We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Hybrid Deep Learning Approach for Marine Debris Detection in Satellite Imagery Using UNet with ResNext50 Backbone
Summary
Despite its title referencing marine debris detection, this paper develops a deep learning computer vision model for identifying marine debris in satellite imagery using a UNet architecture with a ResNext50 backbone — not a study of microplastic pollution itself. It is a remote sensing and machine learning engineering paper, and while the technology could support large-scale ocean plastic monitoring, the paper does not directly examine microplastics or their health effects.
Marine debris is persistent solid stuff in the water. Oceans include several varieties of organic marine debris, but massive levels of man-made marine trash threaten their biological equilibrium. Manually scanning the ocean for garbage is time-consuming and inefficient, making it uneconomical. Deep learning, which is more efficient than manual methods, is used to detect marine debris in satellite imagery in our work. Deep learning algorithms have been successful in semantic segmentation, however marine debris detection using satellite imagery has been underexplored. The lack of comprehensive marine debris datasets until recently and the complexity of multispectral satellite photos are to blame. Our segmentation method using the UNet architecture and a ResNext50 backbone exceeds the existing state of the art on the Marine Debris Archive Dataset (MARIDA), a dataset of 11 band sentinel 2 Satellite image patches. The hybrid solution combines ResNext50's increased feature extraction with UNet's global and local context preservation, which is crucial in satellite photos of floating bodies due to marine debris' movement pattern. We achieved benchmark mean pixel accuracy, IoU, and F1 scores. We achieved an 88% recall, a 10% improvement over the state of the art, in categorizing marine trash pixels in photos. This work attempts to advance deep learning algorithms for remote sensing and move closer to cleaner oceans.
Sign in to start a discussion.
More Papers Like This
Deep-Feature-Based Approach to Marine Debris Classification
This study applied deep learning to classify marine debris from images, demonstrating that feature-based neural network approaches can effectively distinguish plastic types and other debris categories to support automated ocean monitoring.
A Deep Learning Model for Automatic Plastic Mapping Using Unmanned Aerial Vehicle (UAV) Data
Researchers applied a deep learning semantic segmentation model (ResUNet50 based on U-Net architecture) to UAV orthophotos to automatically map floating plastic debris, achieving F1-scores of 0.86-0.92 for specific plastic types including oriented polystyrene, nylon, and PET. Classification accuracy decreased with lower spatial resolution, with 4 mm resolution providing optimal performance for distinguishing plastic types.
An Image Analysis of Coastal Debris Detection -Detection of microplastics using deep learning-
Researchers developed a deep learning-based coastal debris detection system using YOLOv7 and the SAHI vision library to identify microplastics in image data collected from shorelines. The system demonstrated effective detection performance and offers a scalable approach for automated monitoring of microplastic litter in coastal environments.
Enhancing marine debris identification with convolutional neural networks
A deep learning model was developed to identify and classify marine debris components captured by underwater remotely operated vehicle imagery, addressing the challenge of widely distributed ocean waste including microplastics. The convolutional neural network demonstrated improved accuracy for debris detection and classification compared to conventional image analysis methods.
On advances, challenges and potentials of remote sensing image analysis in marine debris and suspected plastics monitoring
This review evaluates the current state of satellite and aerial remote sensing for detecting marine plastic debris, noting that while progress has been made using optical and hyperspectral imaging, significant challenges remain including low detection resolution for small particles, confusion with other floating materials, and the need for better machine learning algorithms. The paper is relevant to the microplastics field as large-scale monitoring tools are needed to track plastic pollution distribution and inform cleanup and policy efforts, though direct detection of microplastics (<5 mm) from orbit remains largely out of reach with current technology.