MambaMPD: A Mamba-Driven Segmentation Framework for Marine Pollution Detection from Remote Sensing Imagery
作者: Shuaiyu Chen, Wei Han, Peng Ren, Chunbo Luo, Zeyu Fu
分类: cs.CV
发布日期: 2026-09-14
💡 一句话要点
提出MambaMPD框架以解决海洋污染检测问题
🎯 匹配领域: 支柱二:RL算法与架构 (RL & Architecture) 支柱九:具身大模型 (Embodied Foundation Models)
关键词: 海洋污染检测 遥感图像 语义分割 深度学习 频率感知增强 边缘引导注意力 U-Net 计算机视觉
📋 核心要点
- 现有海洋污染检测方法在低信噪比和视觉相似性导致的边界模糊等方面存在显著不足,影响检测精度。
- MambaMPD框架通过引入频率感知增强和多尺度边缘引导注意力,增强了模型对微弱污染信号的捕捉能力。
- 在MADOS和M4D数据集上,MambaMPD分别提高了F1和Oil Spill IoU,且计算效率显著高于基础模型方法。
📝 摘要(中文)
准确的海洋污染检测对于保护沿海生态系统和海洋生物多样性至关重要。尽管Vision Mamba模型在遥感语义分割中表现出色,但其在海洋污染检测中的潜力尚未得到充分探索。海洋污染检测面临低信噪比、污染模式碎片化及污染物与海水之间视觉相似性导致的边界不清晰等挑战。为此,本文提出了MambaMPD框架,结合了频率感知增强(FAA)和多尺度边缘引导注意力(EGA)两个互补的结构先验,显著提高了对微弱污染信号的敏感性,同时保留了细致的边界结构。实验结果表明,MambaMPD在多个基准数据集上表现优异,计算效率高于现有基础模型方法。
🔬 方法详解
问题定义:本文旨在解决海洋污染检测中的低信噪比、污染模式碎片化及视觉相似性导致的边界不清晰等具体问题。现有方法在这些方面表现不佳,影响了检测的准确性和效率。
核心思路:MambaMPD框架的核心思路是结合频率感知增强(FAA)和多尺度边缘引导注意力(EGA),以更好地捕捉低对比度和不规则的污染模式,同时提高边界的清晰度。这样的设计使得模型能够有效地处理复杂的海洋污染场景。
技术框架:MambaMPD的整体架构包括一个编码器和一个U-Net风格的解码器。编码器中集成了频率感知增强模块,通过小波变换将特征分解为多尺度频率子带;解码器则通过边缘引导注意力模块和深度监督逐步恢复和精细化语义和空间信息。
关键创新:本文的关键创新在于引入频率感知增强和多尺度边缘引导注意力两个模块,这与现有方法的主要区别在于能够更好地处理低对比度和复杂边界的污染检测任务。
关键设计:在参数设置上,模型采用了小波变换进行特征分解,损失函数设计上注重边界清晰度,网络结构上则结合了注意力机制和深度监督,以提升模型的整体性能。
🖼️ 关键图片
📊 实验亮点
MambaMPD在MADOS数据集上相比OSDMamba提高了3.6%的F1分数,在M4D数据集上提高了6.82%的Oil Spill IoU,且在计算效率上显著优于基础模型方法,展示了其在海洋污染检测中的优越性能。
🎯 应用场景
该研究的潜在应用领域包括海洋环境监测、生态保护及海洋资源管理等。通过提高海洋污染检测的准确性,MambaMPD框架能够为政策制定者和环境保护组织提供更可靠的数据支持,进而促进可持续发展和生态保护。未来,该技术有望扩展到其他领域,如城市环境监测和水质评估等。
📄 摘要(原文)
Accurate marine pollution detection (MPD) is essential for protecting coastal ecosystems and marine biodiversity. Vision Mamba models have shown promise in remote-sensing semantic segmentation by efficiently capturing long-range dependencies and global context, yet their potential for MPD remains underexplored. MPD is particularly challenging because of low signal-to-noise ratios, fragmented pollution patterns, and indistinct boundaries caused by the visual similarity between pollutants and the surrounding sea. To address these issues, we propose MambaMPD, an enhanced Mamba-based framework incorporating two complementary structural priors: Frequency-Aware Augmentation (FAA) and multi-scale Edge-Guided Attention (EGA). FAA integrates wavelet transforms into the encoder to decompose features into multi-scale frequency subbands, enabling the model to capture low-frequency contextual semantics and high-frequency structural details needed to identify small, low-contrast, and irregular pollution patterns. EGA adaptively fuses hierarchical, Laplacian-derived boundary cues with deep semantic representations, refining encoder features before decoding to sharpen boundaries and reduce ambiguity in visually confusing, spatially fragmented scenes. Together, these modules improve sensitivity to subtle pollution signals while preserving fine boundary structures. A U-Net-style decoder with squeeze-and-excitation attention and deep supervision progressively restores and refines semantic and spatial information across scales. Extensive experiments on two benchmark MPD datasets show that MambaMPD achieves higher mIoU than competing methods while requiring substantially less computation than foundation-model-based approaches. On MADOS, it improves F1 by 3.6% over OSDMamba; on M4D, it raises Oil Spill IoU by 6.82% over TransOilSeg.