WO2009097974A1

WO2009097974A1 - Authentication of objects

Info

Publication number: WO2009097974A1
Application number: PCT/EP2009/000410
Authority: WO
Inventors: Markus Gerigk; Wolfgang Borchers; Ludger BRÜLL; Martin Friedrich; Jürgen Focke; Simon Vougioukas; Wolfgang Joa; Raimund Klein; Manfred Moers
Original assignee: Bayer Technology Services Gmbh
Priority date: 2008-02-05
Filing date: 2009-01-23
Publication date: 2009-08-13
Also published as: RU2493968C2; RU2010136838A

Abstract

The invention relates to a method for identifying/authenticating objects based on a production-related and/or processing-related feature.

Description

Authentication of objects

The invention relates to a method and a device for identifying and / or authenticating objects based on production-related and / or processing-related, random features.

Identification is the process that is used to uniquely recognize an object. If an object is uniquely recognized, it can be uniquely assigned or it can be uniquely assigned to the detected object. For example, An item (object) can be assigned a price or its destination. The identification takes place on the basis of characteristics characterizing the object and distinguishing from other objects.

Authentication is the process of verifying (verifying) an alleged identity. The authentication of objects, documents or data is the statement that they are authentic - that is, they are unchanged, not copied originals.

Authentication also takes place on the basis of features characterizing the object and distinguishing from other objects. The features that are used for the authentication are preferably non-transferable, not copyable and not forfeitable.

Preferably, mechanically processable data are determined from the physical features for identification and / or authentication by means of physical methods, so that objects can be detected, assigned and authenticated by machine.

Today, omnipresent barcodes, 2D codes or OCR text (collectively referred to as optical codes) printed on objects are used to identify the objects. The identification of objects based on optical codes will be referred to as recognition in the following in order to linguistically distinguish them from the method according to the invention for identification.

Methods for authenticating objects are also known in the art. For much of these procedures will not be the authenticity of the object itself tested but the authenticity of a preferably inseparably connected to the object aid. An example is the use of magnetic particles which, depending on the distribution and / or orientation under electromagnetic radiation, show a characteristic signal that can be used for authentication (WO1998026379A). Here, the magnetic particles are used as aids and their distribution and / or arrangement used as an authentication feature for the object. A disadvantage of the mentioned method is that the magnetic particles must be connected to the object or introduced into the object. This is a step that is not generally used to make the object. It is thus an additional step, which means additional costs. Furthermore, not every object can be provided with magnetic particles; Metallic objects can prevent the measurability of the characteristic signal; in jewelry or design objects, the addition of such substances may be undesirable.

In another example, dyes are used which, depending on the composition, also show a characteristic signal (WO1999014416A1) which can be used for authentication. In this example, the authentication is done by a chemical reaction of a component A with a component B. This is a complex process that can not be used everywhere and not multiple times to determine the authenticity of an object.

US 5325167 describes a method for authenticating documents. The document contains a visible seal point on which opaque toner particles with reflection facets are applied. The toner particles have a unique microscopic grain structure, which is used as a seal to authenticate the document. The seal site is scanned and the result of the scan is stored in a database. During certification, the seal site is scanned again and the result is compared with the reference result in the database. A disadvantage of the described method as well as all previously described examples is that an additional element (here: seal point) must be introduced into the object or applied to the object. Each additional element means at least one additional step in the production and / or processing of the object and means additional costs. WO2005088533A1 describes a method which manages to authenticate an object without an additional element and allows objects to be uniquely assigned directly on the basis of their nature. In this case, a laser beam is focused on the surface of the object, moved over the surface and detected by means of photodetectors at different points of the surface at different angles different degrees of scattered rays. The detected scattered radiation is characteristic of a variety of different materials and individually for each surface and is very difficult to imitate because it is due to coincidences in the production of the object. For example, paper-like objects have a manufacturing fiber structure that is unique to each manufactured object. The scatter data for the individual objects are stored in a database in order to be able to authenticate the object at a later time. For this purpose, the object is measured again and the scatter data compared with the stored reference data. A disadvantage of the described method as well as all previously described examples is that for authentication, a special device must be used, which is solely for authentication. With a large number of points in the logistics chain of an object on which an authentication makes sense, such a device would have to be installed, which means additional costs.

It is thus based on the prior art, the task of providing a method for authenticating objects that without special devices, only the

Purpose of authentication serve, get along. It continues to set the task, a

To provide a method that can be used for both authentication and identification, without the additional resources on the object, especially only the

Serve authentication is required. It is also an object to provide a method that can be carried out with a device that, depending on the needs, only an identification, only an authentication, a sequential identification and

Authentication or parallel identification and authentication allowed. The process sought should be fast, cost-effective and intuitive. But it should also ensure a high level of security, so that the counterfeiting of objects is prevented or at least made more difficult.

