Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Compositional Oil Spill Detection Based on Object Detector and Adapted Segment Anything Model from SAR Images

Wenhui Wu, Man Sing Wong, Xinyu Yu, Guoqiang Shi, Coco Yin Tung Kwok, Kang Zou

TL;DR

This work tackles oil spill detection in SAR imagery, where distinguishing spills from look-alikes is challenging and pixel-level annotations are costly. It introduces SAM-OIL, a compositional framework that blends a detector (YOLOv8), an Adapted Segment Anything Model (HQ-SAM adapter), and a parameter-free Ordered Mask Fusion module to generate category-aware oil spill masks from bounding boxes. The authors show that this first application of SAM to oil spill detection, along with the Adapter and OMF components, yields competitive performance on the M4D dataset and reduces annotation burdens compared to fully supervised segmentation. The approach highlights the potential of integrating SAM-based segmentation with SAR-specific adaptations for robust, annotation-efficient oil spill detection in remote sensing applications.

Abstract

Semantic segmentation-based methods have attracted extensive attention in oil spill detection from SAR images. However, the existing approaches require a large number of finely annotated segmentation samples in the training stage. To alleviate this issue, we propose a composite oil spill detection framework, SAM-OIL, comprising an object detector (e.g., YOLOv8), an Adapted Segment Anything Model (SAM), and an Ordered Mask Fusion (OMF) module. SAM-OIL is the first application of the powerful SAM in oil spill detection. Specifically, the SAM-OIL strategy uses YOLOv8 to obtain the categories and bounding boxes of oil spill-related objects, then inputs bounding boxes into the Adapted SAM to retrieve category-agnostic masks, and finally adopts the OMF module to fuse the masks and categories. The Adapted SAM, combining a frozen SAM with a learnable Adapter module, can enhance SAM's ability to segment ambiguous objects. The OMF module, a parameter-free method, can effectively resolve pixel category conflicts within SAM. Experimental results demonstrate that SAM-OIL surpasses existing semantic segmentation-based oil spill detection methods, achieving mIoU of 69.52\%. The results also indicated that both OMF and Adapter modules can effectively improve the accuracy in SAM-OIL.

Compositional Oil Spill Detection Based on Object Detector and Adapted Segment Anything Model from SAR Images

TL;DR

This work tackles oil spill detection in SAR imagery, where distinguishing spills from look-alikes is challenging and pixel-level annotations are costly. It introduces SAM-OIL, a compositional framework that blends a detector (YOLOv8), an Adapted Segment Anything Model (HQ-SAM adapter), and a parameter-free Ordered Mask Fusion module to generate category-aware oil spill masks from bounding boxes. The authors show that this first application of SAM to oil spill detection, along with the Adapter and OMF components, yields competitive performance on the M4D dataset and reduces annotation burdens compared to fully supervised segmentation. The approach highlights the potential of integrating SAM-based segmentation with SAR-specific adaptations for robust, annotation-efficient oil spill detection in remote sensing applications.

Abstract

Semantic segmentation-based methods have attracted extensive attention in oil spill detection from SAR images. However, the existing approaches require a large number of finely annotated segmentation samples in the training stage. To alleviate this issue, we propose a composite oil spill detection framework, SAM-OIL, comprising an object detector (e.g., YOLOv8), an Adapted Segment Anything Model (SAM), and an Ordered Mask Fusion (OMF) module. SAM-OIL is the first application of the powerful SAM in oil spill detection. Specifically, the SAM-OIL strategy uses YOLOv8 to obtain the categories and bounding boxes of oil spill-related objects, then inputs bounding boxes into the Adapted SAM to retrieve category-agnostic masks, and finally adopts the OMF module to fuse the masks and categories. The Adapted SAM, combining a frozen SAM with a learnable Adapter module, can enhance SAM's ability to segment ambiguous objects. The OMF module, a parameter-free method, can effectively resolve pixel category conflicts within SAM. Experimental results demonstrate that SAM-OIL surpasses existing semantic segmentation-based oil spill detection methods, achieving mIoU of 69.52\%. The results also indicated that both OMF and Adapter modules can effectively improve the accuracy in SAM-OIL.
Paper Structure (11 sections, 3 equations, 4 figures, 4 tables, 1 algorithm)

This paper contains 11 sections, 3 equations, 4 figures, 4 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Methodology
  3. YOLOv8-SAM
  4. Ordered Mask Fusion module
  5. Adapted SAM
  6. Experiments
  7. Experimental Setups
  8. Comparative Experiments
  9. Ablation Study
  10. Qualitative Analysis
  11. Conclusion

Figures (4)

  • Figure 1: Architecture of the proposed method. The Adapted SAM is composed of a frozen SAM and a learnable Adapter module. SAM-OIL consists of an object detector, an Adapted SAM, and an OMF module.
  • Figure 2: The epoch time for each model.
  • Figure 3: The influence of object detector classification scores on the accuracy.
  • Figure 4: Qualitative examples of YOLOv8-SAM and SAM-OIL. (a) SAR Images, (b) Ground Truth, (c) YOLOv8-SAM, (d) SAM-OIL.