| 安天浩,王幻,张仁杰,翟司浔,李伟滔.基于深度学习的滑坡遥感制图研究进展[J].矿产勘查,2025,16(S2):134-144 |
| 基于深度学习的滑坡遥感制图研究进展 |
| Deep learning-based landslide remote sensing mapping:A review |
| 投稿时间：2025-05-20 |
| DOI：10.20008/j.kckc.2025s2020 |
| 中文关键词: 滑坡遥感制图 深度学习 应用场景 |
| 英文关键词: landslide remote sensing mapping deep learning application scenarios |
| 基金项目: |
|
| 摘要点击次数: 37 |
| 全文下载次数: 28 |
| 中文摘要: |
| 滑坡是破坏性极强的地质灾害，遥感技术的发展为其高效、精准制图提供了重要支持。近年来，深度学习以其卓越的特征提取能力与端到端学习机制，正逐步取代传统人工解译与浅层机器学习方法，成为滑坡遥感制图的核心手段。本文系统回顾了深度学习在滑坡遥感制图中的研究进展，重点围绕遥感数据源、特征体系构建与多源融合策略，阐述了主流模型在滑坡检测、易发性评价与位移预测研究主题中的关键技术与优化路径。进一步探讨了滑坡灾害链分析、灾后评估和智能预警等延伸应用场景，并总结当前面临的主要挑战。最后，展望了未来深度学习模型发展方向，为构建智能化、实时化的滑坡监测与预警体系提供理论支持与实践路径。 |
| 英文摘要: |
| Landslide is a highly destructive geological disaster，and the development of remote sensingtechnology provides important support for its efficient and accurate mapping. In recent years，deep learning，with its excellent feature extraction capability and end-to-end learning mechanism，is gradually replacing the traditionalmanual interpretation and shallow machine learning methods as the core means of landslide remote sensing map.ping. This paper systematically reviews the research progress of deep learning in landslide remote sensing map.ping，focusing on remote sensing data sources，feature system construction and multi-source fusion strategy，and describes the key technologies and optimisation paths of mainstream models in landslide detection，susceptibilityassessment and displacement prediction research topics. Extended application scenarios such as landslide hazardchain analysis，post-disaster assessment and intelligent warning are further explored，and the main challenges cur. rently faced are summarised. Finally，the future development direction of deep learning models is envisioned to provide theoretical support and practical paths for the construction of intelligent and real-time landslide monitoring and early warning system. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |