Deep learning-based Methods for Detecting and Quantifying floating litter in Riverine Environments
T. Jia (TU Delft - Water Systems Engineering)
Riccardo Taormina – Copromotor (TU Delft - Water Systems Monitoring & Modelling)
Z. Kapelan – Promotor (TU Delft - Water Systems Engineering)
More Info
expand_more
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Abstract
Litter, particularly plastic, accumulating in water bodies is a challenging environmental issue that affects ecosystems, human health and the economy. Rivers are the main pathways of land-based plastic waste to the ocean, but they also act as potential temporary and long-termplastic sinks, where significant amounts of plastic waste accumulate, and even remain trapped for decades. The detection and quantification of floating litter in rivers and urban waterways is thus essential for evaluating pollution levels and informing mitigation actions. However, traditional monitoring methods, such as sampling with nets and booms, are not suitable for large-scale structured monitoring across multiple geographic locations in extensive river systems. Deep Learning (DL) methods have shown great promise in automatic detection and quantification of floating litter from images or videos. Given that this specific field is still in its early stages, this thesis aims to enhance the understanding of DL-based litter detection and quantification in riverine environments, identify key knowledge gaps, and explore methodologies to address these gaps and drive further advancements in this field....