基于哨兵2号时序影像的水华提取  

Algae Bloom Extraction Based on Sentinel-2 Time-series Images

作　　者：敦力民[1] 尹海丹 马嵩 DUN Limin;YIN Haidan;MA Song(Shenyang Geotechnical Investigation&Surveying Research Institute Co.,Ltd.,Shenyang 110004,China;Sichuan Institute of Territorial and Spatial Planning,Chengdu 610081,China;Guangdong Shenzhen Ecological Environment Monitoring Center Station,Shenzhen 518038,China)

机构地区：[1]沈阳市勘察测绘研究院有限公司,辽宁沈阳110004 [2]四川省国土空间规划研究院,四川成都610081 [3]广东省深圳生态环境监测中心站,广东深圳518038

出　　处：《地理空间信息》2024年第1期44-48,共5页Geospatial Information

基　　金：国家自然科学基金资助项目(41971295)。

摘　　要：水华现象给人类生活和社会发展造成了负面影响,而目前设计的水华指数较难应用于哨兵2号影像。以哨兵2号遥感影像为数据源,通过分析水华、水体、水田等地物的时序光谱特征,提出从时序特征的角度区分水体与水田,再利用多元逐步线性回归的思想进行水华提取的方法;并与常用的浮游藻类指数(FAI)进行了对比实验。结果表明,该方法的提取精度优于FAI,总体精度达到90%以上,Kappa系数达到0.8以上,并能有效减少水田、水生植物、薄云等虚警,具有良好的提取精度与普适性。The phenomenon of algae bloom has caused negative impact on human life and social development,and the currently designed algae bloom indexes are difficult to be applied to Sentinel-2 images.Based on Sentinel-2 images,we analyzed the time-series spectral features of algae bloom and other features,and proposed a new algae bloom extraction method,which distinguished water bodies and water fields from the per-spective of time-series features,and then used the idea of multivariate stepwise linear regression to extract algae bloom.We conducted a compari-son experiment between this method and FAI in the study area.The results show that the extraction accuracy of method in this paper is better than FAI,the overall accuracy reaches more than 90%,and the Kappa coefficient reaches more than 0.8,which can prove the good extraction ac-curacy and universality of this method.

关 键 词：哨兵2号 水华提取 水华指数 

分 类 号：P237[天文地球—摄影测量与遥感] P237[天文地球—测绘科学与技术]