Personal documents, such as passports, IDs, and driver’s licenses, are extremely important in the political, social, and economic aspects of life because they allow to determine a person’s identity. Counterfeiting documents enables a variety of crimes: from underage drinking to financial crimes, human trafficking, and terrorism. Therefore, there is a need for tools that allow to determine a document’s authenticity confidently . One possibility is hyperspectral imaging (HSI), which is already used in many different fields, where a non-invasive analysis is required. Unfortunately, HSI systems are bulky, expensive, and difficult to operate.
In this work, we present a slightly different approach to hyperspectral imaging in which a standard smartphone camera in combination with a multicolour illumination system is applied. Such a system would be affordable, user-friendly, and portable for document examiners while ensuring proper document authentication. Furthermore, this system could be adapted to be used in other fields that take advantage of HSI. The aim of the work was to develop an MCU-controlled multicolour flash prototype. In the process, sensitivity curves and other camera properties of a commercial smartphone were investigated. Moreover, an enhanced colour discrimination algorithm was developed and tested on different samples. DNG filetype data analysis tools were employed and tailored for automated data analysis and hyperspectral imaging.
To conclude, we have developed a prototype hyperspectral imaging system based on LEDs which can be used for document authentication and other fields of potential applications such as medical diagnostics.