International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 11 Issue: 02 | Feb 2024
p-ISSN: 2395-0072
www.irjet.net
Marine Plastic Debris Detection Rishitha Reddy Gudipati , Sruthika Bachu , Dr. Narayana Rishitha & Student, Dept of AI&ML, Chaitanya Bharathi Institute of Technology Sruthika & Student, Dept of AI&ML, Chaitanya Bharathi Institute of Technology Dr. Narayana & Asst. Professor, Dept of AI&ML ,Chaitanya Bharathi Institute of Technology ---------------------------------------------------------------------***--------------------------------------------------------------------commonly in the form of food containers, such as plastic Abstract - The health of oceans and marine ecosystems is seriously threatened by marine plastic trash. Every year, at least 14 million tonnes of plastic enter the ocean. At now, plastic waste dominates the marine litter population, accounting for 80% of all debris found in the ocean, ranging from surface waters to deep-sea sediments. Every continent has plastic beaches, with higher concentrations of plastic debris found close to well-known tourist attractions and densely populated places. Finding and keeping an eye on plastic debris in marine habitats is essential to creating mitigation plans that work and protecting aquatic ecosystem biodiversity.
bags and bottles, and packaging materials. The other 20% stems from shipping vessel discharges and discarded commercial fishing gear . Studies have shown that removing plastic from the oceans will exponentially benefit the ecosystems. This includes the prevention of the movement of invasive species between regions, the prevention of its degradation into micro-plastics , and the decrease in emissions of greenhouse gases (thereby decelerating climate change) ].
Keywords— Yolov3, Yolov4, Yolov5, object detection, deep
[11]The research paper by Hipolito et al. published in 2021 delves into the pressing issue of marine debris and proposes a solution using machine vision, specifically the YOLOv3 method, to detect underwater marine plastic waste. The authors effectively contextualize the environmental significance of the problem and advocate for technological advancements to mitigate its impact. The methodology is robust, employing a well-structured approach involving dataset preparation, image annotation, deep transfer learning algorithms, and data augmentation. The dataset utilized in this inquiry came from the Data Repository for the University of Minnesota (DRUM), specifically the Dataset of Underwater Trash. There are 8580 pictures in the collection, split into two categories: bio images (4290) and non-bio images (4290). The proponents chose a small sample size of 300 non-bio images. The chosen image dataset was used to construct training and validation sets. The training dataset featured 80% of the data, while the validation set only had 20% of the images. Results and discussions comprehensively present training, validation outcomes, and model evaluation using mean Average Precision (mAP). The study concludes by emphasizing the potential of machine vision in addressing marine plastic debris, highlighting an impressive mAP of 98.15% for model 19. Overall, the paper
2. LITERATURE SURVEY
learning, transfer learning, and marine plastic detection
1.INTRODUCTION Marine plastic debris (MPD) is a major environmental problem with devastating impacts on marine life, beaches, tourism, and fishing. Detecting MPD is essential for effective monitoring and management, but traditional methods are time-consuming and labour-intensive. New technologies for automated MPD detection are being developed, including image processing, machine learning, and sensor-based technologies. Each technology has its own advantages and disadvantages, but all have the potential to revolutionize MPD detection. MPD detection technologies can be used for a variety of purposes, including monitoring pollution levels, tracking the movement of MPD, and assisting with cleanup efforts. By developing more accurate, robust, and scalable MPD detection technologies, we can reduce the impact of MPD on the marine environment. 1.1.
PROBLEM STATEMENT
Plastic pollution poses an imminent threat to the marine environment, food safety , human health, eco-tourism, and contributes to climate change . Global plastic production has exceeded 500 million tons of plastic, and projections indicate that 30% of all produced plastic will end up discarded in the oceans. Researchers have documented a five-fold increase in plastic debris within the Central Pacific Gyre and have shown that plastic pieces now outnumber the native plankton 6:1 in terms of abundance . A significant amount of marine plastic (about 80%) originates from land-based sources : Most
© 2024, IRJET
|
Impact Factor value: 8.226
effectively communicates the gravity of the issue and demonstrates the efficacy of YOLOv3 in automated marine debris detection. [10]An article by Nur Athirah Zailan and Anis Salwa Mohd Khairuddin published in 2021 effectively tackles the pressing issue of plastic debris pollution in riverine environments by proposing a YOLOv4-based algorithm for debris detection. It begins with a strong introduction,
|
ISO 9001:2008 Certified Journal
|
Page 740