Selection for Reuse in WEEE Reverse Logistics with Text-Based Model Identification

Abstract

Recupel, the producer responsibility organisation for Waste Electrical and Electronic Equipment (WEEE) in Belgium, is actively exploring novel approaches to increase the number of reused EEE by optimising their reverse logistics operations. Today, repair and reuse organisations are allowed to manually select devices eligible for reuse at several distribution centres contracted by Recupel. In this article, first, the challenges and limitations of this selection procedure are discussed based on stakeholder interviews. Then, the SmartRe technology, developed in prior research to automatically identify the product model by capturing a picture of the label and comparing text information with a database is introduced. The benefits of the incorporation of this technology in Recupel’s reverse logistics are discussed. Finally, a quantitative study is presented on two annotated datasets collected at one of Recupel’s contracted distribution and sampling centres. This dataset comprises 571 product label images of six product categories that are often repaired and resold, namely coffee machines, vacuum cleaners, microwaves, irons, refrigerators and washing machines. An overall matching efficiency of 69.5% was achieved by comparison with a database of 75k product models, which shows that the proposed strategy is a promising tool to more systematically perform and to scale up selection for reuse.