Artificial Intelligence-Powered Solid Waste Detection System (#669)

Authors

Granda Cruz, Luis Enrique

Agurto Marchena, Fernando Jair

Abstract

Efficient solid waste treatment is a growing challenge for ever-expanding urban populations. One innovative solution is the application of artificial intelligence (AI), which seeks to improve the detection and categorization of waste in population settings. The aim of this review is to evaluate the effectiveness of AI algorithms, especially Deep Learning models, in the automated identification of urban solid waste in different population contexts. Using the PICO and PRISMA methodologies, a comprehensive research was carried out in scientific repositories such as Scopus, Dialnet and Redalyc. The keywords used included "solid waste", "AI", "Deep Learning", "solid waste" and "waste management". The findings suggest that Deep Learning-based models show promising performance in the detection and classification of solid waste in various urban settings, with accuracy rates exceeding 90% in some cases. It was observed that the effectiveness of these systems is highly dependent on the quality and diversity of the preparation data sets, which must reflect the variability of waste in different types of populations. The review concludes that AI has significant potential to achieve optimisation of solid waste treatment in local environments, although further research is required to address challenges such as variability in lighting conditions and sorting in complex environments.