A variety of methods of security marking of objects have been proposed or are currently in use. For instance, objects can carry one or more labels overtly or covertly (or both) which carry a code identifying the object. That code is stored in a database. The label can carry the code in the form of printed numbers or bar codes, DNA or specific antibodies, or various electronic methods such as microchips, etc. The disadvantage with these type of systems is that once one knows the identification code, it is relatively easy to reproduce identical identifying labels.

It is also known for preventing counterfeiting to provide objects such as banknotes or credit notes with parts which are difficult to reproduce. The notes can use patterns which use special colours or resolutions, or can carry holographic labels.

It is an object of the present invention to provide an improved identifier label and labelling system in which it

is very difficult, to the extent of being practically impossible, to reproduce the label, even given the knowledge of the identification code which it carries, and therefore to provide an uncounterfeitable unique label.

Accordingly, the present invention provides a unique identifier (tag or label) which is based on random physical discontinuities which can be incorporated into or onto the item to be uniquely labelled.

For instance, the invention can provide a label carrying an identification code, the label comprising first and second mutually distinguishable components, the second component being fixedly-and randomly-distributed non- uniformly with respect to the first, the position of the second component forming the identification code of the label.

The second component may be distinguishable from the first optically or in other ways, e. g. by being magnetic, radioactive, or giving a change in conductance, resistance or inductance. The requirement is that the two components are heterogeneous after the label has completed manufacture and are measurably distinguishable.

The first component can be a matrix which can be of solidified plastic or resin. The second component can be a

plurality of code elements, for instance beads/particles of the same or different sizes, fixed within the matrix.

There may be a multiplicity of beads and the beads may be visually distinguishable (e. g. by opacity, colour, refractive index or size) to allow reading of the label using an optical reader. The beads may be, for instance, of 10 to 100 ym in diameter and the label can be, for instance, less than 1 mm thick.

Instead of beads, bubbles could be used. They could contain a substance to facilitate a particular type of measurement, e. g. osmium salts for electron microscopy or barium salts for X-ray measurement.

Alternatively, the code elements could be incorporated into a liquid aerosol spray (e. g. lacquer) sprayed onto a surface and as the spray dries the elements will be held in position. Variations in this technique such as brushing on, painting on, dipping, electrostatically attaching or making each element or the material sticky are also possible.

The position of the code elements within the matrix can be measured in two dimensions, for instance by noting their position above or below a line, as indicative of a binary code, or in three dimensions.

The matrix forming the first component can be the material of the article to be labelled. For instance, if a glass or plastic bottle is to be labelled, the second component, e. g. beads, could be mixed with the glass or plastic, but preferably in a defined region of the bottle.

Clearly, other types of container or article can be labelled in this way, for instance: car windscreens, plastic, cellophane, the lacquer coating of packaging on a cardboard base etc.

The two components need not be beads in a matrix. For instance, they could be two distinguishable solidifiable liquids, e. g. two differently coloured resins. Incomplete mixing to produce a non-uniform mixture followed by formation into a label and solidification would give a unique pattern. The pattern could be stored by pattern recognition apparatus or a code generated e. g. by measuring the proportion of one component in given regions and designating proportions above 50% as a binary"1"and below 50% as a binary"0".

The invention also provides a method of manufacturing such a label, the method comprising providing the first and second components and adding them together to fixedly, randomly and non-uniformly distribute the second component with respect to the first.

Thus, an almost infinite number of unique labels can easily be produced just by mixing and stirring the two components, e. g. the beads into the matrix (e. g. resin), then forming the mixture into the label. The mixture may be such that it can form a label itself, or be attached to a carrier to constitute a label, or incorporated as an integral part of the article.

The label can be used by identifying the position of the second component within the first, e. g. the elements within the matrix, and recording that position as an identifier of an object to be labelled. The label is then attached to the object. The code is stored in a database so that it is always possible to check the identity of the object by reading the label attached to it and checking it with the database. The object could also be provided with a serial number against which the identification code could be checked, or two labels according to the invention could be attached to an object, for instance one overtly and one covertly, and both codes recorded in the database.

An advantage of the invention is that even with knowledge of the code carried by the label, it is very difficult to produce a second label with that code. For instance, where beads are used it is very difficult to reproduce the exact random conditions which led to the particular dispersion of beads in the first place. It

should be noted that with the invention, it is not a case of generating the identification code and then forming the label according to that code, but forming the label first, reading its code and then recording that code for the label and labelled object.

Other aspects of the invention provide a corresponding method of labelling an object and a system for reading, analysing and storing identifiers and collating subsequently read identities.

