Marine Debris Detection in Satellite Surveillance using Attention Mechanisms

Ao Shen,Yijie Zhu,Richard Jiang

2023-07-09

Abstract:Marine debris is an important issue for environmental protection, but current methods for locating marine debris are yet limited. In order to achieve higher efficiency and wider applicability in the localization of Marine debris, this study tries to combine the instance segmentation of YOLOv7 with different attention mechanisms and explores the best model. By utilizing a labelled dataset consisting of satellite images containing ocean debris, we examined three attentional models including lightweight coordinate attention, CBAM (combining spatial and channel focus), and bottleneck transformer (based on self-attention). Box detection assessment revealed that CBAM achieved the best outcome (F1 score of 77%) compared to coordinate attention (F1 score of 71%) and YOLOv7/bottleneck transformer (both F1 scores around 66%). Mask evaluation showed CBAM again leading with an F1 score of 73%, whereas coordinate attention and YOLOv7 had comparable performances (around F1 score of 68%/69%) and bottleneck transformer lagged behind at F1 score of 56%. These findings suggest that CBAM offers optimal suitability for detecting marine debris. However, it should be noted that the bottleneck transformer detected some areas missed by manual annotation and displayed better mask precision for larger debris pieces, signifying potentially superior practical performance.

Computer Vision and Pattern Recognition

What problem does this paper attempt to address?

The paper aims to address the issue of marine debris detection, particularly in locating marine debris in satellite surveillance images. Current methods for locating marine debris are limited, inefficient, and have a narrow scope of application. To improve detection efficiency and applicability, researchers attempted to combine YOLOv7 instance segmentation with different attention mechanisms to explore the optimal model. By using an annotated satellite image dataset containing marine debris, the study compared three attention models: lightweight coordinate attention, CBAM (Convolutional Block Attention Module), and self-attention-based bottleneck transformers. Experimental results showed that CBAM performed best in box detection evaluation (F1 score of 77%), while coordinate attention and YOLOv7/bottleneck transformers had F1 scores of approximately 71% and 66%, respectively. Mask evaluation also showed CBAM leading with an F1 score of 73%, with coordinate attention and YOLOv7 performing similarly (around 68%/69%), while the bottleneck transformer had an F1 score of only 56%. This indicates that CBAM has the best applicability in detecting marine debris, but the bottleneck transformer shows potential superiority in detecting certain manually annotated missing areas and the mask accuracy of larger debris.

The paper also discusses the hazards of marine debris to the environment and ecosystems, including pollution of water appearance, impacts on marine ecosystems, and health risks to humans through the food chain. Existing mainstream detection methods such as traditional manual methods, image analysis from drones and onboard cameras, and satellite imagery analysis all have their limitations. To address these issues, researchers proposed using YOLOv7 instance segmentation and innovative attention mechanisms to enhance the efficiency and accuracy of marine debris detection. Specifically, the study used high-resolution satellite images and improved the ability to identify and classify marine debris through the application of instance segmentation and different attention mechanisms, especially in the context of satellite images.