Laser identification

Counterfeit and smuggled goods are said to be one of the fuels that drive organised crime, so it is essential that products can be identified to determine if they are genuine and where they came from. Some manufacturers go to extreme lengths to incorporate hard to replicate anti-counterfeit labels or devices into their products, but it's a game of catch up and it isn't long before the criminals find a way of defeating those measures. So how about using no anti-counterfeit measures at all? This is the approach being promoted by Russell Cowburn at Imperial College in London, who is researching a method called Laser Surface Authentication (LSA), which relies on unique, microscopic identifiers already inherent in products or their packaging.

Fingerprints

In a recent review paper¹, Cowburn gives a thorough explanation of how LSA works with examples of how it might be used in practice, providing far greater detail than was possible in the Nature paper in which he and his colleagues reported their work in July of 2005². Plastic, paper and cardboard surfaces are some of the materials on which they have demonstrated their technique. These surfaces possess unique and entirely random characteristics made up from the tiny lumps and bumps that are only apparent when viewed on the microscopic scale. Individual surfaces can be identified from these features by reading them with a laser.

The technique relies on the phenomenon of laser speckle, an effect that can be seen when an expanded laser beam - even one from a simple laser pointer - is projected onto an opaque surface. The resulting pattern does not appear homogeneous across the surface but instead looks like a random collection of dark and bright spots. This is due to coherence, a property of laser light which means the individual light waves are all in phase, like soldiers marching in step. When laser light is reflected from a surface containing minute imperfections, some of the light waves will still be in step, while other will be out of step. Think of the surface like the Himalayas: some of the light reflects from the mountain tops and some from deeper down in the valleys. The dark and bright spots in a speckle pattern are the result of these phase relationships which occur on reflection.

The speckle pattern of common items, such as office paper, cardboard and plastic cards, possess a unique fingerprint which can be used for identification purposes. This fingerprint, or signature, can be read by scanning the item with relatively unsophisticated hardware: a low power laser, which is focused onto the sample surface, and four photo-detectors, each positioned to detect the light scattered at different angles. These components are assembled into one device which is scanned across a surface - perhaps only a small section of it - detecting and recording the diffusely reflected light from the laser at each position. These data are digitised so that dark and bright readings are converted into either ones or zeros; these digital sequences make up the signature of an item. According to the author, it is possible to read the same signature for an item regardless of which scanner is used, assuming the scanners are of the same design and construction.

The surprising thing is that this technique works at all. The short wavelengths of light are such that the smallest of displacements can change the laser speckle pattern. The patterns are sensitive to movement of the laser, vibrations of the surface onto which it is focused, or refractive index changes in the intervening space. One might reasonably expect that slight displacements in the position of the same item beneath the scanner, from one measurement to the next, would result in vastly different signatures each time, but it turns out that the technique is not as fickle as one might expect.

Items must be placed under the scanner in approximately the same position each time, but there is a certain amount of tolerance for misplacement which makes the technique practical for real world applications, obviating the need for micro-positioning tools. To counteract the effects of misplacing the same item between scans, the digital sequences which make up the signatures are effectively moved with respect to each other during the comparison process, in an effort to obtain the best possible match. If the probability of a match is high, the result from the matching algorithm will identify it as the same item.

Methods of identifying an item can be generalised into two categories: one-to-one matching and one-to-many. LSA can be used in both ways. An example of the former would be when the signature from a scanned item, such as a passport, is compared with the one that is on file for that document, in order to see if it is genuine. An example of the latter case would be when a package in a warehouse is scanned and the signature is ran through a database to find a match, thereby obtaining further information about it. Because it is so highly improbable that LSA signatures from two different items could ever match (a false positive), due to the complex and random nature of the identifiers, the technique could be used confidently, even with the largest of databases.

One problem is wear and tear. After all, paper packaging can be soiled, credit cards scratched and the family dog might chew on your homework. This system might not withstand everything, but the matching algorithm can tolerate a good deal of mechanical and chemical damage which might occur between scans. In experiments where office paper was either scorched, crumpled, scribbled on, or soaked in water, the before and after comparisons identified the items as being the same, with the probability that it could be a false positive match being at most 1 in 10¹⁸ (1 followed by 18 zeros - that's a big number!).

Laser Surface Authentication may lead to obvious applications such as in credit card or passport security, for providing proof of their authenticity. It would be impossible to replicate a signature of this type with any existing technology, due to the complexity of the identifier, so forging a passport might become a thing of the past. LSA might also become a low cost solution for tracking inventory, enabling products to be identified and their history traced, without ever having to individually label them with serial numbers or bar codes. Items could also be tracked covertly because the identifier is an inherent part of its surface and need not be explicit, which would be another benfit this technique offers in the ongoing battle against smuggling and the trade in replica goods.

1) Russell Cowburn (2008). Laser surface authentication - reading Nature's own security code Contemporary Physics, 49 (5), 331-342 DOI: 10.1080/00107510802583948

2) James D. R. Buchanan, Russell P. Cowburn, Ana-Vanessa Jausovec, Dorothée Petit, Peter Seem, Gang Xiong, Del Atkinson, Kate Fenton, Dan A. Allwood, Matthew T. Bryan (2005). Forgery: ‘Fingerprinting’ documents and packaging Nature, 436 (7050), 475-475 DOI: 10.1038/436475a