ORIGINAL RESEARCH article
Front. Mar. Sci.
Sec. Marine Pollution
Volume 12 - 2025 | doi: 10.3389/fmars.2025.1604055
This article is part of the Research TopicAdvances in Marine Environmental Protection: Challenges, Solutions and Perspectives Volume IIView all 12 articles
A machine learning-based detection, classification, and quantification of marine litter along the central east coast of India
Provisionally accepted
Mallela Pruthvi Raju1
Veerasingam Subramanian2*
V Suneel3
Fahad Syed Asim2
Hana Ahmed Khalil2Mark Chatting2
P Suneetha1
P. Vethamony2- 1Andhra University, Visakhapatnam, Andhra Pradesh, India
- 2Qatar University, Doha, Qatar
- 3National Institute of Oceanography, Council of Scientific and Industrial Research (CSIR), Dona Paula, Goa, India
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
Globally, the growth of plastic production has increased exponentially from 1.5 million metric tons (Mt) in 1950 to 400.3 Mt in 2022, resulting in a substantial increase of marine litter along the coastal region. Presently, there is a growing interest in using an artificial intelligence (AI) based automatic and cost-effective approach to identify marine litter for clean-up processes. This study aims to understand the spatial distribution of marine litter along the central east coast of India using the conventional method and AI based object detection approach. From the field survey, a total of 4588 marine litter items could be identified, with an average of 1.147 ± 0.375 items/m 2 . Based on clean coast index, 37.5% of beaches were categorized as 'dirty' and 62.5% of beaches as 'extremely dirty'. For the machine learning approach 'You Only Look Once (YOLOv5)' model was used to detect and classify various types of marine litter items. A total of 9714 images representing seven categories of marine litter (plastic, metal, glass, fabric, paper, processed wood, and rubber) were extracted from eight field videos recorded across diverse beach settings. The efficiency of the trained machine learning model was assessed using different metrices such as Recall, Precision, Mean average precision (mAP) and F1 score (a metric for forecast accuracy). The model achieved a F1 score of 0.797, mAP@0.5 of 0.95, and mAP@0.5-0.95 of 0.76, and these results show that YOLOv5 model could be used in conjunction with conventional marine litter monitoring, classification and detection to provide quick and accurate results.
Keywords: marine litter, machine learning, Computer Vision, Clean coast index, Andhra
Received: 01 Apr 2025; Accepted: 05 May 2025.
Copyright: © 2025 Raju, Subramanian, Suneel, Asim, Khalil, Chatting, Suneetha and Vethamony. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
* Correspondence: Veerasingam Subramanian, Qatar University, Doha, Qatar
Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.