Surprisingly, it has been found that objects are simple and inexpensive and using existing equipment, not just the purpose of authentication serve, can be identified and authenticated based on manufacturing and / or processing-related feature. The production-related and / or processing-related features are preferably part of a marking which is applied to or on the object, the marking not being used solely for authentication, but fulfilling further functions, as described below.

The subject matter of the present invention is therefore a method for identifying and / or authenticating an object by means of a detection unit on the basis of a production-related and / or processing-related feature which is part of a marking, characterized by the following steps:

- positioning of object and detection unit to each other,

optically detecting the mark or a portion of the mark, thereby producing a digital image file comprising the mark or a portion of the mark in machine processable form,

Localization of at least one characteristic feature within the image file,

Extraction of the at least one characteristic feature,

Comparison of the at least one characteristic feature with at least one characteristic feature of at least one object that has been detected at an earlier time and whose at least one characteristic feature has been stored in a database.

A marker is understood to mean a visible feature on the object, eg a text, a logo, a line, a graphic, a pattern, a photograph, a combination of the named elements or the like. The marking is applied to a carrier or introduced into a carrier. The carrier may be the object to be identified and / or authenticated itself or a part of the object. It may also be a package in which the object is packaged or it may be a label or tag attached to the object. Preferably, the carrier is inseparably connected to the object. Alternatively, it is conceivable to design the carrier in such a way that the An attempt to separate the carrier and the object leads to the destruction of the carrier, the object and / or the marking, whereby destruction is understood to mean that after the destruction an authentication of the object based on the marking is no longer possible.

The mark can be printed on the carrier. But it can also be e.g. be introduced by laser inscription, treatment with water jet or sandblast, by etching or engraving or other known method for marking in the carrier.

The marking has random production-related irregularities. This means that two markings applied by the same method to comparable supports or incorporated in comparable supports do not look identical, but have features that distinguish them. These distinctive ones

Characteristics have a random character, i. they are unpredictable. This results in a very high protection for the object: the irregularities can be reproduced, falsified and / or copied only at great expense.

These irregularities may be e.g. Variations in the ink distribution on the support be (splashes or the like). Also, the ink itself preferably has certain irregularities, e.g. in the size distribution of the pigments. Laser markings have e.g. Irregularities in the form of fringes in the border area between marked and unmarked material.

The carrier for receiving the marking may be made of paper, wood, textile, plastic, metal, ceramic, glass or other solid material. The carrier may also be a composite material of at least two of said materials. Preferably, the carrier also has random manufacturing and / or processing-related irregularities. If the marking is produced by printing technology, the carrier of the marking preferably consists of a fibrous material (paper, cardboard, textile, wood). Fibrous carriers show random production-related irregularities in the arrangement, size and orientation of the fibers. In the case of metallic carriers, the irregularity exists, for example, in grinding marks. It was surprisingly found that random manufacturing and / or processing-related irregularities in the carriers lead to an increase in the irregularities in the markers (synergy effect). For example, if a paper-like carrier has a fibrous structure that is on the order of the ink droplets applied to the carrier by inkjet technology to create the printed image, the fibrous structure will cause increased capillary force on the droplets causing the droplets to form Droplets continue to be distributed in and on the carrier as if the fiber structure of the carrier were smaller than the size of the ink droplets.

The carrier structure and material, marking agent and the process parameters for the production of the markings are preferably matched to one another in such a way that characteristic irregularities arise in the marking which are of the same order of magnitude as the spatial resolution of the registration unit or larger. Particularly preferably, the characteristic irregularities in the marking are greater than the spatial resolution of the detection unit.

Preferably, each object has at least one individual characteristic irregularity in the marking, by means of which it can be uniquely identified and / or authenticated. The irregularities are preferably robust, i. they do not change over time or only insignificantly.

From these individual, characteristic irregularities in the marking at least one manufacturing and / or processing-related feature, which is used for identification and / or authentication, extracted.

For this purpose, the object or, better, the marking associated with the object is positioned in a first step with respect to a detection unit. A detection unit for detecting a production-related and / or processing-related feature comprises at least one source which generates electromagnetic radiation, preferably with at least one wavelength in the range from 300 nm to 1000 nm. With the aid of the at least one radiation source, the marking or a part of the marking is irradiated. The detection unit further comprises at least one optical system, with which the radiation reflected by the marking or a part of the marking is directed onto at least one detector, and at least one detector, which is guided by the marking or a part the mark reflects reflected radiation into an electronic signal. The detector unit acts as a sensor for the electromagnetic radiation used. It preferably permits a spatial resolution of the electromagnetic radiation returned by the marking. The detector unit is eg a photoelement (phototransistor, photodiode) or an arrangement of photoelements. The arrangement may be cellular (line sensor) or planar (area sensor). Suitable optical sensors include, for example, CCD chips (CCD = charge coupled device, charge coupled device) or CMOS sensors (CMOS = comementary metal oxide semiconductor).

