基于支持向量机的一类水域叶绿素a浓度反演研究

Retrieve the oceanic chlorophyll-a concentration in case I water by support vector machine

指出了用叶绿素a的浓度估计海洋初级生产力的重要作用;分析了目前采用的浓度反演方法的不足;尝试将基于统计学习理论的最小二乘支持向量机用于浓度反演,SeaBAM的数据实验结果表明该方法可以获得更高的反演精度;可以有效避免过学习的情况出现;不像神经网络那样需要确定网络结构。

That chlorophyll-a is an important factor when estimating ocean＇s primitive productivity. The shortcomings of current methods used to retrieve chlorophyll-a concentration were analyzed. For the first time, least squares support vector machine based on statistic learning theory was used to do this work Experiments based on this method. SeaBAM data shows that using SVM method can receive the lowest error, avoid over fitting, and not need to design net-structure.