本数据集为“资兴降雨诱发型滑坡数据集”，包含一个滑坡编录图（RLZX-LIM）与一个滑坡检测数据集（RLZX-LDD）。RLZX-LIM记录了2024年7月26日至28日中国湖南省资兴市由台风“格美”引发极端降雨所触发的19,403个浅层滑坡的空间位置与形态（多边形）。RLZX-LDD则基于卫星与无人机影像构建，为滑坡智能检测提供高质量的训练样本。该数据集旨在为数据驱动的区域滑坡研究及智能滑坡检测方法的发展提供基础数据支持。", "ds_source": "
数据来源于https://doi.org/10.6084/m9.figshare.27960762 。", "ds_process_way": "
先收集并处理事件前后的多时相遥感影像（卫星与无人机数据）以划定研究区。其次，专家通过三维时空场景的目视解译，初步绘制滑坡边界多边形，形成RLZX-LIM初稿。然后，通过野外实地调查（包括道路沿线核查与无人机精细测绘）获取验证数据，对初编成果进行定量质量评估与修正，形成最终版RLZX-LIM。最后，基于高精度的RLZX-LIM及配准的影像与辅助图层（如DEM），通过切片生成用于模型训练的特征-标签对，构建出RLZX-LDD。", "ds_quality": "
通过多源数据与实地验证予以控制。RLZX-LIM的质量以实地调查和无人机测绘获取的参考数据为基准进行了定量评估，确保了滑坡边界与属性的准确性。RLZX-LDD源于高质量且经过验证的RLZX-LIM，并与影像精确配准，从而保证了其作为检测训练样本的可靠性与鲁棒性。该数据集有效填补了降雨诱发型滑坡检测数据集的空白，其局限性在于主要依赖特定单次极端降雨事件，其时空普适性有待更多案例验证，且自动化生成方法尚未成熟，目前主要依靠专家人工解译。", "ds_acq_start_time": "2024-07-01 00:00:00", "ds_acq_end_time": "2024-07-31 00:00:00", "ds_acq_place": "中国资兴市", "ds_acq_lon_east": null, "ds_acq_lat_south": null, "ds_acq_lon_west": null, "ds_acq_lat_north": null, "ds_acq_alt_low": null, "ds_acq_alt_high": null, "ds_share_type": "login-access", "ds_total_size": 18899318314, "ds_files_count": 10552, "ds_format": "*.tif,*.png", "ds_space_res": "", "ds_time_res": "日", "ds_coordinate": "无", "ds_projection": "", "ds_thumbnail": "33fbee67-a6a9-4bba-83ef-503da28af436.png", "ds_thumb_from": 0, "ds_ref_way": "", "paper_ref_way": "", "ds_ref_instruction": "", "ds_from_station": null, "organization_id": "52b7b79b-860c-49a5-9083-9a70cf8bed5a", "ds_serv_man": "敏玉芳", "ds_serv_phone": "0931-4967596", "ds_serv_mail": "ncdc@lzb.ac.cn", "doi_value": "", "subject_codes": [ "170.4510" ], "quality_level": 3, "publish_time": "2025-12-29 10:19:01", "last_updated": "2025-12-29 10:19:01", "protected": false, "protected_to": null, "lang": "zh", "cstr": null, "i18n": { "en": { "title": "Records of shallow landslides triggered by extreme rainfall in July 2024 in Zixing, China", "ds_format": "*.tif,*.png", "ds_source": "
The data is sourced from https://doi.org/10.6084/m9.figshare.27960762 .", "ds_quality": "
Control through multi-source data and field validation. The quality of RLZX-LIM has been quantitatively evaluated based on reference data obtained from field investigations and drone mapping, ensuring the accuracy of landslide boundaries and attributes. RLZX-LDD is derived from high-quality and validated RLZX-LIM, and is accurately registered with images to ensure its reliability and robustness as a training sample for detection. This dataset effectively fills the gap in rainfall induced landslide detection dataset, but its limitation lies in mainly relying on specific single extreme rainfall events. Its spatiotemporal universality needs to be verified by more cases, and the automated generation method is not yet mature. Currently, it mainly relies on expert manual interpretation.", "ds_ref_way": "", "ds_abstract": "
This dataset is the \"Zixing Rainfall Induced Landslide Dataset\", which includes a landslide catalog map (RLZX-LIM) and a landslide detection dataset (RLZX-LDD). RLZX-LIM recorded the spatial location and morphology (polygons) of 19403 shallow landslides triggered by extreme rainfall caused by Typhoon \"Gemei\" in Zixing City, Hunan Province, China from July 26 to 28, 2024. RLZX-LDD is based on satellite and drone imagery, providing high-quality training samples for intelligent landslide detection. This dataset aims to provide fundamental data support for data-driven regional landslide research and the development of intelligent landslide detection methods.
", "ds_time_res": "日", "ds_acq_place": "Zixing City, China", "ds_space_res": "", "ds_projection": "", "ds_process_way": "Collect and process multi temporal remote sensing images (satellite and drone data) before and after the event to delineate the study area. Secondly, experts used visual interpretation of three-dimensional spatiotemporal scenes to preliminarily draw landslide boundary polygons, forming the initial draft of RLZX-LIM. Then, verification data was obtained through field investigations (including road inspections and precise drone mapping), and the preliminary results were quantitatively evaluated and revised to form the final version of RLZX-LIM. Finally, based on high-precision RLZX-LIM and registered images and auxiliary layers (such as DEM), feature label pairs for model training are generated through slicing to construct RLZX-LDD.", "ds_ref_instruction": "" } }, "submit_center_id": "ncdc", "data_level": 0, "recommendation_value": 0, "license_type": "https://creativecommons.org/licenses/by/4.0/", "doi_reg_from": "reg_outside", "cstr_reg_from": "reg_outside", "doi_not_reg_reason": null, "cstr_not_reg_reason": null, "is_paper_in_submitting": false, "ds_topic_tags": [ "极端降雨", "滑坡", "数据驱动" ], "ds_subject_tags": [ "自然地理学" ], "ds_class_tags": [], "ds_locus_tags": [ "资兴市" ], "ds_time_tags": [ 2024 ], "ds_contributors": [ { "true_name": "汪发武", "email": "wangfw@tongji.edu.cn", "work_for": "同济大学", "country": "中国" } ], "ds_meta_authors": [ { "true_name": "汪发武", "email": "wangfw@tongji.edu.cn", "work_for": "同济大学", "country": "中国" } ], "ds_managers": [ { "true_name": "汪发武", "email": "wangfw@tongji.edu.cn", "work_for": "同济大学", "country": "中国" } ], "category": "灾害" }