Red tide identification in coastal regions based on the logical operation of multi-band spectral indices.

Optical remote sensing is an applicable approach for red tide monitoring in wide observation range. This study proposed a red tide identification method using the logical operation of multi-band spectral indices (MBSIs). Five candidate MBSIs from two categories that could potentially indicate red tide outbreaks are selected and integrated into different combinations. The reflectance spectra of red tide measured in the Xinghai Bay, China, are used to determine the optimal combination of MBSIs through the self-organized map network. According to the result, the combination of fluorescence baseline height and the spectral difference indices at red and near-infrared bands achieves the highest accuracy at 0.95 among all the combinations of MBSIs. Therefore, the logical operation of these three MBSIs is applied to Sentinel-2 satellite images to identify red tide in coastal regions. It is found that the logic operation method of MBSIs can work robustly in different sea areas and provides more reasonable results than using any single MBSI. The proposed method is able to detect the red tide that are often undetectable by human eyes with low computational complexity, and can be applied to the rapid detection and daily surveillance on the spatial distribution of red tide outbreaks. [ABSTRACT FROM AUTHOR]