We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Transformer Assisted U-Net for Marine Litter Detection on Sentinel-2 Imagery
AI summary Read the abstract
This study presented a deep learning scheme for detecting marine litter in Sentinel-2 satellite imagery using a Deep UNet architecture enhanced with self- and cross-attention transformer mechanisms in the decoder, incorporating all Sentinel-2 bands alongside NDVI and FDI indices. The model was trained on the FloatingObjects benchmark dataset and compared against state-of-the-art approaches for marine debris segmentation from space.
The contamination of marine environments with man-made litter is a growing nation-wide concern. Satellite imagery combined with deep learning–based detection models has emerged as a robust and cost-effective solution for large-scale marine litter monitoring. In this article, we present a novel deep learning-based scheme to detect marine litter using Sentinel-2 imagery based on the Deep UNet architecture, introducing self- and cross-attention mechanisms into the decoder via transformer layers. The model leverages all Sentinel-2 bands except B10, and the NDVI and FDI indices are additionally incorporated to better guide the segmentation process. To evaluate the proposed model, we train it on the FloatingObjects dataset, a widely used benchmark for marine debris detection, and compare its performance against state-of-the-art approaches.
More Papers Like This
Marine Debris Detection in Satellite Surveillance using Attention Mechanisms
AI summary Read the abstract
This research develops attention-based deep learning models to automatically detect marine debris, including plastic waste, in satellite imagery. By leveraging large-scale remote sensing data and advanced neural network architectures, the approach enables scalable monitoring of ocean pollution across areas that would be impractical to survey manually.
Synergistic integration of vision transformers and advanced segmentation algorithms for panoptic mapping of marine litter
AI summary Read the abstract
A panoptic segmentation approach combining vision transformers with instance and semantic segmentation accurately classifies and localizes marine litter objects in challenging underwater and surface imagery. Improved automated detection of marine debris is essential for quantifying plastic pollution distribution and informing targeted cleanup and regulatory interventions.
Towards Detecting Floating Objects on a Global Scale with Learned Spatial Features Using Sentinel 2
AI summary Read the abstract
Researchers developed a machine learning approach using Sentinel-2 satellite imagery to detect floating plastic debris and marine litter on a global scale, demonstrating that learned spatial features can improve detection of large aggregations of floating objects on water surfaces.
Advanced Classification of Marine Pollutants Using Sentinel-2 Multispectral Thermal Imaging and Vision Transformer for Enhanced Water Quality Assessment
AI summary Read the abstract
This study used satellite multispectral imaging from the Sentinel-2 platform combined with a Vision Transformer machine learning model to automatically classify different types of marine pollutants — including plastics, algae, and oil — from aerial imagery. The AI-based approach significantly outperformed traditional classification methods and could detect plastic debris patches across large ocean areas. Automated large-scale detection of marine plastic pollution from satellites could transform the way we monitor and respond to ocean plastic contamination.
Detection of Microplastics in Coastal Environments Based on Semantic Segmentation
AI summary Read the abstract
Researchers developed a deep learning semantic segmentation approach for detecting microplastics on sandy beaches at the pixel level, evaluating 12 models including U-Net variants and transformer architectures under real-world conditions.
Research digests by email
When a large batch of papers lands in the Atlas, we read through it and send a short write-up of what stood out.