banknote, they are added line by line scanned pixels in a certain way in a test data memory written so that the image of the banknote in the memory is always the normal position occupies. During the test, the image data with a predetermined pattern compared. With the specified algorithm for storing the test data, the Determination of the angle of the inclination and rotation of the image in the memory be avoided.

DE 195 17 194 A1 discloses a device for checking banknotes, in which the banknote is evenly distributed over the entire width using mirrors is illuminated. For example, two to four linear, parallel to each other arranged CCD arrays used, with a filter for a certain wavelength is arranged. For the evaluation the are in a Column values of the different CCD arrays combined. For note recognition is not only in the visible but also in the non-visible spectral range measured.

A method of checking the authenticity of money is e.g. in US 5,615,760. In a multi-dimensional space (which is determined by the measured parameters is established) it is determined whether a combination of parameter values within or lies outside an n-dimensional ellipsoid.

The known methods are relatively complex on the one hand and react on the other sometimes sensitive to banknotes not supplied. In the latter Case needs to be designed with the corresponding effort ensure that the banknotes are always in a well-defined position on the Readers are led past.

Presentation of the invention

The object of the invention is to provide a method of the type mentioned at the outset, that reliable identification and authenticity check with relatively little Hardware effort enables.

The solution to the problem is defined by the features of claim 1. According to the Invention are the samples arranged in rows and columns of a representative Area with simple mathematical operations on a line vector pictured. An area that is a characteristic identifier is therefore reduced to a line in such a way that all samples are taken into account in terms of content are.

Experiments have shown that the line vector according to the invention is within certain limits Limits is relatively robust against variations in angular position. So will the banknote is conveyed slightly at an angle and read, the recognition performance in the not significantly affected.

The sections are chosen in advance so that - in terms of area - the most of the surface of the security is taken into account. If not all cutouts This fact can have the same meaning for the authenticity of the security if necessary e.g. be taken into account by appropriate weighting. In principle, a single contiguous section can suffice. Preferably but at least two areas of approximately the same size in the inventive Processed way. Corresponding are at least two line vectors to build.

The sections are advantageously defined in such a way that more elements or samples per line are present as per column. If the data reduction is too strong, detection performance may be undesirably throttled.

The part of the image taken into account for the recognition and authenticity check should be a predetermined one minimum distance (e.g. 5 mm) from the edge of the security. Namely the edge is heavily worn on older banknotes and rather for analysis annoying as helpful.

According to a particularly preferred embodiment, the becomes a vector component total value by adding up all samples of the corresponding Column created. Other calculations would also be possible. Tailored to that recognizing image could e.g. certain line values added up and others subtracted so that the total value reflects a difference value. Could continue weighted averages are also generated (with the choice of weights based on those Can turn off image characteristics which are important or which are all recognizable Securities are common).

The hardware scanning area is advantageously larger than the dimension of the Securities. When recognizing banknotes of different sizes based on the format of the largest banknote. This makes it position-independent Security detection possible. The edge of the security becomes software determined. In this way, the angular position of the security can be determined become.

A particularly simple and quick method to compensate for the angle dependency of the line vector according to the invention consists of several angular ranges are given with their own reference vectors. When processing the The security is first determined in which angular range the security lies, to then simply access the corresponding reference. The reference vectors are permanently stored in a memory.

The invention is based on a surface scanning of the security. It in principle, it does not matter what type of sensors are used to record the samples become. Optical scanning is common. Banknotes are used, for example, at a certain wavelength in the visible range of the spectrum.

The samples can also be difference values. For each sampling point e.g. the Subtracted intensity in red from that in green.

Certain banknotes also have characteristic features of electrical and / or magnetic type. Under these circumstances, the samples can also be corresponding represent physical quantities.

The method according to the invention can be implemented using known methods Units. The processing is typically implemented in software.

From the following detailed description and the entirety of the claims there are further advantageous embodiments and combinations of features of the Invention.

Brief description of the drawings

Fig. 1

Eine beispielhafte Darstellung der Lage der Auswertebereiche;

Fig. 2

eine schematische Darstellung zur Erläuterung der erfindungsgemässen Datenreduktion;

Fig. 3

eine schematische Darstellung der Abtastung einer schräg eingezogenen Banknote;

Fig. 4

ein Blockschaltbild einer Schaltungsanordnung zur Durchführung der erfindungsgemässen Verarbeitung.

The drawings used to explain the exemplary embodiment show:

Fig. 1

An exemplary representation of the location of the evaluation areas;

Fig. 2

a schematic representation to explain the data reduction according to the invention;

Fig. 3

a schematic representation of the scanning of an obliquely drawn banknote;

Fig. 4

a block diagram of a circuit arrangement for performing the processing according to the invention.

In principle, the same parts are provided with the same reference symbols in the figures.

Ways of Carrying Out the Invention

Fig. 1 shows a banknote 1. This is completely scanned so that the edge software can be determined. The scanning area 14 is larger than the one to be processed Banknote 1.

