基于神经网络算法的Sentinel-1和Sentinel-2遥感数据联合反演土壤湿度研究
投稿时间:2019-04-19  修订日期:2019-05-13  点此下载全文
引用本文:
摘要点击次数: 71
全文下载次数: 0
作者单位E-mail
吴善玉 南京信息工程大学大气物理学院 15261827862@163.com 
鲍艳松 南京信息工程大学大气物理学院 ysbao@nuist.edu.cn 
蔡僖 南京信息工程大学大气物理学院  
何颖 南京信息工程大学大气物理学院  
朱柳桦 南京信息工程大学大气物理学院  
陆其峰 中国气象局卫星气象中心中国遥感卫星辐射测量和定标重点开放实验室  
刘旭林 北京市气象探测中心  
李林 北京市气象探测中心  
侯岳 格尔木市气象局  
雷红玉 格尔木市气象局  
李广文 格尔木市气象局  
马军 格尔木市气象局  
唐维尧 南京信息工程大学大气物理学院  
基金项目:国家重点基础研究发展计划(973计划);国防科工局十三五预研项目;研究生创新项目
中文摘要:土壤水分是生态环境的重要参数及水分循环的重要组成部分,多源遥感数据联合反演地表土壤水分是近年来研究的热点和趋势。作为新一代Sentinel系列卫星,Sentinel-1 SAR数据与Sentinel-2光学数据联合具有广泛应用前景。本文以西班牙萨拉曼卡地区为研究区域,联合Sentinel-1后向散射系数和入射角信息、Sentinel-2光学数据提取的植被指数、地面实测数据,构建BP神经网络土壤湿度反演模型,并将该模型用于试验区土壤湿度反演。研究结果表明:(1)基于Sentinel-1卫星VV和VH极化雷达后向散射系数、雷达入射角和Sentinel-2提取的植被指数数据构建的BP神经网络土壤湿度反演模型,能够实现对该地区土壤湿度高精度反演;(2)在光学与微波数据联合反演植被覆盖区土壤湿度中,Sentinel-2的NDVI、NDWI1和NDWI2指数都可以用于削弱植被对土壤湿度反演的影响,但基于SWRI1波段的NDWI1能够获得更高精度的土壤湿度反演结果(RMSE=0.049cm3/cm3、ubRMSE=0.048cm3/cm3、Bias=0.008cm3/cm3、r=0. 681);(3)对比各种试验方案,相比于Sentinel-1 VH极化模式,Sentinel-1 VV极化模式在土壤湿度中表现出更大的优势,说明Sentinel-1 VV极化更适用于土壤湿度反演研究。
中文关键词:土壤水分  Sentinel-1  Sentinel-2  BP神经网络
 
Joint retrieval of soil moisture by Sentinel-1 and Sentinel-2 remote sensing data based on neural network algorithm
Abstract:Soil moisture is an important parameter of ecological environment and an important part of water cycle. Multi-source remote sensing data retrieval of surface soil moisture is a hotspot and trend in recent years. As a new generation of Sentinel satellites, Sentinel-1 SAR data combined with sentinel-2 optical data has broad application prospects. Based on Salamanca area of Spain, combined the Sentinel-1 backscattering coefficient and the incident angle information, the vegetation index extracted from the Sentinel-2 optical data, and the ground measured data to construct the BP neural network soil moisture retrieval model. The model was used to retrieve the soil moisture. Finally the model retrieval results were tested and evaluated. The results showed that: (1) The BP neural network soil moisture retrieval model based on Sentinel-1 satellite VV and VH polarized radar backscattering coefficient, radar incident angle and Sentinel-2 extracted vegetation index data could realize high precision retrieval of soil moisture in Salamanca area; (2) In the joint retrieval soil moisture of optical and microwave data in vegetation area, NDVI, NDWI1 and NDWI2 indexes from Sentinel-2 could be used to weaken the influence of vegetation on soil moisture retrieval, but NDWI1 based on SWRI1 band could obtain more accurate results (RMSE=0.049cm3/cm3, ubRMSE=0.048cm3/cm3, Bias=0.008cm3/cm3, r=0.681); (3) Comparing the various experimental schemes, the Sentinel-1 VV polarization mode showed a greater advantage than the Sentinel-1 VH polarization mode, indicating that Sentinel-1 VV polarization was more suitable for soil moisture retrieval studies.
keywords:soil moisture  Sentinel-1  Sentinel-2  BP neural network
  查看/发表评论  下载PDF阅读器