The invention may be used, for instance, to identify documents, e. g. bank notes, bearer bonds, travellers cheques, passports, prescriptions, security papers/documents, credit cards, coupons, certificates, gift certificates/vouchers, and other documents of value.

The invention will be further described by way of non- limitative example, with reference to the accompanying drawings, in which:- Figure 1 diagrammatically illustrates a label according to a first embodiment of the invention ; Figure 2 diagrammatically illustrates a label according to a second embodiment of the invention ;

Figure 3 diagrammatically illustrates a third embodiment of the invention; Figure 4 diagrammatically illustrates a security/ validation system according to the invention; and Figure 5 diagrammatically illustrates an embodiment of the invention incorporated in a bottle; Figures 6 and 7 illustrate schematically examples of the reader/recorder system; Figure 8 illustrates the control system for the reader/recorder; Figures 9 and 10 are images obtained from example labels; and Figures 11 to 13 are images obtained from further example labels.

As shown in Figure 1, the label 1 is composed of a matrix 3, typically of translucent plastic (e. g. acrylic polymer) which forms the body of the label. The body can be, for example, up to 1 mm thick. Embedded within the matrix are beads/particles 5 of predefined diameter (e. g.

10 ym polystyrene beads). These are incorporated in the

matrix in a random, non-uniform distribution. This is achieved by first mixing a plurality of the beads with liquid matrix material, forming the mixture into the required shape of the label and then allowing the matrix to solidify. As the matrix solidifies, the beads are trapped and held in position. The plastic used is preferably flexible when solidified.

In one example, the matrix. can be formed, for instance, of an epoxy resin which is solidified by mixture with a setting agent.

The mixing process results in the random positioning of the beads. Thus, there is no requirement to physically "place"the beads in a given position. The random nature of the process generates a large number of unique labels.

Figure 2 shows a variation of the embodiment of Figure 1 in which beads 7 and 9 of different diameters (two in this case) are used in the label 20. This increases the number of variations achievable by a label containing a given number of beads. Differences in other physical differences can be used, e. g. colour or shape.

An alternative way of producing the labels shown in Figures 1 and 2 is by incorporating the beads into a liquid to be painted, sprayed or brushed onto an object before

solidifying. Or the beads, or the label or object could be made sticky so that they adhere on application. In a further alternative, the beads could be applied before being covered by a solidifying material to fix their position.

The code formed by the positioning of the beads can be read in a number of ways, using any number of a range of standard visual (or other) properties to record the position of the dispersed beads. For example, as shown in Figure 1, the distribution of the beads about a central line (which may or may not be marked on the label itself) can be recorded, with those above the line scored and recorded as a binary"1"and those below scored as a binary "0". Where the line is not marked on the label a fixed reference point is chosen, e. g. a corner of the label, and the line is defined relative to that point by the software.

Either way, the reading can be done using a known image analysis system. This generates a binary code for each label. The code for the label illustrated in Figure 1 would be 10100101010000010100Q. Clearly, ensuring a sufficient number of codes is a matter of statistical choice which is dependent on the number of beads analysed over a given length or region of the label. The maximum number of codes for n beads in a label read as illustrated in Figure 2 is 2n. 40 beads would yield 1.1XI012 combinations; 50 would provide 1. lx1015 combinations.

Varying the sizes of the beads as shown in Figure 2 increases the number of combinations. Another possibility is to read the position of the beads in more detail than just above or below a centre line. For instance, it would be possible to read how many beads are in given zones or, as illustrated in Figure 2, to define two lines and read the beads as indicative of a 0,1 or 2 in a system to give codes of base 3.

It should be noted that normally the software controlling the reading process is adapted to exclude difficult situations where beads overlap or touch as shown at A or B in Figure 1. It can also be adapted to exclude duplicate labels if they occur.

Figure 3 illustrates a further alternative version of the label. In Figures 1 and 2, the position of the beads was recorded essentially two dimensionally, i. e. along and vertically on the label. However, the label 30 could be given a significant depth as shown in Figure 3 and the position of the beads 5 in three dimensions measured and recorded. While this increases the number of codes, it also increases the image processing and software requirements.

It should be noted that the number of beads used should be sufficient to give a sufficient number of unique

codes, but also not so great that the beads cannot be individually read. Too many beads could result in an essentially uniform distribution compared to the resolution of the reading device. So, the number and size of beads should be chosen according to the required application.

Thus in the label (or in the field of view of the reader) the area occupied by the beads should be small, e. g. less than 10%, preferably less than 1%. The beads are spaced apart by distances large compared to their diameter.

It is possible for the beads to be large enough to be read unaided, or small beads could be used and magnification incorporated into the reader. Clearly, the reader can be adapted to scan the labels (reader movement) or the labels can be scanned past fixed readers (label movement).

