基于高光谱遥感的上海市黑臭水体特征水质指标反演模型构建

Construction of retrieval model for characteristic water quality indicators of black and odorous water in Shanghai based on hyperspectral remote sensing

快速、准确地进行黑臭水体水质监测十分重要。化学需氧量(COD)、氨氮、总磷等是黑臭水体判别的常用化学指标,叶绿素a(Chl-a)、总悬浮物、透明度等是黑臭水体判别的常用光学指标。遥感分析技术可以进行快速大范围监测,因此探索建立基于卫星高光谱遥感数据与黑臭水体特征水质指标浓度的反演模型具有现实意义。以上海市2017年黑臭河道整治目录中的105条(段)河道为研究对象,同步开展光谱监测和水质监测,分析各指标与遥感反射率的相关性,选出适合进行反演的特征指标,并建立反演模型。结果表明:(1)光学指标中的Chl-a和化学指标中的COD可作为上海市黑臭水体反演的特征指标;(2)以三波段模型构建Chl-a反演模型,宜选用667.47、710.75、766.21 nm 3个波长的多项式回归模型,但是Chl-a质量浓度最好≥16μg/L;(3)以最佳波段或波段组合模型构建COD反演模型,宜选用637.14 nm波长的线性回归模型。

Rapid and accurate monitoring for black and odorous water quality is of great significance.Chemical oxygen demand(COD),ammonia nitrogen and total phosphorus are common chemical indicators for distinguishing black and odorous water,while chlorophyll a(Chl-a),total suspended solids and transparency are common optical indicators.Remote sensing analysis technology can monitor large scale of water areas rapidly.Therefore,it is of great practical significance to construct a retrieval model based on satellite hyperspectral remote sensing data and characteristic water quality indicators of black and odorous water.105 river samples in the list of black and odorous water bodies in 2017 in Shanghai were taken as the research object,and their spectrum and water quality indicators were monitored simultaneously.The correlation between indicators and remote sensing reflectance were analyzed and suitable indicators were selected to construct retrieval models for black and odorous water in Shanghai.Results showed that:(1)Chl-a of optical indicators and COD of chemical indicators could be selected as characteristic indicators for the retrieval of black and odorous water in Shanghai.(2)The Chl-a retrieval model was constructed by three-band model with polynomial regression,whose wavelengths were 667.47,710.75 and 766.21 nm.But the Chl-a mass concentration usually needed to be over 16μg/L.(3)The COD retrieval model was constructed by best band or bands combination model with linear regression,whose wavelength was 637.14 nm.