One important task of food manufacturers is to avoid contamination in their products. This applies, for example, to foods containing nuts, such as mueslis, muesli bars, trail mixes or cookies.
Our sorting machine SAFE-IDENT Sort distinguishes individual spectra of IO and NIO components using a hyperspectral camera and indirect homogeneous illumination. The various types of food to be sorted, such as nuts or almonds, travel at a speed of 150 mm per second along a conveyor belt integrated into the machine. Inside the machine, our specially developed image processing software controls 32 integrated blow-out nozzles, which ensure precise ejection of NIO components such as shell residues, foreign bodies or spoiled parts. The pure IO components are all that remain. Additionally, the system can detect quality flaws on good products – for example, mold, drying or residues on the surface.
Hyperspectral Cameras make the difference
Very fast Analysis
An ordinary RGB camera is based on the three primary colors red, green and blue, just like the human eye. All the colors of human vision can be reproduced using the appropriate mixture of these colors. However, the necessary differentiation between slightly varying shades of brown, as in the case of nuts and their shells, can hardly be realized with this type of camera. By contrast, a hyperspectral camera records a spectrum of 250 spectral bands in the wavelength range from visible to near infrared. This allows the camera to detect individual spectra of light. It can distinguish whether the same shade of brown originates from one wavelength or several superimposed ones.
Since the hyperspectral camera can only record the light spectra emitted by the lighting and reflected by the product to be inspected, special lighting is required. In Safe-Ident Sort, four halogen lamps provide indirect, continuous and uninterrupted illumination for the corresponding light area via a reflector screen. The illumination has a special heat sink which conducts away the heat that is generated.
The nuts to be processed (almonds, hazelnuts, walnuts, cashews, macadamia, peanuts and more) each have uniquely identifiable spectra in their IO parts. The software analyzes the images captured by the hyperspectral camera very quickly and detects the faulty spectra in them, thus uniquely identifying each particle and assigning it to the IO or NIO category.