On the Acoustic Difference of Metal and Glass When Falling on an Impact Plate

We are interested in the separation of metal and mirror/glass components in recycling operations. Conventional optical separation of glass/mirror and metal parts lacks accuracy due to the reflective properties of glass/mirror. This is especially true for metal-coated mirrors, where even hyperspectral imaging fails. A cheap and effective way to separate these materials is to drop the parts onto an impact plate and record the resulting sounds with microphones. In this paper, we analyze the acoustic emissions from the impact of the materials mirror and metal. The results should provide a basis for further processing of the data via machine learning.