Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Beyond Segmentation: An Oil Spill Change Detection Framework Using Synthetic SAR Imagery

Chenyang Lai, Shuaiyu Chen, Tianjin Huang, Siyang Song, Guangliang Cheng, Chunbo Luo, Zeyu Fu

TL;DR

The paper tackles reliable oil-spill detection under variable sea states by shifting from single-image segmentation to bi-temporal change detection. It introduces Oil Spill Change Detection (OSCD) and the Temporal-Aware Hybrid Inpainting (TAHI) framework to synthesize pre-spill SAR images from post-spill data, enabling supervised bi-temporal learning. Through the OSCD dataset and benchmarking of state-of-the-art change-detection models, it shows that temporally guided supervision reduces false positives from look-alike phenomena and improves detection accuracy compared with traditional segmentation. The work provides a scalable offline data-generation pipeline and a practical pathway to integrate temporally aware oil-spill monitoring into real-world surveillance, with potential applicability to other rare marine events.

Abstract

Marine oil spills are urgent environmental hazards that demand rapid and reliable detection to minimise ecological and economic damage. While Synthetic Aperture Radar (SAR) imagery has become a key tool for large-scale oil spill monitoring, most existing detection methods rely on deep learning-based segmentation applied to single SAR images. These static approaches struggle to distinguish true oil spills from visually similar oceanic features (e.g., biogenic slicks or low-wind zones), leading to high false positive rates and limited generalizability, especially under data-scarce conditions. To overcome these limitations, we introduce Oil Spill Change Detection (OSCD), a new bi-temporal task that focuses on identifying changes between pre- and post-spill SAR images. As real co-registered pre-spill imagery is not always available, we propose the Temporal-Aware Hybrid Inpainting (TAHI) framework, which generates synthetic pre-spill images from post-spill SAR data. TAHI integrates two key components: High-Fidelity Hybrid Inpainting for oil-free reconstruction, and Temporal Realism Enhancement for radiometric and sea-state consistency. Using TAHI, we construct the first OSCD dataset and benchmark several state-of-the-art change detection models. Results show that OSCD significantly reduces false positives and improves detection accuracy compared to conventional segmentation, demonstrating the value of temporally-aware methods for reliable, scalable oil spill monitoring in real-world scenarios.

Beyond Segmentation: An Oil Spill Change Detection Framework Using Synthetic SAR Imagery

TL;DR

The paper tackles reliable oil-spill detection under variable sea states by shifting from single-image segmentation to bi-temporal change detection. It introduces Oil Spill Change Detection (OSCD) and the Temporal-Aware Hybrid Inpainting (TAHI) framework to synthesize pre-spill SAR images from post-spill data, enabling supervised bi-temporal learning. Through the OSCD dataset and benchmarking of state-of-the-art change-detection models, it shows that temporally guided supervision reduces false positives from look-alike phenomena and improves detection accuracy compared with traditional segmentation. The work provides a scalable offline data-generation pipeline and a practical pathway to integrate temporally aware oil-spill monitoring into real-world surveillance, with potential applicability to other rare marine events.

Abstract

Marine oil spills are urgent environmental hazards that demand rapid and reliable detection to minimise ecological and economic damage. While Synthetic Aperture Radar (SAR) imagery has become a key tool for large-scale oil spill monitoring, most existing detection methods rely on deep learning-based segmentation applied to single SAR images. These static approaches struggle to distinguish true oil spills from visually similar oceanic features (e.g., biogenic slicks or low-wind zones), leading to high false positive rates and limited generalizability, especially under data-scarce conditions. To overcome these limitations, we introduce Oil Spill Change Detection (OSCD), a new bi-temporal task that focuses on identifying changes between pre- and post-spill SAR images. As real co-registered pre-spill imagery is not always available, we propose the Temporal-Aware Hybrid Inpainting (TAHI) framework, which generates synthetic pre-spill images from post-spill SAR data. TAHI integrates two key components: High-Fidelity Hybrid Inpainting for oil-free reconstruction, and Temporal Realism Enhancement for radiometric and sea-state consistency. Using TAHI, we construct the first OSCD dataset and benchmark several state-of-the-art change detection models. Results show that OSCD significantly reduces false positives and improves detection accuracy compared to conventional segmentation, demonstrating the value of temporally-aware methods for reliable, scalable oil spill monitoring in real-world scenarios.
Paper Structure (15 sections, 4 equations, 5 figures, 3 tables)

This paper contains 15 sections, 4 equations, 5 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. SAR-based Oil Spill Detection
  4. Change Detection in Remote Sensing
  5. Methodology
  6. Problem Formulation
  7. Temporal-Aware Hybrid Inpainting
  8. High-Fidelity Hybrid Inpainting
  9. Temporal Realism Enhancement
  10. Dataset Construction and Ablation Study
  11. OSCD Dataset
  12. Restoration Ablation Study
  13. OSCD Evaluation
  14. Results and Discussion
  15. Conclusion

Figures (5)

  • Figure 1: Comparison between our proposed approach and current segmentation-based methods. Our temporally aware change detection approach effectively isolates new oil spills, reducing false positives caused by visually similar look-alikes.
  • Figure 2: TAHI Component 1: High-Fidelity Hybrid Inpainting (HFHI) restores spatial continuity to generate a pre-event estimate.
  • Figure 3: TAHI Component 2: Temporal Realism Enhancement (TRE) refines radiometric and textural properties to produce the final reference image.
  • Figure 4: Visual comparison of ablation results. From left to right: Post-Spill (Original), +HFHI (PM only), +HFHI (Full), and +HFHI (Full) + TRE. The proposed pipeline progressively enhances restoration quality, effectively removing oil slicks while preserving sea-surface texture.
  • Figure 5: Visual comparison of CD and segmentation outputs on the OSCD dataset. CD better avoids false positives on look-alikes. GT shows oil spills (cyan) and look-alikes (red).