摘要
土壤水分是影响作物生长的重要因素,也是监测旱情、估算作物产量的重要参量。为及时、准确地掌握土壤水分,在利用水云模型(Water Cloud Model, WCM)对Sentinel-1 A的后向散射系数校正的基础上,联合地面土壤水分数据,采用线性回归、BP神经网络和支持向量回归三类模型进行了地表土壤水分反演实验研究。实验结果表明:线性回归、BP神经网络和支持向量机回归模型均有较好的反演效果,其决定系数(R;)分别为0.87、0.83和0.8,均方根误差(R;)分别为1.1%、1.9%和2.1%。3种模型基本能够满足Sentinel-1A土壤水分的反演要求,对比来看,BP神经网络的反演效果最好。
Soil moisture is an important factor affecting crop growth, as well as an important parameter for monitoring drought conditions and estimating crop yields. In order to grasp the soil moisture timely and accurately, based on the correction of the Sentinel-1 A backscattering coefficient using the Water Cloud Model(WCM), combined with the ground soil moisture data, linear regression, BP neural network and three types of support vector regression models have carried out experimental research on the inversion of surface soil moisture. The experimental results show that linear regression, BP neural network and support vector machine regression models all have good inversion results. The coefficient of determination(R;) is 0.87, 0.83 and 0.8, and the root mean square error(R;) is 1.1%, 1.9% and 2.1%. The three models can basically meet the inversion requirements of Sentinel-1 A soil moisture. In comparison, the BP neural network has the best inversion effect.
作者
焦月正
张安兵
王贺封
于坤
李武乾
JIAO Yuezheng;ZHANG Anbing;WANG Hefeng;YU Kun;LI Wuqian(School of Mining and Geomatics Engineering,Hebei University of Engineering,Handan 056038,China;Handan Polytechnic College,Handan 056001,China;Handan Key Laboratory of Natural Resources Spatial Information,Handan 056038,China)
出处
《测绘与空间地理信息》
2022年第2期40-44,共5页
Geomatics & Spatial Information Technology
基金
国家自然科学基金面上项目(42071246)
河北省自然科学基金面上项目(E2020402006)
河北省高等学校科学技术研究项目重点项目(ZD2018230)资助。
关键词
水云模型
土壤水分
后向散射系数
回归模型
water-cloud model
soil moisture
backscattering coefficient
regression model