The detection unit preferably has means for signal and image evaluation (see, for example, Figure 2, elements 2-1, 2-3 and 2-4).

As a detection unit, e.g. a document scanner or optical code camera (e.g., smart cam) may be used. This has the advantage that a device is used for identification and / or authentication, which not only serves the purpose of authentication but also fulfills at least one additional function.

In a first step of the method according to the invention, the positioning of object and detection unit takes place relative to one another.

The positioning is the method of arranging the object and the detection unit with each other so that the marking or a part of the marking can be optically detected. Here the different functions of the marking become clear. The marking particularly preferably has at least three functions:

1. It has a characteristic feature that can be used for identification and / or authentication.

2. It can be used for the relative positioning of detection unit and object to each other, since it has optically detectable features that allow a clear determination of the position of the object and the marker with respect to the detection unit. 3. It fulfills at least one further purpose not directly related to the identification and / or authentication, eg the tag is an optical code or part of an optical code. Based on the optical code, the object can be detected. Or the mark is a logo or part of a logo on the basis of which the user of the object an assignment of the

Object to make a brand, company, image or the like. Further functions of the marking are conceivable.

Thus, the marking is not only for the purpose of identification and / or authentication. In contrast to many prior art authentication features, the method according to the invention makes use of an existing element of the object to be authenticated in order to perform an authentication and does not introduce an element which exclusively serves the purpose of the authentication. This makes the process according to the invention inexpensive.

The positioning can be done in several steps. In a first step, the object and the detection unit are usually positioned manually to one another. In the case of the use of a document scanner (FIG. 3), for example, the object is placed on a window for this purpose and, if necessary, aligned with markings on the window. In the case of using a camera (eg for the reading of optical codes), for example, this is held manually over the object. A monitor that displays the captured camera image can facilitate manual positioning. It is also possible to carry out the positioning automatically, wherein the marking or a part of the marking is used for positioning. This means that based on the position of the marking or a part of the marking on at least one sensor of a detector unit, a signal is generated which is used to drive an actuator, the object and detection unit to each other in the correct position required for identification and / or authentication brings. After the object or marking and detection unit have been brought into position relative to one another, the marking or a part of the marking is optically detected in a second step of the method according to the invention. For this purpose, the marking or a part of the marking is irradiated by means of electromagnetic radiation, preferably with at least one wavelength in the range of 300 nm to 1000 nm, and that of the marking or a part of the marking reflected radiation by means of optical components such as lenses, mirrors, diaphragms etc. directed to at least one detector. The detector converts the electromagnetic signal from the marker into an electronic signal.

Today's CCD or CMOS detectors have sensor resolutions (number of photoelements) of e.g. nominal 3 megapixels (2048 x 1536), 5 megapixels (2560 x 1920), 7 megapixels (3072 x 2304) and other resolutions (pixels = picture element). With a field of view size B of 20 mm x 15 mm, which is imaged on an optical detector with a resolution S of 7 megapixels, one pixel on the detector corresponds to an area in the field of view of 6.5 microns x 6.5 microns (spatial resolution). The spatial resolution O results generally from equation 1. It can be seen that a finer spatial resolution can be achieved by increasing the sensor resolution S (increased number of photoelements) or by reducing the field of view size B.

ß O = - (Equation 1)

Preferably, the spatial resolution is chosen to be of the same order of magnitude as the manufacturing and / or processing irregularities in the mark from which the characteristic feature for identification and / or authentication is derived, or chosen to be finer than the irregularities.

If the irregularities in the marking are very small and there is no sensor with a sufficient resolution, it is preferable to use an enlargement unit between the marking and the sensor. The enlargement unit generates an enlarged image of the mark or part of the mark. The enlarged image is then captured by the camera system. As an enlargement unit, e.g. a microscope objective (see, e.g., Figure 4).

It is also conceivable, several shots of the mark or part of the

Mark to make, between the individual recordings the mark and the

Detection unit relative to each other by an amount to be shifted, which causes the image on the photo sensors by an amount shifts smaller than a pixel (photo element) is (subpixel method). The individual recordings are combined into one recording. In this way, an image can be achieved with a resolution that is more accurate than the resolution of the detector used.