Of course, rather than translating the position of the beads into codes, it would be possible to use standard image analysis or image grabbing systems with a predetermined definition together with pattern recognition systems. Then it would be the overall pattern which is recorded, the number of codes being dependent on the number of beads and the resolution of the pattern recognition system.

The physical size and shape of the label depends on

the required application. For instance, for use in banknotes, the label could be similar in size to the familiar"metal"strip and could be incorporated into the notes at the time of note manufacture. Alternatively, the tag could be incorporated into a plastic window set in the note, as recently introduced in Australia, for displaying a hologram. The label could be configured to be incorporated into a credit card. Thus, the size of the label and of the beads would normally be considerably smaller than that illustrated in Figures 1,2 or 3.

The label can be incorporated as an integral part of the object to be labelled. For instance, Figure 5 schematically shows a bottle 50 in a defined region 52 of which the beads 5 are incorporated just as in Figures 1 to 3 but with the matrix being the material of the bottle, e. g. glass. Clearly, the beads are chosen to survive the manufacturing process, e. g. to withstand the temperature of molten glass.

When the label is used for identifying banknotes there is no need to read the label at the time of banknote manufacture. This can be deferred until public release of the banknote as described below.

Figure 4 diagrammatically illustrates a security/ validation system using the invention as applied to an

article such as a banknote. The system includes a recording/logging part 40 which includes a reader 41 and processor 43. When the banknote 45 is ready to be released to the public, the label 10 is read by the reader 41 and the code is analysed by the processor 43 and logged in the store 47. As one example, the code can be logged together with the serial or batch number 49 of the banknote.

Thus, the recording/logging part would just be used once for initial logging of the label code. The store 47 is preferably a secure, centralised database.

If, after the banknote has been released, it is desired to check the validity of the note, the label 10 is read by the reader 51, the code calculated by processor 53 (which is analogous to processor 43) and compared with the code and serial number 49 stored in store 47. The valida- tion process of comparing the code and serial number can occur either in the processor at the validation station, or at the store 47. For instance, the processor 53 can send the code and serial number to the store 47 which checks them against the stored numbers and sends back a positive or negative response to confirm authenticity.

Thus this makes it possible to read a unique, uncounterfeitable label or coded region in a product directly in real time (in a matter of seconds or minutes

depending on the nature of the reader and checking system) to give an immediate validation answer. Thus, it is capable of being an immediate detection system for counterfeited items.

In an alternative application of the invention, the manufacturer of the identification labels 10,20,30 could read and log the code on each label together with a serial number of the label (the serial number also being printed on the label) and then a supply of labels could be sent to the end user of them. Such labels could then be used to label objects which the end user wants to monitor. The end user can be provided with a validation unit 50 so that they can check the authenticity of any labelled object as desired.

In a variation, two or more labels according to the invention can be applied to an object and both codes recorded in store 47. One label could be overt and one covert (i. e. hidden). This could give an increased degree of security.

The following example illustrates an embodiment of the invention.

1) Construction Of An Identifier (Label)

The unique tag or label, generally known as an identifier, is formed from a physical system based on the incorporation of a second component comprised of uniformly dyed microspheres obtained from a suitable supplier (such as Bang Laboratories Inc, 9025 Technology Drive, Fishers, Indiana 46038-2886, USA) incorporated into a first component comprised of solidified polyvinyl alcohol obtained from a suitable supplier (such as Aldrich-Sigma Company Ltd, The Old Brickyard, New Road, Gillingham, Dorset SP8 4XT, UK). Specifically, the microspheres may be black polystyrene supplemented with 2% (wwil) divinylbenzene with a mean diameter of 83 ym and a standard deviation of 75 to 90 gm (stock number D0830000PK, Bangs Laboratories Inc). Alternatively the microspheres may be 165 ym mean diameter with a standard deviation of 150 to 180 Hm comprised of polymethylmethacryate and dyed black (stock number D1650000PK, Bangs Laboratories Inc).

A 5% (wv-1) solution of 80% hydrolysed polyvinylalcohol with an average molecular weight of 9,000 to 10,000 daltons (stock number 36,062-7, Aldrich-Sigma Chemical Company Ltd) was prepared in deionized water and warmed gently to dissolve. An aliquot of stock microsphere suspension was prepared by suspending a suitable amount of the microsphere power or suspension in, typically, 1.0 ml of polyvinylalcohol solution and diluted suitably on a x 10 dilution regime with additional polyvinylalcohol to provide

the desired bead density depending on requirements and applications. Examples, typically, 60 yl if the liquid polyvinylalchol and the suspended microspheres were pipetted onto a suitable surface such as"acetate"sheets used with overhead projector systems. The sample was drawn out over the surface using a microscope slide and the disposed sample allowed to cool and dry. The resultant film of solidified polyvinylalcohol were removed from the acetate sheet and used as the tag. It could be sized according to its purpose. The resultant identifier had contained within the solid matrix of polyvinylalcohol, a patten of randomly dispersed microspheres fixed firmly and permanently in position and were able to be maintained in those positions for extended periods of time.

