基于GEE平台的Sentinel-2内陆复杂湖泊藻华信息提取  

Information extraction of Sentinel-2 inland complex lake algal bloom based on GEE platform

作　　者：宁悦 NING Yue(The First Institute of Surveying and Mapping of Xinjiang Uygur Autonomous Region,Changji 831100,China)

机构地区：[1]新疆维吾尔自治区第一测绘院,新疆昌吉831100

出　　处：《经纬天地》2023年第2期5-9,共5页Survey World

摘　　要：本文基于Sentinel-2遥感影像协同Google Earth Engine平台,以10 m的空间分辨率提取内陆复杂湖泊乌梁素海黄苔藻华分布信息,提出一个可操作的藻华遥感监测方法。利用基于像素的随机森林模型提取乌梁素海准确的黄苔藻华信息分布图,通过定性和定量相结合的精度评估方法,验证了该方法的稳定性和适用性。结果表明:(1)利用Sentinel-2遥感影像可以获得更精细的藻华信息分布,可以实现频繁且精确的藻华制图。(2)分类的总体精度为91.82%,kappa系数为0.91。本文提出的方法有利于研究区藻华暴发预警和生态环境科学管理。Based on Sentinel-2 remote sensing images and Google Earth Engine platform,this paper extracts the distribution information of yellow moss algal blooms in Uliangsuhai,an inland complex lake,with a spatial resolution of 10m,and proposes an operable remote sensing monitoring method for algal blooms.In this paper,the pixel-based random forest model is used to extract the accurate distribution map of algal bloom information in Uliangsuhai,and the stability and applicability of the method are verified by the accuracy evaluation method combining qualitative and quantitative methods.The results show that:(1)Using Sentinel-2 remote sensing images can obtain finer distribution of algal bloom information,and can realize frequent and accurate algal bloom mapping.(2)The overall accuracy of classification is 91.82%,and the kappa coefficient is 0.91.The method proposed in this paper is beneficial to early warning of algal bloom outbreak and scientific management of ecological environment in the study area.

关 键 词：黄苔藻华 随机森林 Google Earth Engine 乌梁素海 Sentinel-2 

分 类 号：P962[天文地球—自然地理学]