In a preferred embodiment of the method according to the invention, the marking or a part of the marking is irradiated areally and the reflected light is registered by a surface sensor as detector unit. Surface irradiation means that the entire part of the label that is used for identification and / or authentication is completely irradiated. The planar irradiation has the advantage that the marking or a part of the marking can be detected optically very quickly. It is likewise conceivable for the marking to be irradiated punctiform or linearly and for a relative movement to be carried out between the marking and the beam, so that all the portions of the marking required for the identification and / or authentication are irradiated one after the other and the radiations reflected by the different positions be sequentially collected / registered by a suitable detector unit.

The result of the optical detection is a digital image file that includes the mark or part of the mark in machine processable form. For this purpose, the signals occurring at the sensors of the detector (radiation intensities) are digitized. For example, gray levels can be assigned to the analog intensity signals. In an 8-bit representation, the gray values 0 to 255 are assigned to the various occurring intensity values, where 0 generally means black (dark) and 255 generally white (maximum intensity value). Of course, other representations are possible, in particular a binary representation in which all intensity values below a threshold value are assigned the value 0 (black) and all intensity values above a threshold value are assigned the value 1 (white). Preferably, the irradiation and / or detector parameters are adjusted so that a bimodal intensity value distribution results in which the two modes do not overlap. This can be determined empirically by irradiation experiments in which the parameters are optimized accordingly. In such a case, the threshold may be placed somewhere between the two modes. If the modes overlap, the threshold is preferably placed in the minimum between the modes. The gray-scale or black-and-white data can be stored as an image file (eg in bitmap format or a compressed format such as the JPEG format in a database (2-5 in FIG. 2)) or further processing (data reduction, feature localization / -extraction, adjustment) are supplied.

In a third step of the method according to the invention, the localization of at least one characteristic feature takes place in the image file. Preferably, the image file includes a graphic or it can be converted into a graphic or it can create a graphic from the image file. The graphic refers to a two-dimensional arrangement of pixels. A pixel (abbreviation of picture element) is a point of a graphic with a defined color or gray value. A graphic can be output on an output device (e.g., printer or screen) to make it visible. The visualized graphic shows the marking or a part of the marking of the object in a two-dimensional representation, optionally in enlarged form, optionally in a false-color representation.

The graphic of a marking or of a part of a marking preferably has at least three uniquely identifiable, characteristic points which do not lie on a straight line. These three points are preferably to be found in each graph of each mark of each object subjected to a method according to the invention. These three points serve as references to compare markers with each other. These three points span a triangle. One of the three points can be assigned the zero point of a coordinate system. From this go out two rays; one beam passes through one of the other two points, the other beam passes through the other of the other two points. The rays indicate the directions of the x and y axes of a coordinate system. On the basis of the coordinate system, the position of the marker in relation to the detection unit can be determined. Furthermore, the position of one or more characteristic features within the graphic can be localized on the basis of the coordinate system. In order to localize at least one characteristic feature within the image file, it is always necessary to determine the position of the coordinate system in the corresponding graphic. If the position of the coordinate system is determined, localization of at least one characteristic feature takes place within the graph. In this case, a characteristic feature may be part of the graphic; but it can also be that the entire graphic is used as a characteristic feature. One Characteristic feature is characterized by the fact that it is as unique as possible for the respective mark, so that it allows a distinction of the mark on / at the object under consideration from the mark of another object. The localization is based on the coordinate system determined for the respective graphic.

In a fourth step of the method according to the invention, at least one feature is extracted by selecting it after the localization and processing it independently of the data of the remaining graphic. From the one or more features a so-called signature is generated. The signature is a digitally storable and machine processable representation of the characteristic feature (s). The signature is unique, i. Objects with the same characteristics produce the same signature; Objects with different characteristics produce different signatures. The signature is robust and continuous, i. small changes in the features lead to small changes in the signatures, larger changes in the features lead to major changes in the signatures. The process of signature creation is not necessarily reversible, i. a signature does not necessarily mean that the features can be reproduced; but the signature allows a clear assignment to the characteristics. In principle, different signatures can be generated for the existing features. There is at least one process for their generation that is reproducible for each signature. Examples of the signature creation are given below.

Preferably, signatures of already acquired objects are stored in a database (reference data).

In a fifth step of the method according to the invention, at least one localized and extracted characteristic feature is compared with at least one further characteristic feature. The comparison is preferably carried out by means of the signatures mentioned. Based on the signature, the object can be identified and / or authenticated. For this purpose, the signature of the object is compared with at least one further signature which was generated at an earlier time. If the signatures are identical, the probability is very high that the objects from which the signatures were generated are identical. If the signatures are different, there is a high probability that the objects from which the signatures were generated are also different. In general, from the comparison of the signatures a probability for the conformity of the signatures and thus for the match of the objects can be determined.

