Research On Automatic Extracting Method Of Algal Bloom Information In Typical Lakes Using ETM+ Imagery

Algal bloom in eutrophication lakes may lead to the deterioration of water quality, thus threatening the safety of drinking water. Remote sensing monitoring of algal bloom in the lake can effectively map the information about the bloom occurrence and change. However, the existing algal bloom extraction from remote sensing image is of low automation and the extraction result was mixed with the aquatic plants, therefore it is necessary to further optimize the algorithm to improve the accuracy and efficiency of the algal bloom extraction.Based on the Landsat ETM+ imagery covering Taihu lake, Chaohu lake and Dianchi lake, the paper first analyzed the spectral characteristics of algal bloom and aquatic plants as well as their difference and seasonal variation, then determined the extraction index of algal bloom information and its threshold, distinguished the algal bloom and aquatic plant, and finally built the information extraction system about lake algal bloom.35 ETM+ images were used in this study, including 20 images with heavy algal bloom distribution, with image date from May to December. Five images in the summer and winter of Taihu lake were used for algorithm development and others were used to validate. Conclusions are summarized as follows:(1) Distinction between algal bloom and water bodyBuilt a new algal bloom extraction index NFAI, based on the baseline of the reflectance between red and SWIR band and the max FAI pixel in individual image. The NFAI>0 is defined as algal bloom area. The results show that the water and bloom can be distinguished and the bloom area can be determined.(2) Distinction between algal bloom and aquatic plantBuilt the baseline height index BL123 and BL457. The algal bloom and aquatic plant can be detected and distinguished by the index’s quantile and spatial information.(3) Automatic extraction of lake algal bloomCombined NFAI with BL123 and BL457, an automatic remote sensing extraction flow and extraction system of algal bloom was built and validated by the 15 images in Taihu lake, Chaohu lake and Dianchi lake in the year from 2000 to 2012, results show that the extraction result by the proposed method in this study kept great consistency with visual interpretation, indicating this new method and extraction information system provided the more automatic and better extraction efficiency.