Abstract: Accurate remote sensing monitoring of cyanobacteria blooms in lakes can provide scientific basis for prevention and control of lake pollution. Based on Landsat8 multispectral data, Chaohu Lake, China, was taken as research objects, and normalized difference vegetation index (NDVI) model and floating algae index (FAI) model were used to extract the cyanobacteria blooms, then the intensity of cyanobacteria blooms extracted was divided into three levels. Furthermore, the two extraction methods were compared, and the distribution of cyanobacteria blooms in Chaohu Lake were analyzed. The research results show that: (1) the thresholds and grading thresholds determined for NDVI method and FAI method in this workcan effectively extract the cyanobacteria blooms in Chaohu Lake, (2) compared with NDVI method, FAI method can reduce the influence of clouds on the extraction accuracy, (3) the eutrophication in the western half of Chaohu Lake and its coastal waters has been serious, and the eutrophication in the eastern half of the Chaohu Lake is becoming serious. The extraction thresholds and classification thresholds of cyanobacteria blooms determined in this paper not only provide method support for the treatment and early warning of cyanobacteria blooms in Chaohu Lake, but also provide reference for remote sensing monitoring of cyanobacteria blooms in other lakes.

Key words: cyanobacteria bloom, floating algae index, normalized difference vegetation index, Chaohu Lake; threshold