Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
June 14, 2026Remote SensingOpen Access

Two-Stage Oil Spill Detection in SAR Using a Domain-Adapted Segment Anything Model

View Full Paper
Ask AI
Bookmark
Share

Authors

GGGeorge GiannopoulosKingston UniversityMKMaria KremeziNational Technical University of AthensVKVassilia KarathanassiNational Technical University of Athens

Discussion

Loading...

Member takes

Implication

Randomized trial demonstrates improved oil spill mapping in SAR imagery, highlighting advanced segmentation techniques.

Key Points

  • The aim is to develop an accurate two-stage deep learning methodology for detecting and delineating oil spills in SAR imagery.
  • Developed a ConvNeXt-T classifier for initial oil slick screening in image patches.
  • Applied a domain-adapted Segment Anything Model for single-shot segmentation of spill boundaries.
  • Combined preprocessed Sentinel-1 VV backscatter with GLCM texture measures for enhanced input representation.
  • Achieved an overall accuracy with an F1-score of 0.86, outperforming UNet and CBDNet which scored 0.83.
  • Traditional models like DeepLabV3, SegNeXt, and OFCNet scored 0.82, indicating superior performance of the proposed method.
  • Wind speed analysis revealed it affects detectability but does not independently dictate segmentation quality.

Cite This Study

Giannopoulos et al. (2026) studied this question.

synapsesocial.com/papers/6a2e4608b1cc60ccdea8ae30https://doi.org/10.3390/rs18121948
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 3 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Textural Features for Image Classification1973 · 22,933 citations
  2. 2Robust Tests for the Equality of Variances1974 · 2,949 citations
  3. 3Marine oil spill detection using improved polarimetric feature based on polarization SAR image2024 · 18 citations