A System for Identification of Marine Phytoplankton

摘要

Identification of marine phytoplankton microscopic images is one of most popular topics in recent research. In this article, we present a system that assists scientists in identifying kinds of marine phytoplankton. First, we built a large image database which represented common phytoplankton species of China's seas areas. More than that, our database will keep expanding because we allow users to submit their own phytoplankton images. Second, we provided an automatic recognition module. In this module, we completed image segmentation with improved region growing, feature extraction based moments, finally classification via SVM and so on. Third, we developed some services to serve the specialist and nonspecialists. In one word, with this system, nonspecialists without professional knowledge can identify any species in given images, and specialists have a brand-new working platform to try their own ideas in marine phytoplankton identification.