In general, the compared signatures will not match 100%. The reason for this is, for example, the fact that the carrier of the marking and / or the marking itself undergoes an aging process and changes as a result of environmental influences. Furthermore, the process of optical detection of the mark is subject to certain variations. It is in the optical detection e.g. It is not always possible to irradiate exactly the same area so that, if necessary, a slightly varying signal is determined during each acquisition process. As a rule, therefore, a threshold value S is set. If the degree of correspondence between the signatures is S or more, then a match is considered given, if the degree of correspondence is below S, the compared data sets are considered different.

A number of algorithms are known for comparison (see, e.g., Image Analysis and Processing: 8th International Conference, ICIAP '95, San Remo, Italy, September 13-15, 1995. Proceedings (Lecture Notes in Computer Science)). It is conceivable to resort to the extensive work on the recognition of biometric features of humans (fingerprint recognition, face recognition, iris recognition, venous recognition, retina recognition, etc.), because even in the case of biometric features, it is necessary to record characteristic features and reference data to compare. Here, too, optical methods are used for capturing and image evaluation during adjustment. Examples of methods for balancing e.g. Fingerprints can be found in US Pat. No. 7333641 B2, DE10260642 A1, DE10260638 A1, EP1435586B1.

It is conceivable to carry out the comparison with the image files or the graphics themselves. Due to the above-mentioned fluctuations in each optical detection process, however, the image files are subjected to image processing in which at least one characteristic feature which makes up the respective marking is determined. Such a method is also referred to as fingerprinting, since a fingerprint is virtually determined from each digitally present marking (or marking part). Compared to a comparison based on the image files or graphics has fingerprinting also the advantage of higher speed, as smaller data sets are used for the mathematical calculations. Likewise, the amount of data used to store the reference data during fingerprinting is smaller, so that less storage capacity is required.

In a method according to the invention for the identification of an object, a so-called 1: n alignment takes place, i. the currently determined signature (1) is compared with all the (s) stored in the database signatures to determine the signature that largely coincides with the currently determined signature. This signature belongs to the object that is likely to be identical to the currently detected object. Typically, the object is identified if there is a single signature in the database for which the degree of correspondence with the currently determined signature is greater than a threshold S. The threshold value S is preferably determined empirically: If there are several signatures in the database for which the degree of correspondence with the currently determined signature is greater than the threshold value S, the threshold value S is obviously set too low; If an object is detected twice in succession and the signatures thus obtained have a degree of coincidence which is smaller than the threshold value S, the threshold value S is obviously set too high.

In a method according to the invention for authenticating an object, a so-called 1: 1 adjustment takes place. The object has already been identified, it being immaterial to the method according to the invention how the object has been identified, it is crucial that an identity has been established. The method according to the invention serves to verify the alleged identity. There is only a match between the signature of the currently detected object and the signature of that object with which the object is due to match the identity. If the degree of coincidence is above a threshold value S, an authenticity is considered given; if the degree of agreement is below the threshold S, the objects are considered different and the claimed identity could not be verified. The threshold value S is preferably determined empirically. The comparison can be made by means of a microprocessor, which may be part of the detection unit. It is also conceivable that the detection unit generates image files or extracted data which are fed to a separate unit for adjustment. The detection unit may be connected to the unit for adjustment; However, it is also conceivable that the generated data are first stored in a memory in the detection unit and the detection unit is only connected at a later time with a unit for adjustment to transmit the data. The connection or connection can be made electronically via cable, via radio, optically, acoustically or via another channel of signal transmission.

It is conceivable that the reference database is part of the registration unit.

The overall process, which preferably passes through every object that is to be supplied to an identification and / or authentication at some point, is summarized as follows:

I. First Assessment: At least one manufacturing and / or processing-related feature on the object is optically detected by means of a detection unit.

π. The optically detected by the detection unit feature is digitized and possibly digitally processed. The result is a signature.

HI. The signature is stored in a reference database along with other signatures of other objects.

IV. Identification / Authentication: An already detected object becomes the

Identification and / or confirmation of its identity again supplied to a detection unit that optically detects at least one manufacturing and / or processing-related feature on the object.

V. The optically detected by means of the detection unit feature is digitized and possibly digitally processed. The result is a signature to be checked.

VI. The signature to be checked is compared with at least one existing signature from the reference database. VII. A probability of matching the signature to be checked with at least one signature from the reference database is determined.

