Banknote inspection with JAI Multispectral Cameras

Banknote inspection with JAI Multispectral Cameras

Digital payment has become more popular than ever over recent years due to the global COVID19 pandemic. This has resulted in far fewer cash payments, yet money is still printed on a regular basis. Digital payments offer a completely different levels of security when compared to physical cash payments; it’s very difficult to forge digital money.

How does this differ from cash?

banknotes are produced across several production sites and must be of high quality – or else they simply wouldn’t last when implemented into circulation. More importantly, they must have, “top dollar” security features to uphold their value and prevent counterfeits.

There are hundreds of thousands of counterfeit banknotes in circulation… Typically, these numbers are largely made up by few sources. One of the main counterfeited currencies globally is the U.S. dollar (USD). This is, for the most part, because it is distributed and used all over the globe while having relatively low security standards. In contrast, the Swiss franc (CHF) is one of the toughest currencies to counterfeit. Less than 1000 counterfeited banknotes with a total of 220 thousand CHF were reported in 2019 for example.

How can Machine Vision reduce counterfeit currencies?

Automated visual inspection methods can verify most security features on physical currencies. Unique serial numbers encoding the origin of banknotes are widely used as one of these features, which can be checked by using a more traditional RGB/MONO camera.

Further inspection can be carried out with RGB and MONO cameras, including colour, size, print quality, labelling, and OCR/OCV inspection. However, there are limitations to these cameras – and the largest of which is the cameras sensitivity to UV and IR wavelengths.

What is a Multispectral Camera?

A multispectral camera uses more than one sensor to capture data from different wavelengths.

Sensor arrangements can be Red, Green, Blue and Near-IR. This 4 chip sensor, when compared to a standard RGB camera which uses a Bayer pattern, will generate far more accurate data, and it will also allow for advanced reading of colour, anti-aliasing, denoising and sharper image quality.

The sensors are split out inside the camera body and are placed at different angles in order for the process to operate correctly and accurately. When light first enters the camera it hits a specially designed prism causing the wavelengths to be split apart, hence, causing these wavelengths to exit the prism at the correct angle and direction – towards the correct sensor.

Multispectral cameras are also able to conduct security checks not visible to the human eye. This is due to its ability to see non-visible wavelengths. By using a single device for quality inspection, the setup of the system is very likely to be more effective, robust, and cost-efficient.

To learn more about multispectral imaging and how it can improve your security checks, download the white paper.

Download White Paper