Hyperspectral Imaging for the On-line Identification and Classification of End-of-Life Lamps

The lighting market is expected to have revenues of over 100 billion euros in 2020. Despite the evident economical potentialities of this sector, the high products variability, complexity and the presence of materials with potential negative environmental impacts, make lamps recycling processes a challenging task. The Directive 2002/96/EC defines 80% as minimum recovery target for lamps containing mercury. Recovery processes currently adopted for End-Of-Life lamps treatment consist of a preliminary manual phase for sorting different lamp typologies, strictly correlated to the operator ability of recognize the lamps category. The present work proposes a new approach based on the application of hyperspectral vision systems for the automatic identification of End-of-Life lamps, aimed at optimizing recycling processes. These solutions could facilitate the securing of lamps containing hazardous materials, with the aim of maximizing the degree of purity of the recovered key metals and rare earths, stimulating potential secondary raw materials markets.