Abstract: In order to improve the evaluation accuracy of the eutrophication state of lake water quality, a new evaluation method for the eutrophication state of lake water quality was proposed, which combined hyperspectral remote sensing and firefly algorithm（FA） and improved extreme learning machine （ELM）. Because the performance of ELM model was affected by its initial input weight and hidden layer bias parameter selection, firefly algorithm was applied to ELM model parameter optimization. The results showed that compared with PSO-ELM, GA-ELM, DE-ELM and ELM, FA-ELM could effectively improve the accuracy of water quality eutrophication evaluation, providing a new method for water quality eutrophication evaluation of lakes.

Key words: water quality, hyperspectral, eutrophication, extreme learning machine, firefly algorithm（FA）