AI-Driven Detection of Waste Hotspots in Surface-Water Bodies: A Comprehensive Review

DOI

10.2991/978-94-6239-723-1_20How to use a DOI?

Keywords

Floating litter; riverine plastics; surface-water monitoring; Sentinel-2; SAR; UAV imagery; CNN; YOLO; transformer models; hotspot mapping

Abstract

Any form of floating debris in our waterways within rivers, lakes and reservoirs is an increasing environmental concern that has vast ecological and social economic consequences. We have witnessed recent advances in remote sensing and artificial intelligence that offers us the opportunity of automatic detection and mapping of floating waste hence we can accomplish large scale near real time water quality assessment. In the review, we consider articles that were conducted between 2019 and 2025 that employed optical satellite data, UAV data recorded by RGB sensors, hyperspectral sensors, SAR data and deep learning systems to the problem of debris detection and hot spot delineation. In this case, a structured selection process was employed, which took into account sensor selection, algorithmic strategies as well as reported results of the evaluation. According to the literature, high-resolution images of the UAV and multispectral satellite missions such as Sentinel-2 are the most suitable to combine time frequency and spatial characteristics. The deep-learning approaches (CNN classifiers, YOLO series detectors, and transformer systems), tend to perform highly, compared to the classical machine-learning and spectral-index approaches, in the detection of small and fine-scale debris. Even though SAR and hyperspectral data respond well in the clear atmosphere, as well as in turbidity and low light conditions, they have not been used in the study of inland waters. The metrics of precision and recall and F1 and the use of Streamlit were chosen in the selection of the set of data, augmentation, and training to CNN/transfer learning and transformer, basing our experimental pipeline on previous studies. Finally, the literature review makes recommendations of how the operational and scalable systems can be developed to identify the hotspots of debris.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Nishchay Malhotra
AU  - Sheetal S. Patil
AU  - Amit Kumar
AU  - Vishal Menia
AU  - Kanav Dogra
AU  - Nishchal Singh Bishnoi
PY  - 2026
DA  - 2026/07/14
TI  - AI-Driven Detection of Waste Hotspots in Surface-Water Bodies: A Comprehensive Review
BT  - Proceedings of the International Conference on Responsible, Risk-aware, and Regulated AI (RRRAI 2026)
PB  - Atlantis Press
SP  - 216
EP  - 227
SN  - 1951-6851
UR  - https://doi.org/10.2991/978-94-6239-723-1_20
DO  - 10.2991/978-94-6239-723-1_20
ID  - Malhotra2026
ER  -