Scrap metal sorting with colour vision and inductive sensor array

This paper presents a novel automatic scrap metal sorting system which employs a colour vision based optical sensing system and an inductive sensor array. The operation of the system is verified in a real metal recycling plant. The long period test results indicate that 80 % purity can be achieved when the feeding conveyor speed is limited below 1,5 m/s. The described system is not designed for any particular metal. However, the above separation result can only be achieved when reddish (brass, copper) and bright metals (stainless steel) are separated. The properties of aluminium, zinc, and magnesium are too similar for the current sensing principle. The results do not only depend on the sensing system, but also optimal work flow, lighting, dust and vibrations have to be considered in a practical sorting machine. The achieved purity and capacity is sufficient for industrial use. Efficient use of sensor fusion provides good performance despite the diversity of the scrap metals.