Red tide detection based on high spatial resolution broad band optical satellite data

Red tide, one of the major ecological disasters in world, exerts great impact on marine environment. The ocean color satellite data with low spatial resolution and high spectral is often used for red tide detection, but insufficient fine-scale detection due to its coarse resolution. To address this problem, a new method based pseudo hue angle (PHA-RI) broad band proposed paper. Different from standard International Commission Illumination (CIE) system, false-color bands near infrared (NIR), green are calculate CIE tristimulus X, Y, Z, instead true-color red, blue. With method, can be easily differentiated non-red water, distinction between water doubled compared that composite. Experiment results show PHA-RI effectively detect an averaged overall accuracy, F1-score precision 92%, 0.92 respectively. Compared traditional machine learning support vector (SVM) index algorithm Gaofen-1 (GF-1) (GF-RI), has obvious advantages especially strip distributed tide. Besides, good applicability, it proved suitable tides different dominant species including Noctiluca scintillans, Skeletonema costatum Heterosigma akashiide. Also, applicable sensors, such as Chinese satellites GF-1 Wide Field View (WFV), Haiyang 1D (HY-1D) Coastal Zone Imager (CZI), Sentinel-2 MultiSpectral Instrument (MSI) Landsat 8 Operational Land (OLI). application experiment indicates effective successfully events, which have not been detected by Moderate Resolution Imaging Spectroradiometer (MODIS) Aqua also shows detection. This work provides using data.