If the relative position of the object to the detection unit in the identification / authentication with respect to the position at the first detection different (rotation, displacement, compression, extension), so the amount of displacement and / or rotation and / or compression and / or extension in relation to the situation at the initial registration on the basis of the three clearly identifiable, characteristic points of the diagram are determined and if necessary corrected (see eg Alfred Nischwitz et al., computer graphics and image processing, Vieweg publishing house, 2nd edition, 2007). It is thus possible to feed the recorded graph from the three points of a coordinate transformation to make the graph coincide with a graph created earlier.

It is conceivable to use a higher spatial resolution and an enlarged image detail in the initial detection than in the later identification and / or authentication. The quality of the recorded data during the initial assessment significantly determines the quality of the later matching in the identification and / or authentication (hit rate, accuracy of the match). It is conceivable to make several recordings of the marking and / or a part of the marking in the initial recording and to generate an improved reference data material from an image averaging.

To carry out a method according to the invention, a device is required, which is referred to here as a detection unit. This detection unit comprises at least one source of electromagnetic radiation for irradiating the marking, at least one optical system for imaging the radiation reflected by the marking onto at least one detector unit and at least one detector unit for detecting the reflected radiation and for converting the radiation into electronic signals. For carrying out a method according to the invention, a unit for digitizing the electronic signals, an evaluation unit for digital data processing and for feature extraction, as well as a unit for comparing a characteristic feature with features stored in a database are also required. These mentioned units can be integrated in the registration unit or in one or a plurality of units, which may be connected to the detection unit and / or to each other, be carried out separately.

Preferably, a document scanner, a copier or a camera system, which is preferably used for optical codes, used as a detection unit. If required, this existing device will be supplemented with means that can perform image manipulation, feature localization and extraction. The supplementation can take place by means of microchips (hardware) and / or suitable algorithms (software).

The detection unit according to the invention may have means for comparing data (FIG. 2). It is also conceivable that the detection unit according to the invention generates images or extracted data which are fed to a separate unit for adjustment. The detection unit according to the invention may be connected to the unit for comparison; However, it is also conceivable that the generated data is first stored in a memory in the detection unit according to the invention and the detection unit according to the invention is connected to a unit for comparison at a later time to transmit the data. The connection or connection can be made electronically via cable, via radio, optically, acoustically or via another channel of signal transmission. It is conceivable that the reference database is part of the detection unit according to the invention.

In a particular embodiment, the detection unit according to the invention has a magnification unit which can be optionally mounted between the object and the detector. It preferably has different magnification factors, e.g. the factors 2, 5, 10, 20, 50, 100 or others.

Hereinafter, a concrete embodiment of the present invention will be described, but without limiting the invention to this embodiment.

Figure 1 shows as an object to be identified a package (1-1) carrying a printed optical code of the type DataMatrix as a marker (1-2).

Figure 2 shows schematically a detection unit comprising a control unit (2-1), a source of electromagnetic radiation (2-2), a detector unit with built-in optics (2-3) and an evaluation unit (2-4). The detection unit is connected via the control unit to a database (2-5).

FIG. 3 shows schematically a part of the detection unit according to the invention in the form of a scanner. The object to be authenticated (3-1) is placed on a window (3-2) and is irradiated from below by means of a source of electromagnetic radiation (3-3). The radiation reflected at the object is directed via mirrors (3-4, 3-5) by means of optics (3-6) onto a detector (3-7). One of the mirrors (3-4) and the radiation source (3-3) are made movable (direction of movement is illustrated by the arrows), so that different parts of the object (3-1) can be irradiated without moving the object. This will cause the object to be scanned and scanned. Figure 3 (a) is a perspective view of the scanner. Fig. 3 (b) and Fig. 3 (c) show the scanning apparatus and the object from the side at different timings of the scanning operation.

FIG. 4 shows a schematic side view of an object (4-1) with an optically detectable marking (4-2) as well as the beam path during the imaging of the marking on an area detector (4-8). In the illustration, the mark (4-2) is magnified by means of a lens (4-3). The result is an enlarged intermediate image (4-4). This is imaged on the detector (4-8) using an eyepiece (4-5), a camera object (4-6) and an aperture (4-7).

FIG. 5 shows, as an example of a marking according to the invention, the printed image of an optical code of the type DataMatrix (5-1) as well as the approximately 25-fold magnification (5-2) of the printed image. Magnification indicates production-related irregularities (e.g., a fuzzy line at the bottom of the L-finder, indicated by the arrow (5-3)), which can be used as characteristic features for identification and / or authentication.

