Abstract:Deep learning techniques have been explored within the marine litter problem for approximately 20 years but the majority of the research has developed rapidly in the last five years. We provide an in-depth, up to date, summary and analysis of 28 of the most recent and significant contributions of deep learning in marine debris. From cross referencing the research paper results, the YOLO family significantly outperforms all other methods of object detection but there are many respected contributions to this field that have categorically agreed that a comprehensive database of underwater debris is not currently available for machine learning. Using a small dataset curated and labelled by us, we tested YOLOv5 on a binary classification task and found the accuracy was low and the rate of false positives was high; highlighting the importance of a comprehensive database. We conclude this survey with over 40 future research recommendations and open challenges.

What problem does this paper attempt to address?

The paper primarily explores the application of deep learning technology in marine debris detection and tracking, and reviews and analyzes related research in recent years. Its core objective is to summarize and evaluate the effectiveness of deep learning methods in addressing the marine debris problem and the challenges faced. Specifically, the paper first introduces the background and severity of the marine debris issue, including the negative impact of debris on marine ecosystems and the potential threat to human health. Subsequently, it provides an overview of the application history of artificial intelligence, especially deep learning technology, in this field, from early attempts to recent research progress. The core content of the paper is an in-depth summary and analysis of 28 recent important studies on the application of deep learning in marine debris detection. These studies cover a variety of methods and technologies, including but not limited to: - **Remote Sensing**: Utilizing satellite images and drone data to detect and track the distribution of debris on the ocean surface. For example, using algorithms such as Support Vector Machine (SVM) and Random Forest (RF) for image classification, and using Synthetic Aperture Radar (SAR) for plastic debris detection. - **Object Detection**: Directly applying object detection technology on the ocean surface or in underwater environments. Research shows that the YOLO series algorithms perform excellently in object detection tasks, but perform poorly on small datasets, highlighting the importance of high-quality databases. - **Other Applications**: Such as using intelligent vision systems for underwater shooting and inspection tasks, and developing methods for automatic identification and counting of marine organisms. By cross-referencing the results of these studies, the authors point out that the YOLO family of algorithms significantly outperforms other methods in object detection. However, they also emphasize the current lack of a comprehensive underwater debris database. Additionally, the authors tested the performance of YOLOv5 on a binary classification task and found that the accuracy was low and the false alarm rate was high, which again underscores the importance of establishing a comprehensive database. Finally, the paper proposes more than 40 future research suggestions and open challenges to promote the development of this field. Overall, the paper aims to provide an effective deep learning-based solution to the marine debris problem and points out the key issues and future directions in current research.

State of the art applications of deep learning within tracking and detecting marine debris: A survey

YoloXT: A Object Detection Algorithm for Marine Benthos

Harnessing Deep Learning for Underwater plastic Trash Identification

MDM-YOLO: Research on Object Detection Algorithm Based on Improved YOLOv4 for Marine Organisms.

Survey on Deep Learning-Based Marine Object Detection

An Efficient Deep-Sea Debris Detection Method Using Deep Neural Networks

Research Challenges, Recent Advances, and Popular Datasets in Deep Learning-Based Underwater Marine Object Detection: A Review

Research Challenges, Recent Advances and Benchmark Datasets in Deep-Learning-Based Underwater Marine Object Detection: A Review

YOLOv7t-CEBC Network for Underwater Litter Detection

Is the use of Deep Learning and Artificial Intelligence an appropriate means to locate debris in the ocean without harming aquatic wildlife?

Instant Deep Sea Debris Detection for Maneuverable Underwater Machines to Build Sustainable Ocean Using Deep Neural Network

Deep Learning-Based Visual Detection of Marine Organisms: A Survey

Deep Neural Networks for Marine Debris Detection in Sonar Images

Research on the Identification and Classification of Marine Debris Based on Improved YOLOv8

Marine Litter Detection Based on YOLOv7 Algorithm and Data Encryption Protection.

Marine Debris Detection Model Based on the Improved YOLOv5

YOLOv8-C2f-Faster-EMA: An Improved Underwater Trash Detection Model Based on YOLOv8

Deep-Sea Organisms Tracking Using Dehazing and Deep Learning

MLDet: Towards Efficient and Accurate Deep Learning Method for Marine Litter Detection

Deep Learning Innovations for Underwater Waste Detection: An In-Depth Analysis

An Experimental Study on Marine Debris Location and Recognition using Object Detection