Waste Material Classification With a Single-Pixel Dual-Band Ge/Si Photodetector

Fast and easy waste classification is of the uttermost importance for recycling activities. Several industrial systems are available on the market to achieve efficient waste segregation, but state-of-the-art classification methods require expensive and cumbersome instrumentation. Here we propose a low-cost and facile approach to waste identification exploiting a single Ge/Si photodetector with bias-tunable spectral response. We theoretically demonstrate the effectiveness of the proposed system and we provide a preliminary experimental demonstration reaching 95% classification accuracy between seven different waste materials.