Analyzing The Hyperspectral Features Of Two Algae In Water Blooms: Microcystis Aeruginosan And Chlorella Vulgaris

With its advantages of rapidness, simpliness and easisness on long-termmonitoring, hyperspectral remote sensing has been widely used in waterquality monitoring work. In recent years, the situation of eutrophication in ourcountry is very severe, the lake algae blooming has happened frequently, thestudy of monitoring the eutrophication and algal blooming through remotesensing technology is also in progress.About using remote sensing to monitor water eutrophication and algalblooming, there are still many mechanism aspects not clear, such as theinternal mechanism of algal blooming, the spectral features of blooming water,the response relation between spectrum and water quality, the interrelationshipbetween optical active materials in water. In order to ultimately achieve thereal-time monitoring and forcasting the algal blooming through remote sensingtechnology, it is essential to understand the characteristics of water qualityand the spectral features and the response relation between them.This study focuses on two algal species, Microcystis aeruginosa andChlorella vulgaris. Their growing data and hyperspectral reflecting data werecollected under laboratory training. Their growing patterns, the influence fromdifferent nutrients on their growth and the hyperspectral characteristics wereexplored and analysed. The prediction models based on relations betweenalgae cell and chlorophyll-a concentrations and hyperspectral reflectrancewere established through the principal component analysis and linearregression method, which can be used to predict the concentrations of algaeand chlorophyll-a.According to the results, the spectrum curve of Microcystis aeruginosashows a pattern of three peaks and three valleys in the visible band. The threepeaks appear at bands around550nm,660nm and710nm while the threevalleys emerge at bands around440nm,620nm and675nm. The curve ofChlorella vulgaris shows a different pattern of two peaks and two valleys inthe visible band. The two peaks appear at bands around550nm and710nmwhile the two valleys emerge at bands around440nm and670nm. Microcystis aeruginosa and Chlorella vulgaris show different spectral features at the bandaround620nm. The peaks and valleys of spectral curves locate exactly in theregion of reflection and absorption of chlorophyll. These bands are alsoclearly the sensitive bands of Microcystis aeruginosa and Chlorella vulgaris.It was also found that Microcystis aeruginosa has similar spectrumcharacteristics with the one of algae blooming from400nm to710nm.The principal components analysis of sensitive bands reduced thedimensions of spectral data effectively while remained most of the spectruminformation. Based on the built relation models between the principlecomponent information and algal and chlorophyll-a concentrations, thealgae-cell-concentration model is better than the chlorophyll-a-concentrationmodel.Through the indoor research on spectrum of two algal species, it can helpus to identify the spectral features of the specific algae species and toovercome many interference factors of outdoor research. This can providefoundmental support for future research.