Estimation of Ocean Water Chlorophyll-A Concentration Using Fuzzy C-Means Clustering and Artificial Neural Networks

A system incorporating a fuzzy c-means clustering and an ensemble of artificial neural networks (ANNs) is proposed to estimate chlorophyll-a (Chl a) concentration from remotely sensed reflectance (Rrs) measurements. Fuzzy c-means is used to measure and define multiple spectral clusters from a pre-specified training set. A radial basis function (RBF) neural network is used to emulate the function of the fuzzy c-means clustering to determine the cluster and grade of membership for previously unseen spectral measurements. Next, a feed forward multi-layer perceptron (MLP) neural network is incorporated and used for Chl a estimation. The proposed method can be used to estimate Chl a concentration from Rrs measured at various global oceanic locations representing heterogeneous water types. The performance of the proposed method is presented in two experiments representing a proof of concept and a potential global Chl a prediction model. The two experiments are compared to the traditional approach, where a single ANN is used for all water types. It is shown that the cluster-based approach has the potential to build a more global Chl a prediction model