Characterization of fine metal particles using hyperspectral imaging in automatic weee recycling systems

Abstract

Waste from electric and electronic equipment (WEEE) represents the fastest growing waste stream in EU. The large amount and the high variability of electric and electronic products introduced every year in the market make the WEEE recycling process a complex task, especially considering that mechanical processes currently used by recycling companies are not flexible enough. In this context, hyperspectral imaging systems (HSI) can represent an enabling technology able to improve the recycling rates and the quality of the output products. This study shows the preliminary results achieved using a HSI technology in a WEEE recycling pilot plant, for the characterization of fine metal particles derived from WEEE shredding.

Characterization of Fine Metal Particles Derived from Shredded WEEE Using a Hyperspectral Image System: Preliminary Results

“…Considering these results, it can be reasonably stated that methods that take into account the spectral correlation through a covariance matrix, such as the MhlD and MD algorithms, applied to compensated reflectance spectra ( ) are reliable and efficient methods for the classification of metal particles derived from shredded WEEE. The accuracy assessment presented in this paper was in overall agreement with the previous results achieved for the same topic, and presented in [ 13 ]. The lower results in pixel-wise classification were somehow expected, since the introduction of new classes tends to increase classification errors, especially when some materials show similar spectral signatures within the considered VNIR spectral range.…”

“…The accuracy assessment presented in this paper was in overall agreement with the previous results achieved for the same topic, and presented in [13]. The lower results in pixel-wise classification were somehow expected, since the introduction of new classes tends to increase classification errors, especially when some materials show similar spectral signatures within the considered VNIR spectral range.…”

“…This work features an extension of the results obtained in [ 13 ], introducing a new dataset, which increases the number of metals to be characterized. In particular, this study evaluates the capability of the VNIR hyperspectral camera to characterize mixtures composed of particles from five different metals: besides brass (CuZn), iron (Fe) and copper (Cu), particles of aluminum (Al) and nickel (Ni) were added to the dataset.…”

“…In Candiani et al [ 13 ], the authors tested the performance of an HSI technology for WEEE sorting, comparing several combinations of (i) classification algorithms and (ii) methods for the compensation of illumination conditions. The latter is a very crucial point since such conditions highly affect the acquired signal of metal particles, as explained in the following sections.…”

Highly Evolvable E-waste Recycling Technologies and Systems

“…Secondly, the economic assessment relied on data coming from literature and on the estimation of some costs derived from experimental tests and pilot infrastructure investment, which could differ from real industrial environment conditions. Further details on this research are reported in [31,[45][46][47].…”