资源一号02D卫星高光谱数据水体透明度反演研究

Water Transparency Retrieval Based on Hyperspectral Data of ZY-1-02D Satellite

水体透明度能直观反映水体清澈和混浊程度,是评价湖泊富营养化,衡量水质优劣的重要指标之一。资源一号02D卫星搭载的高光谱相机(AHSI),具有与水质参数相关的精细光谱特征,为水体透明度反演提供了丰富波段选择。文章以我国官厅水库为研究区,采用资源一号02D卫星高光谱影像,开展水体透明度反演研究。通过构建半经验模型,并基于半分析算法(第五版)(QAAv5)和半分析算法(第六版)(QAAv6)的半分析模型,对官厅水库的透明度进行估算,结果均方根误差分别为0.07m、0.55m、0.19m,平均相对误差分别为9.5%、94.7%和31.4%。研究结果表明:资源一号02D卫星高光谱数据能够实现较高精度的水体透明度反演,相比于多光谱卫星数据具有更高的反演精度,可用于内陆湖库的水质遥感监测工作。

Water transparency can reflect the clarity and turbidity of the water body and is one of the important indicators for evaluating lake eutrophication and measuring water quality.The hyperspectral imager(AHSI)onboard ZY-1-02D satellite has specific spectral characteristics related to water quality parameters and can provide a wealth of band options for water transparency retrieval.In this paper,the Guanting Reservoir is chosen as the study area and ZY-1-02D satellite hyperspectral imagery data is used to estimate its water transparency.A semi-empirical model is established,as well as the QAAv5 and QAAv6 semi-analytical models are used,to estimate the transparency of Guanting Reservoir.The RMSEs of the results are 0.07m,0.55m,and 0.19m,respectively,whereas the corresponding mean relative errors are 9.5%,94.7%,and 31.4%.The results show that the ZY-1-02D satellite hyperspectral data can achieve high-precision for water transparency estimations.Compared with multispectral satellite data,it can yield higher accuracy for water transparency retrieval of inland lakes and reservoirs.