南海水色遥感的主因子分析

分析了叶绿素、黄色物质和无机悬浮颗粒等要素对海水光谱反射率的贡献，并由这些要素的光学特性正演光谱反射率，与实测结果符合较好。在此基础上，利用主成分分析方法，通过对光谱反射率数据的特征向量变换和主因子回归，建立了反演南海海水叶绿素和溶解有机碳的遥感算法。与实测结果的比较表明，叶绿素和溶解有机碳的相对误差分别达１７５％和３７４％。

Principal Component Analysis for Ocean Color Remote Sensing in South China Sea

A three component ocean color model for South China Se a is presented, which takes into account the contributions of chlorophyll, yel low substance and inorganic particles. It is then used to simulate the spectral reflectance just below sea surface based on optical properties of these colored matters. The reflectance spectra are also calculated from in situ measured downw elling and upwelling irradiance. The simulated spectra are validated by the comp arison with those measured in situ. Principal component analysis is carried out on a large number of reflectance sp ectra, and it is used to develop the retrieval algorithms of chlorophyll and d issolved organic carbon. The results indicated that the retrieval accuracy of ch lorophyll and dissolved organic carbon is 17% and 37.4% respectively. It appea r s possible that this full spectral method can improve ocean color remote sensin g accuracy.