Whilst size, shape, colour, and surface texture of food can be easily measured by conventional machine vision systems, the prediction of many other parameters, such as fat, sugar and moisture content, protein, material homogeneity, or materials hazardous to the consumer can all be detected quickly and simultaneously by hyperspectral imaging systems.
HSI implementation into food production systems also enables accurate high speed sorting. The spectral signature (much like a human "fingerprint") of the material, combined with the spatial information, allows excellent determination of substandard product, contamination, presence of foreign bodies, as well as defects by colour, shape and size.
Our advanced machine learning and classification algorithms built into our
software then allow discrimination methods to be set up for the process line and the output can be applied in real time to interface to gravity, paddle separators, or robotic arm pickers etc.