Detection of contaminants in raw materials such as sunflower seeds, berries, and certain vegetables like beans, is a well-known problem in the agrifood industry. Ejection systems exist but their detectors are often costly and limited to certain types of food. Furthermore, detecting contaminants with the same shape and/or color as raw materials is often a problem.
We have developed a new hyperspectral imaging camera. This affordable device was designed for easy integration into existing production lines in order to complement or replace existing systems. Furthermore, it allows rapid switching from one type of production to another. Tests on a simulated production line demonstrated efficient detection of contaminants present in sunflower seeds, peanuts (salted and BBQ), and blueberries. The prototype will soon be installed on a peanut/cashew/sunflower seed production line. It will be connected to a contaminant ejection device to determine its efficiency under normal conditions of use.
Applications for this new technology are wide ranging. A feasibility study even found that it may well be possible to determine the concentration of certain antioxidants like tomato lycopene in real time, without any contact with the fruit on the conveyor. In preliminary tests, we were also able to detect fat and bones in chicken breasts as well as bones in fish filets.