【摘要】 利用2006年10月14~18日在甘肃省庆阳市境内环江、柔远河和马莲河实测的水体波谱数据,模拟水色卫星传感器波段,根据目前广泛应用于各种水色卫星传感器的五种叶绿素浓度估算模式计算叶绿素含量;利用平均绝对误差(MAE)、均方差(RMSE)、一致性指数(IA)和线性相关系数(LCC)等模式性能检验指标,结合与光谱数据测定同步获取的叶绿素含量分析数据,对模式性能进行评价,试图从地面遥感的角度来深入分析这些模式的应用精度;模式的性能分析结果表明:(1)Clark(1997)具有最小MAE和RMSE,最大的IA和LCC,因而是比较适用研究区域叶绿素浓度的估算模式;(2)虽然Clark(1997)性能相对来说比较高,但是估算误差仍在37%以上。进一步利用本次实验中获取的26个样本数据,对Clark(1997)模式参数进行优化和改进,得到了修正后的模式,并利用预留的8个叶绿素实测样本进行验证,对优化后的模式进行的对比分析,结果表明精度提高了16%。更多还原

【Abstract】 This paper evaluates modes for the estimation of Chlorophyll-a concentration in water based on remote sensing technique including five modes: O’Reilly(1998),Maritorena(2000),O’Reilly(2000),Clark(1997) and Mitchell(2001).The data used in this study were measured at Huanjiang river,Rouyuan river and Malian river in Qingyang city,Gansu Province on October 14th-18th,2006.Reflectance spectra is measured with a portable FieldSpc FR spectoradiometers(ASD Inc.)(350-1 050 nm).Currently water samples for concentration measurement associated with chlorophyll and petroleum contamination and suspended material are processed with the related standard of China.Performance indexes including Mean Absolute Error(MAE),Root Mean Square Error(RMSE),Index of Agreement(IA) and Linear Correlation Coefficient(LCC)are employed.The results show that(1) Clark’s mode(1997) has a good performance;(2) although with more high performance,the estimating error of Clark’s mode(1997) reaches to over 37.Ulteriorly twenty-six samples are used to optimize parameters of Clark’s mode(1997) and the modified mode is obtained.With eight reserved samples the modified mode is validated.The result shows that the accuracy is 25% and it increases 11%.更多还原