Scientists make DIY multispectral imager for fake detection

Researchers in Colombia have created a multispectral imaging system, based on components that are available off the shelf, that could be an alternative to commercial equipment costing thousands of dollars.

Called MEDUSA, the system is based on a web camera, light-emitting diodes, a semi-spherical dome for even lighting, an open-source Arduino development board and free Python programming software.

The setup was put through its paces using applications for which commercial multispectral imagers may be used – including analysing ink changes on a counterfeit banknote – and according to the researchers was able to produce spectral information that was invisible to the naked eye and backed its potential in these uses.

"MEDUSA can be freely reproduced and customised from this research, making it a powerful and affordable analytical tool to analyse a wide range of subtle chemical properties in samples," said the team from National University of Medellin, who have published the work in the journal HardwareX, including detailed build and operating instructions.

The device was able to detect general surface ink differences between fake and authentic 20,000 Colombian peso bills that were not easily noticeable with standard RGB imaging.

Multispectral imaging – which can detect light reflectance from different spectral bands of visible and nonvisible wavelengths – is becoming increasingly practical for counterfeit detection because it is a non-destructive technology, keeping a sample intact for further investigation.

It has become increasingly used in the pharmaceutical and food industries for quality control as well as other analytical applications, including counterfeit detection, and has also started to find applications in the clothing and currency sectors.

"If properly calibrated, [the Medusa] device might find use for the non-invasive analyses on areas like food quality, counterfeit analysis, microbial colony phenotyping, plant chemical analysis, [and] preliminary spectral analyses on multiple surfaces," write the authors.