Abstract: The red tide is a global marine natural disaster．In order to predict and forecast the occurrence of red tide,with the technology of flow cytometry,micro-image and image processing,a real-time harmful algae monitoring system was developed．A method based on background subtraction was used to quickly and accurately segment the algae images.In order to overcome the influence of algal cells in morphology and individual difference brought by the different growth period and environment,geometry features which are invariant to translation,rotation and scale and texture features based on GLCM were extracted.And finally,one-to-one multi-class support vector machine was adopted for identification．The experimental results show that the average recognition accuracy rate is as high as 94.37%.

Key words: Red tide,Image segmentation,Feature extraction,Algae recognition,Support vector machines