Machine vision is a useful tool for detecting defects and controlling the dimensions of agri-food products on a conveyor. Today, hyperspectral imaging (HSI) is emerging to analyse the composition. However, the HSI instruments currently available are not equipped for high-speed food production or new practices in precision agriculture.

The objective of the HS2I project is to combine HSI with high-speed imaging to answer these unmet needs. This involves developing a new type of camera.

In the food industry, HS2I will enable the use of HSI technology for fast, reliable and fully automatic food monitoring – including detection of foreign bodies and management of cooking level, moisture content and product composition.

In precision agriculture, HS2I will benefit ground-based applications, where there is a need for advanced, high-speed detection of plant diseases and product defects, for example.

Lead SME Photon Lines is collaborating with First Light Imaging, which manufactures high-performance cameras, and puff pastry manufacturer, Geldelis.

Final report summary
A high-speed (up to 3000 hz) hyperspectral camera – eyeSPID – has been developed. This is both the first HSI camera to reach such a high speed and the first to be designed and manufactured in Europe. To complement this system, full-featured HSI software – eyeSPICE – has been developed, allowing the control and management of the camera setup and calibration; the hyperspectral acquisition; HSI data visualisation, exploration and analysis; the development of HSI interpretation models; and the online monitoring process. To our knowledge, very few HSI software solutions on the market cover such a wide range of functions.

There is still some work to be done to reach the longer-term objectives. Photon Lines will have to optimise the lighting system associated with the eyeSPID camera to meet the most demanding requirements of food producers like Geldelis. Other versions of the camera are also under development.