基于Landsat-9影像的云龙湖水深反演研究  

Inversion study of Yunlong Lake water depth based on Landsat-9 images

作　　者：任中杰 万永智 王伟 王继龙 宋银燕[1] 张方方 REN Zhongjie;WAN Yongzhi;WANG Wei;WANG Jilong;SONG Yinyan;ZHANG Fangfang(Xuzhou Branche of Jiangsu Province Hydrology and Water Resources Investigation Bureau,Xuzhou 221000,China)

机构地区：[1]江苏省水文水资源勘测局徐州分局,江苏徐州221000

出　　处：《江苏水利》2023年第10期51-55,共5页Jiangsu Water Resources

摘　　要：基于Landsat-9多光谱及全色影像,以云龙湖东湖为研究区域进行水深反演研究。为选取能够有效提高影像空间分辨率且能保持水深波段相关性的融合方法,对多种影像融合方法进行了对比分析;对最佳水深反演波段及反演因子进行了探索,进而构建了多种反演模型,并利用实测水深数据进行了精度评价。结果表明,SFIM融合算法对水深相关性保持性最好,最佳水深反演波段为B5波段,最佳反演因子为B5+B6,相关系数均较高,指数、傅里叶、高斯、多项式等多种水深反演模型中最优反演模型为基于最小绝对残差拟合的二次多项式模型,精度评价决定系数为0.7149,均方根误差为0.094 m。研究结果旨在为云龙湖水库水量调度及生态水位保障提供重要参考。Based on Landsat-9 multispectral and panchromatic imagery,a study on water depth inversion was conducted with the east lake of Yunlong Lake as the study area.To select the fusion method that can effectively improve the spatial resolution while maintain water depth-wave band correlation,a variety of image fusion methods were analyzed and compared.The optimum water depth inversion band and inversion factors are explored,and a variety of inversion models are constructed,and the accuracy is evaluated by using the measured water depth data.The results show that the SFIM fusion algorithm has the best ability to maintain the correlation of water depth.The best water depth inversion band is B5 band,the best inversion factor is B5+B6,and the correlation coefficients are all high.Among the water depth inversion models,such as index,Fourier,Gaussian and polynomial,the optimal inversion model is the quadratic polynomial model based on the least absolute residual(LAR)fit.The determination coefficient of precision evaluation is 0.7149,and the root-mean-square error(RMSE)is 0.094 m.The research results provide an important reference for the water dispatch and safeguards for the ecological water level at Yunlong Lake Reservoir.

关 键 词：Landsat-9 水深遥感 水深反演 SFIM LAR 云龙湖 

分 类 号：TV211.2[水利工程—水文学及水资源]