FIG. 6 shows, as an example of a marking according to the invention, the enlarged printed image (6-1) of a DataMatrix code. One area was selected and enlarged approximately 16 times (6- 2). Again, an area was selected from the enlarged area and enlarged again about 16 times (6-3). The production-related irregularities in the margins of the printed image are clearly visible. Figure 7 shows the effect of an edge detection method applied to the image of a DataMatrix code (7-1). The arrow (7-2) symbolizes the edge detection method. As a result (7-3), the edges were pictorially emphasized.

FIG. 8 shows the image of a DataMatrix code in edge representation (8-1). By averaging the position values of the lower and left edge of the DataMatrix code (so-called L-Finder), the x-axis (8-3) and y-axis (8-4) of the coordinate system can be determined. At its intersection (8-2) lies the origin of the coordinate system.

FIG. 9 shows a characteristic region (9-1) from a printed image (see FIG. 6). From this a line pattern (9-3) can be determined by means of an edge detection method (indicated by the arrow 9-2) which can be used as a signature.

In a preferred embodiment of the present invention, machine-readable optical codes produced by printing technology are used as markings on or on the object or on the packaging of the object and production-related irregularities of the markings are used for identification and / or authentication (FIG. 1). The optical codes primarily serve the purpose of recognizing the corresponding object by known methods. Among machine-readable optical codes, e.g. Bar codes, stacked codes (e.g., "Codablock" or "PDF417"), matrix codes (e.g., "DataMatrix" or "MaxiCode"), or OCR (Optical Character Recognition) text. These are usually printed on the object or the packaging of the object (Figure 1). They have optically detectable irregularities caused by the pressure, the pressure medium and / or the support of the printed image. These irregularities in the printed image or parts of it are to be optically recorded.

For initial detection, the object is fed to a detection unit, which is shown schematically in FIG. Upon acquisition, the DataMatrix code is irradiated by a source of electromagnetic radiation (2-2). The reflected radiation is directed to a detector with built-in optics (2-3). Since the DataMatrix code for the recognition of an object usually with the help of an optical code camera (code camera) is read, such a code camera is preferably used for authentication according to the invention. Code cameras for reading a DataMatrix code usually use an area sensor on which the entire code is mapped at once. If the irregularities in the printed image are smaller than the spatial resolution of the code camera, a microscope object between the code camera and the object can be used to increase the irregularities (FIG. 4).

FIG. 5 shows the captured image of a DataMatrix code. At about 25x magnification significant irregularities in the printed image can be seen. Especially the lower part of the L-Finder shows a fringy line (see arrow 5-3). The further magnification (see Figure 7) to the area where the individual pixels are clearly visible, make it clear that the printed image contains characteristic features that can be used for identification and / or authentication. For this, the features must be localized, extracted and stored.

If the DataMatrix code is completely captured, eg edge detection methods (Sobel operator, Laplace operator, Prewitt operator, Roberts operator or the like) can be used to determine the outer boundaries of the code (FIG. 7). With the DataMatrix code, the so-called L-Finder offers the possibility of placing a coordinate system which is the basis for all further mathematical operations (Alfred Nischwitz et al., Computer Graphics and Image Processing, Vieweg Verlag, 2nd edition, 2007). The average values of the positions of the lower and left edges of the L-finder form the x and y axes of the coordinate system (FIG. 8). The zero point of the coordinate system results from the intersection of the axes. Further characteristic points form the other three corners of the DataMatrix code. It is now possible to extract a feature by, for example, selecting an area of defined size at a defined distance from the zero point in the direction of the x-axis. Such an area could be the area designated 9-1 in FIG. 9, which was arbitrarily selected from the DataMatrix code in FIG. The location and size of the area is clearly defined. The area is characteristic and thus suitable for the creation of a signature. It is possible to digitally save the image 9-1 as a file and to use it as a signature. It is also possible to digitally store only the line pattern (9-2) that determines the characteristic as a file and to use it as a signature. To store the signatures is a database that can be connected to the detection unit (Figure 2). During identification and / or authentication, an image file of the DataMatrix code is created as in the initial capture. From the image file, a graphic can be generated in which the position of the code is determined using the L-finder and from which a characteristic feature is extracted at a defined position of the L-finder. The characteristic is compared with signatures in the database. In the identification, a similarity value (correlation coefficient) is determined for each pair of characteristic to be tested and signature. Thus it can be specified with which probability the characteristic to be tested and the respective signature match. It can be the indication of the object whose signature with the highest probability matches the characteristic to be tested. The object is identified.

The method described is referred to as a 1: n adjustment, since an object to be identified is compared with n different objects in order to identify the object with which the object to be tested is in all likelihood identical. The use of optical codes offers the possibility of speeding up the matching: by means of the optical code, the object can be detected by means of known methods, e.g. detected by an optically stored identification number and thus the number of objects with which must be adjusted considerably reduced. Ideally, the optical code allows the object to be assigned to an object in the database, and it only needs to be checked whether it is actually the identified object. For this purpose, a characteristic feature of the object to be tested is compared with the corresponding characteristic feature of the identified object in the database (verification). This adjustment is called the l: l adjustment.