2) A Reader/Recorder System And A Reader Only Examples of optical reader systems are now described.

The same basic systems can be used for both the reader/recorder and the reader-only systems, the difference being that the image information is only accepted for inclusion in the database from a reader/recorder system and not from the reader-only system. The reader-only system can only use the captured information to compete with the data stored in the database.

It was advantageous that the optical systems used to

read the randomly distributed microspheres was telecentric in design ("Optical System Design"by R Kingslake, page 88- 89, Academic Press, 1983,"Modern Optical Engineering by W J Smith, page 131, McGraw Hill, 1966). That is, the system should preferably be one that will reduce the length positional errors and measurement errors that may occur if the images are not exactly in the focal plane of the optical system. In the telecentric example two lenses 61 of suitable focal length were selected separated by a stop 62 to cover then required field of interest on the identifier. For example, a system was constructed (Figure 6) using Spindler and Hoyer 32 2236 100 mm achromatic doublets (Spindler and Hoyer Ltd, 2 Drakes Mews, Crownhill, Milton Keynes, Buckinghamshire MK8 OER, UK) coupled with a suitable CCD camera 63, in this case, a 1/3 inch Computar PMH200 system (Computar Ltd, Computar House, 6 Garrick Industrial Centre, Irving Way, London NW9 6AQ, UK). The output image in the form of a digital image from the camera 63 was captured by a suitable video captivator 65 connected to a suitable personal computer system 67. In this example a Captivator video capture card was used (Videologic Ltd, Home Park Estate, Kings Langley, Hertfordshire WD4 8LZ, UK).

A telecentric design may be of less importance if the required image is constrained in a fixed position, for example, on a roller system or below a transparent platen

surface. In this second example (Figure 7) a simple TV camera lens 71, for example, a Computar 50 mm F1.3 system (Computar Ltd, UK) attached to an extension tube 73 (e. g.

25ML25 manufactured by Comar Ltd, 70 Hartington Grove, Cambridge CB1 4UK, UK) was used. In turn this system was connected to a Computar CHH200 camera 75, Captivator video capture card and personal computer (as described in the first example).

Finally, in a third example, the CCD camera was replaced by a CCD line scan device, for example, a Sony ILX503A (Sony Semiconductors Ltd, The Crescent, Jays Close, Basingstoke, Hampshire RG22 4DE, UK) and the image of the microspheres obtained by moving the identifier relative to the scan device.

The identifier may be illuminated using an indirect light source 69. However, the most favourable illumination was obtained by incorporating a fluorescent dye into either the first component of the identifier or onto the medium immediately below the identifier when it was placed on a suitable surface. This dye was activated by an ultraviolet florescent lamp 69 producing illumination at 365 mm, for example, a Phillips TJ4W/08 (Philips Lighting, City House, London Road, Croydon CR9 3QR). This illumination system dramatically reduced all shadowing effects due to an indirect illumination system, and enhanced the quality of the digital information captured by the viewing system. In turn, the quality of the information captured enhanced the subsequent steps to validate the information captured.

3) Control System For Data Handling, Database Construction And Identifier Comparison The control system used to handle the digital image information generated by the hardware systems described in Section 2 above, consists of the following core elements (Figure 8). The user interface 81 controls and co- ordinates all aspects of the system during normal use, including hardware operation, interfacing with the database management system in order to perform database related operations. The Scanner Control 83 and Converter 85 coordinate the processing of the digital image into a digital code and controls the physical operation of the hardware system. The database Management System 87 stores and retrieves information in the database 89. The user Interface and the Converter form a single module. All elements of the system operate under an industrial standard operating system, such as Microsoft Windows NT.

An example of a bitmap image obtained with particles in the range 75-120 ßm is shown in Figure 9. In Figure 10 the image has been processed to remove the background.

A second set of bitmap images generated using two sizes of particles (small in the range of 75-120 jum, large in the order if 300 ym) are shown in Figures 11,12 and 13.

The particles in this case were graded dust particles. Figure 12 shows the image of Figure 11 processed to remove background, and Figure 13 shows the image processed with an edge detector to produce only outlines of the particles.