CAS 滑坡数据集：用于山体滑坡检测深度学习的大规模多传感器数据集

在这项工作中，我们提出了CAS滑坡数据集，这是一个用于基于深度学习的滑坡检测的大规模多传感器数据集，由中国科学院（CAS）山地灾害与环境研究所的人工智能小组开发。该数据集旨在解决滑坡识别中遇到的挑战。随着气候变化和地震导致的山体滑坡发生率增加，人们越来越需要一个精确而全面的数据集来支持快速有效的山体滑坡识别。与现有数据集的数据集大小、覆盖范围、传感器类型和分辨率限制相比，CAS 滑坡数据集包括 20,958 张图像，整合了来自 <> 个地区的卫星和无人机数据。为了确保可靠性和适用性，我们建立了一种稳健的方法来评估数据集的质量。我们建议使用CAS滑坡数据集作为构建滑坡识别模型的基准，并促进深度学习技术的发展。研究人员可以利用该数据集获得增强的预测、监测和分析能力，从而推进自动滑坡检测。如果您打算使用我们的数据集，请通过引用我们在您的项目中的工作来感谢我们的研究。

In this work, we propose the CAS Landslide Dataset, a large-scale multi-sensor dataset for deep learning-based landslide detection, developed by the AI team of the Institute of Mountain Hazards and Environment, Chinese Academy of Sciences (CAS). This dataset is designed to address the challenges encountered in landslide identification. With the increasing incidence of landslides caused by climate change and earthquakes, there is a growing demand for an accurate and comprehensive dataset to support rapid and effective landslide identification. Compared to the limitations in dataset size, coverage, sensor types and resolution of existing datasets, the CAS Landslide Dataset contains 20,958 images, integrating satellite and unmanned aerial vehicle (UAV) data from <> regions. To ensure reliability and applicability, we established a robust method to evaluate the quality of the dataset. We recommend the CAS Landslide Dataset as a benchmark for constructing landslide identification models, and to promote the development of deep learning technologies. Researchers can leverage this dataset to acquire enhanced prediction, monitoring and analysis capabilities, thereby advancing automated landslide detection. If you intend to use our dataset, please acknowledge our research by citing our work in your project.