融合多源地理数据与高分辨率遥感影像的尾矿库识别与监测——以云南省个旧市为例  

Identifying and monitoring tailings ponds by integrating multi-source geographic data and high-resolution remote sensing images:A case study of Gejiu City,Yunnan Province

作　　者：刘晓亮 王志华[1,2] 邢江河 周睿 杨晓梅 刘岳明[1,2] 张俊瑶 孟丹 LIU Xiaoliang;WANG Zhihua;XING Jianghe;ZHOU Rui;YANG Xiaomei;LIU Yueming;ZHANG Junyao;MENG Dan(State Key Laboratory of Resources and Environmental Information System,Institute of Geographic Sciences and Natural Resources Research,CAS,Beijing 100101,China;University of Chinese Academy of Sciences,Beijing 100049,China;College of Geoscience and Surveying Engineering,China University of Mining and Technology(Beijing),Beijing 100083,China;Image Sky Beijing Technology Co.,Ltd.,Beijing 100086,China)

机构地区：[1]中国科学院地理科学与资源研究所资源与环境信息系统国家重点实验室,北京100101 [2]中国科学院大学,北京100049 [3]中国矿业大学(北京)地球科学与测绘工程学院,北京100083 [4]京师天启(北京)科技有限公司,北京100086

出　　处：《自然资源遥感》2024年第1期103-109,共7页Remote Sensing for Natural Resources

基　　金：国家重点研发计划项目“我国污染场地时空分布规律及其形成机制”(编号:2018YFC1800100);LREIS自主创新项目“海岸带系统精细刻画与陆海统筹过程模拟调控”(编号:KPI001)共同资助。

摘　　要：尾矿库是具有高势能的重大危险源,对尾矿库空间范围开展快速识别与监测,及时掌握尾矿库数量及分布情况,对我国尾矿库的环境监管与治理具有重要意义。现有单纯基于遥感影像识别提取的思路,因缺乏尾矿库潜在目标针对性,易将裸露地表等混淆为尾矿库,给实际的尾矿库监测应用带来较大误差。为此,提出一种融合企业名录与空间分布点位、数字高程模型(digital elevation model,DEM)、道路网等多源地理数据与高分遥感影像的尾矿库提取方法。以云南省个旧市为研究区的应用验证结果表明:融合多源地理数据可有效排除不存在尾矿库的干扰区域,尾矿库提取结果的精确率和召回率分别为83.9%和72.4%。该技术方案在全国尺度的尾矿库高频次、自动化遥感监测中具有广阔的应用前景。Tailings ponds are considerable hazard sources with high potential energy.Ascertaining the number and distribution of tailings ponds in a timely manner through rapid identification and monitoring of their spatial extents is critical for the environmental supervision and governance of tailings ponds in China.Due to the lack of pertinence for potential targets,identifying tailings ponds based on solely remote sensing images is prone to produce confusion between tailings ponds and exposed surfaces,resulting in significant errors in practical applications.This study proposed an extraction method for tailings ponds,which integrated enterprise directory,multi-source geographic data(e.g.,data from spatial distribution points,digital elevation model(DEM),and road networks),and high-resolution remote sensing images.The application of this method in Gejiu City,Yunnan Province indicates that integrating multi-source geographic data can effectively exclude the interferential areas without tailings ponds,with the precision and recall rates of the extraction results reaching 83.9%and 72.4%,respectively.The method proposed in this study boasts significant application prospects in high-frequency and automated monitoring of tailings ponds nationwide.

关 键 词：多源地理数据 遥感 面向对象分类 尾矿库 多尺度分割 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]