GF-1与Landsat8水体叶绿素a浓度协同反演——以太湖为例

Synergistic inversion method of chlorophyll a concentration in GF-1 and Landsat8 imagery: A case study of the Taihu Lake

不同空间分辨率、光谱分辨率和辐射分辨率传感器数据的协同反演,对于提高水体叶绿素a浓度反演精度具有重要作用。以GF-1 WFV和Landsat8 OLI数据为对象,分别以单波段替代、单波段融合和三波段融合的协同方法,分析空间分辨率和光谱分辨率在多源遥感数据协同反演过程中对于提高水体叶绿素a反演精度的主导特征;在此基础之上,进一步探索GF-1 WFV和Landsat8 OLI数据协同反演的最优组合方式,以提高叶绿素a浓度的反演精度。结果表明,在GF-1 WFV和Landsat8 OLI协同反演过程中,近红外波段光谱分辨率和辐射分辨率对精度的影响占据主导,近红外波段光谱分辨率的提高更有利于提高叶绿素a浓度的反演精度;在蓝光波段与红光波段,则是空间分辨率越高叶绿素a浓度反演精度越高;GF-1 WFV和Landsat8 OLI最优叶绿素a协同反演光谱指数组合因子为:Landsat8 OLI近红外波段、GF-1 WFV和Landsat8 OLI融合红光波段、GF-1 WFV和Landsat8 OLI融合蓝光波段。通过实测数据验证表明,协同前GF-1 WFV和Landsat8 OLI单独反演结果的平均相对误差分别为41.93%和38.37%,优化协同反演后平均相对误差降低到17.35%。

Different spatial resolutions,spectral resolutions and radiation resolutions influence the accurate estimation of remotely sensed chlorophyll a concentration of water.In this study,GF-1 WFV and Landsat8 OLI imagery was used as objects,and the cooperative methods of single-band substitution,single-band fusion and three-band fusion were respectively used to analyze dominant characteristics of spatial resolution and spectral resolution for improving the precision of chlorophyll a concentration inversion in multi-source remote sensing data.On such a basis,the optimal combination of GF-1 WFV and Landsat8 OLI data was further explored so as to improve the inversion accuracy of chlorophyll a concentration and promote the application of domestic high-resolution satellite GF-1 imagery.The results show that,in the GF-1 WFV and Landsat8 OLI cooperative inversion process,the spectral resolution and radiation resolution of near infrared band dominate the characteristics,and the influence of the near infrared band spectrum resolution enhancement is more favorable for improving the inversion accuracy of chlorophyll a concentration,whereas in the blue and red bands,the higher the spatial resolution,the higher the accuracy of chlorophyll a concentration inversion.The combination factors of GF-1 WFV and Landsat8 OLI optimal chlorophyll a concentration synergistic inversion spectral index are as follows:Landsat8 OLI near infrared band,GF-1 WFV and Landsat8 OLI fused red band,GF-1 WFV and Landsat8 OLI fused blue band.The GF-1 WFV and Landsat8 OLI separate inversion accuracy with average relative errors of 41.93%and 38.37%,respectively.After optimization,the average relative error of synergistic inversion is reduced to 17.35%.This study preliminarily explored the spectral resolution and spatial resolution of GF-1 WFV and Landsat8 OLI imagery of water chlorophyll a concentration cooperative inversion dominant characteristics and the optimal coordinated way.The authors are in the hope of providing reference for the channel design of the following domestic satellites and the cooperative inversion of multi-source satellites.