Landslide4Sense
收藏资源简介:
Landslide4Sense数据集是由Institute of Advanced Research in Artificial Intelligence (IARAI)创建的滑坡检测基准数据集,包含3799个图像补丁。该数据集通过融合Sentinel-2传感器的光学层和ALOS PALSAR的数字高程模型及坡度层,提供了丰富的地理和地形信息。数据集涵盖了四个不同时间和地理位置的滑坡事件,旨在支持深度学习在滑坡检测领域的研究和方法验证。每个图像像素都被标记为是否属于滑坡区域,通过多种数据源和细致的手动标注确保准确性。该数据集的应用领域包括灾害评估、预警系统支持和滑坡现象的深入理解,为远程传感、计算机视觉和机器学习社区提供了重要的资源。
The Landslide4Sense dataset is a landslide detection benchmark dataset created by the Institute of Advanced Research in Artificial Intelligence (IARAI), consisting of 3799 image patches. It fuses optical layers from the Sentinel-2 sensor, as well as digital elevation model and slope layers from ALOS PALSAR, to provide rich geographic and topographic information. The dataset covers four landslide events with different times and geographic locations, aiming to support research and method validation for deep learning in the field of landslide detection. Each image pixel is labeled as either belonging to a landslide area or not, with its accuracy ensured by multiple data sources and meticulous manual annotation. Its application areas include disaster assessment, early warning system support and in-depth understanding of landslide phenomena, providing an important resource for the remote sensing, computer vision and machine learning communities.
- 1Landslide4Sense: Reference Benchmark Data and Deep Learning Models for Landslide DetectionInstitute of Advanced Research in Artificial Intelligence (IARAI) · 2022年