It makes sense to shift the position of the line in positive and negative x- and y-direction and to determine the correlation coefficient for each shift. Although the location points of the beginning and the end of the line can be determined unambiguously by determining the x- and y-axis and the zero point from the L-finder, there are always slight distortions in the image which leads to errors in the position determination , The inventive method is suitable for the identification and / or authentication of persons, animals and all conceivable items such as packaging, letters, parcels, documents, money, identity cards, jewelry, medicines, electronic and mechanical components, intermediates, end products, other valuables, etc.

The invention is characterized by a high degree of robustness, is stationary and mobile use, intuitive, inexpensive in use and allows the combination with existing methods for identification using optical codes.

reference numeral

1-1 object

1-2 mark in the form of a printed optical code of type DataMatrix

2-1 control unit

2-2 Source of electromagnetic radiation

2-3 detector unit

2-4 evaluation unit

2-5 database

3-1 object

3-2 windows

3-3 Source of electromagnetic radiation

3-4 mirrors

3-5 mirrors

3-6 optics

3-7 detector

4-1 object

4-2 mark

4-3 lens

4-4 intermediate image

4-5 eyepiece

4-6 Camera obj. Ect

4-7 aperture

4-8 detector 5-1 Printout of a DataMatrix code

5-2 25x magnification of a print image of a DataMatrix code

5-3 Characteristic, production-related, random features

6-1 Enlarging a Print Image of a DataMatrix Code 6-2 Enlarged section of a print image of a DataMatrix code

6-3 Enlarged section of a print image of a DataMatrix code

7-1 Printout of a DataMatrix code

7-2 arrow symbolizing an edge detection procedure

7-3 Result of an edge detection method on the printed image of a DataMatrix code

8-1 Edge representation of a DataMatrix code

8-2 Origin of a coordinate system spanned by the axes of the L-finder of the DataMatrix code

8-3 x-axis 8-4 y-axis

9-1 characteristic features from the edge region of the printed image of a DataMatrix code in an enlarged view

9-2 arrow symbolizing an edge detection procedure

9-3 Result of an edge detection method on a characteristic feature from the edge region of the printed image of a DataMatrix code in an enlarged view

Claims

claims

Method for identifying and / or authenticating an object by means of a detection unit on the basis of a production-related and / or processing-related feature which is part of a marking, characterized by the following steps:

- positioning of object and detection unit to each other,

optically capturing the mark or a portion of the mark printed image, thereby producing a digital image file comprising the mark or a portion of the mark in machine processable form,

- Localization of at least one characteristic feature within the

Image file

Extraction of the at least one characteristic feature,

2. The method of claim 1, with the additional step

Determining at least one match probability between the characteristic feature and the at least one characteristic feature of the at least one previously detected object.

3. The method of claim 1 or 2, characterized in that from the image file, a graph is generated, which contains at least three uniquely identifiable points that are not on a straight line and form a coordinate system that is used to determine the position of the object and detection unit to each other is used.

4. The method according to claim 1 to 3, characterized in that the three points are used for a coordinate transformation in order to bring the graphic with a previously generated graph to cover.

5. The method according to any one of claims 1 to 4, characterized in that the three points uniquely define the position of the at least one characteristic feature.

6. The method according to any one of claims 1 to 5, characterized in that the marking is the printed image of an optical code.

7. The method according to any one of claims 1 to 6, characterized in that for detecting a reader for optical codes, a camera or a

Document scanner is used.

8. The method according to any one of claims 1 to 7, characterized in that between object and detection unit, a microscope objective is mounted.

9. The method according to any one of claims 1 to 8, characterized in that a subpixel method is applied to increase the spatial resolution.

10. The method according to any one of claims 1 to 9, characterized in that by means of an edge detection method, a line pattern is extracted as a characteristic feature.

11. An apparatus for identifying and / or authenticating an object based on a manufacturing and / or processing-related feature that is part of a

A marking is, comprising at least one source of electromagnetic radiation for irradiating the marking, at least one optical system for imaging the radiation reflected by the marking on at least one detector unit, at least one detector unit for detecting the reflected radiation, a unit for digitizing the detected radiation

Evaluation unit for digital image processing and for feature extraction and a unit for comparing a characteristic feature with features that are stored in a database.

12. The device according to claim 11, characterized in that between the marking and the detector unit optionally a magnification unit can be attached.