Detection of Urban Residues Using Deep Learning Algorithms

Waste management in urban areas is an essential key to environmental sustainability and maintenance, as well as public health. The rapid growth of cities and the population has led to an increase in waste generation and has become a global challenge, as well as an interesting field for research and adoption of innovative technologies is essential since many aspects of waste management are still manually executed in cities world-wide. This manual approach comes with typical limitations like human error, fatigue from repetitive tasks, and economic costs. This paper proposes an innovative system leveraging computer vision and artificial intelligence, specifically neural networks, to automate the detection of waste management anomalies in urban areas. This is aimed at improving efficiency in the collection stage. The study also presents tests carried out in real environments, analyzing the results obtained and the process of improvement in the detections. The uniqueness of this work lies in the specific application of deep learning methodologies for waste anomaly detection, which presents a significant step towards automation in urban waste management.