In the present example, two cutouts 2.1, 2.2 are defined for banknote 1. Together, these cover the essential area of banknote 1 (e.g.> 50% of the area). You have e.g. the shape of elongated rectangles. A given one Edge area 3 of the banknote is deliberately not taken into account.

As can be seen from FIG. 2, a scanning area (and thus each partial section thereof) can be understood as a matrix of samples. The matrix of a section 2.1, 2.2 has a predetermined number n of columns S1, ..., Sn and one (in generally smaller) number m of lines Z1, ..., Zm.

The samples correspond e.g. the intensity of a particular color at the scanning location or the difference of different color intensities. Other combinations can also be used of physical measurements.

According to a preferred embodiment of the invention, all samples a column summed up to a component of a row vector. According to the existing number n of columns becomes a row vector ZV with n Components created. One line vector results for each section 2.1, 2.2.

Each line vector ZV can then use a known pattern recognition method are analyzed (cf. the prior art mentioned at the beginning, e.g. US 5,615,760).

An important aspect of the invention is that the banknote is not currently being read must become. If the banknote is read in obliquely as shown in FIG. 3, then the angle β can be determined on the basis of the course of the edge 4. The cutout 5 is rotated accordingly with respect to the edge 4 of the banknote.

According to the invention, different angles β differ in advance Reference vectors generated and saved. Each reference vector applies to a specific one Angular range.

Example: The entire angular range from -10 ° to + 10 ° is divided into five sections, namely -10 ° to -6 °, -6 ° to -2 °, -2 ° to + 2 °, + 2 ° to + 6 °, + 6 ° to + 10 °. Becomes a An angle of, for example, 3 ° is determined, then the reference vector of the partial area + 2 ° to + 6 ° used as the basis for pattern recognition.

With small angles, it is very easy to determine the angle from the course of the edge close because the angular functions sin (), cos () and tan () are known per se Can be approximated linearly.

4 shows a block diagram of an arrangement for carrying out the method. The bank note 6 is preferably transverse to the conveying direction with its longitudinal direction 7 past a reader 8. The reader 8 has a length greater than that Banknote 6 is. The bank note 6 can thus be read in regardless of the position.

The reader (e.g. an optical scanner unit) sends the sample values line by line to one Preprocessing unit 9. This can be designed so that it scanned the whole Can take picture. But it is also conceivable that the lines in the same necessary evaluation (detection of the edge and the angle, Generation of a running total of the columns).

From the preprocessing unit 9, the samples of the predetermined sections become passed on to the column summer 10. This generates the characteristic ones Row vectors ZV. In the subsequent pattern recognition 11, the Line vectors compared with the reference vectors. The result is e.g. a probability value for correct identification or for the authenticity of the banknote.

The angular position β is also extracted from the preprocessing unit 9. She will be in the selection circuit 12 used to store the relevant reference vector from the reference vector 13.1, ..., 13.4 for pattern recognition.

It is understood that the described embodiment is in many respects can be modified without departing from the scope of the invention. Of the Reader 8 can by several different scanning units in different wavelength ranges be supplemented. For example, also a unit for measuring capacitance be used in a predetermined grid. The samples must thus do not necessarily correspond to optical pixels.

In summary, it can be stated that the invention proposes a method which Reliable and location-independent testing and detection with simple means of characteristic features.

Claims (9)

Method for recognizing and / or checking securities, wherein a) the security is scanned in a predetermined grid and b) identified by a sample comparison or checked for authenticity,
characterized in that c) a line vector (ZV) is generated from at least one areal section containing a characteristic identification feature, the components of which each represent a total value taking into account all the sample values of a column of the section (2.1, 2.2). A method according to claim 1, characterized in that the at least one Section (2.1, 2.2) is selected so that the essential part of the area Is covered. A method according to claim 2, characterized in that the at least one Cutout (2.1, 2.2) does not cover a predetermined edge area (3). Method according to one of claims 1 to 3, characterized in that the Total value of a column essentially the sum of all samples of the column is. Method according to one of claims 1 to 4, characterized in that for Scanning the security is selected a scanning area that is larger than that Security and that an angular position of the security through an evaluation of the scanning area is determined. A method according to claim 5, characterized in that determined first is in which of several predetermined angular ranges the angular position and that for the pattern comparison is assigned to the selected angular range Reference pattern is used. Method according to one of claims 1 to 6, characterized in that the Security is scanned in two different spectral ranges and that the Difference of the measured intensities is used as the basis for the evaluation. Circuit arrangement for performing the method according to claim 1, comprising a) a reader (8) for scanning the security line by line in a scanning area which is larger than the dimensions of the security, b) a summer for summing the sample values in columns in a section of the sample area containing a characteristic identifier, c) a pattern recognition circuit for comparing the line vector with a reference vector. Circuit arrangement according to claim 8, characterized by a circuit to determine the angular position of the security and a selection circuit (12) for selecting the reference vector assigned to the angular position.

Images