Pixel-level detection and classification of marine oil spills in aerial imagery with annotation uncertainty handling

Oil spills pose significant environmental threats, particularly in coastal and marine ecosystems, requiring fast and accurate detection methods to support mitigation efforts. In this study, we propose a deep learning-based framework for automatic oil spill detection and classification using convolut...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Marine pollution bulletin Jg. 223; S. 118975
Hauptverfasser: Tavares, Luiz G.C., Passos, Wesley L., Martins, Bruno C., Rodrigues, Mauricio de P., da Silva, Eduardo A.B., Netto, Sergio L.
Format: Journal Article
Sprache:Englisch
Veröffentlicht: England Elsevier Ltd 01.02.2026
Schlagworte:
Aerial images
Annotation uncertainty
Convolutional neural networks
Image and video processing
Image segmentation
Oil spill
Oil spill
Annotation uncertainty
Image segmentation
Aerial images
Image and video processing
Convolutional neural networks
ISSN:0025-326X, 1879-3363, 1879-3363
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Oil spills pose significant environmental threats, particularly in coastal and marine ecosystems, requiring fast and accurate detection methods to support mitigation efforts. In this study, we propose a deep learning-based framework for automatic oil spill detection and classification using convolutional neural networks. We developed two complementary RGB image datasets annotated at the pixel level: one captured from real maritime monitoring operations and another curated from online sources. These datasets are employed to train and evaluate semantic segmentation models based on the U-Net architecture. The models are capable of detecting oil presence and classifying different oil types based on appearance characteristics (e.g., silvery, rainbow, emulsified). To improve performance and generalization, a series of ablation studies have been conducted, examining the effects of network backbone and pre-training, data augmentation strategies, and the handling of visually ambiguous regions. Our best-performing model achieved a mean Intersection over Union (mIoU) of 0.738 and an accuracy of 86.9% for binary oil detection, while achieving an mIoU of 0.351 and an accuracy of 81.6% for multiclass oil type classification. These results demonstrate the feasibility of using pixel-level segmentation for oil spill monitoring in diverse marine conditions, highlighting the importance of dataset quality and model configuration in real-world deployment. •Introduced two novel datasets for pixel-level oil spill segmentation.•Proposed uncertainty-aware training using an “undefined” class.•Compared five U-Net backbones for oil detection and multiclass oil type mapping.•Achieved mIoU 0.738 and 0.351 in binary and multiclass segmentation, respectively.•Method enables robust oil spill mapping for environmental surveillance.
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
ISSN:0025-326X
1879-3363
1879-3363
DOI:10.1016/j.marpolbul.2025.118975