Extracting Chlorophyll Concentration of water body based on spectral unmixing model

摘要

Remote sensing (RS) data can be used to estimate the chlorophyll-a concentration of water body, which has become a key issue of water quality monitoring. In this paper the field water reflectance spectra have been measured and used for monitoring the water quality of Taihu Lake according to reflectance spectra of water varying with concentrations. After analyzing the spectra characters, the optimal band range (670-710nm) were selected and the chlorophyll-a concentration by linear spectral unmixing model were retrieved using known water and chlorophyll-a endmember spectra of in situ samples. TM and MODIS image was processed and the distribution of chlorophyll-a concentration was retrieved by linear spectral unmixing method. The result showed that a fine linear correlation existed between the abundance of chlorophyll-a from the spectral unmixing model and the concentration of chlorophyll-a from the analytical result in the laboratory, and also shows that spectral unmxing model is a feasible model in the practical application for water quality monitoring.