A Fuzzy Integral Model for Estimating Chlorophyll Concentrations in Tai Lake from Thematic Mapper Imagery

摘要

A model for Tai Lake chlorophyll a concentrations is addressed herein using the Choquet fuzzy integral that combines a set of different kind of BP neural network water quality classifiers. Firstly each neural network is employed to model the transfer function between the chlorophyll a concentrations and the radiances of Landsat Thematic Mapper data, then the degree of belief of chlorophyll a concentrations distribution and accuracy (as fuzzy density of Choquet fuzzy integral) is obtained. After that, Choquet fuzzy integral is used to fuse with outputs of each neural network, resulting in the chlorophyll a concentrations evaluation. Experiment results on Tai Lake of China show that the performance of the proposed model out-performs that of other conventional techniques.