Automated identification of red tide phytoplanktonProrocentrum triestinum in coastal areas by image analysis

Abstract

An image analyzing system was developed for automated identification and cell counting of a red tide phytoplankton,Prorocentrum triestinum, in coastal areas. Using an interference filter, a photograph of phytoplankton was taken through a fluorescence microscope. The photograph was then fed into an image analyzing system consisting of an image sensor, a front mini-computer, and a host computer. The system identifiedP. triestinum as effectively as the conventional method in one twentieth of the time required by the latter. The system was used to examine preserved samples collected fromP. triestinum-dominated red tides in Tokyo Bay. It was found that there were two phytoplankton populations in the red tide:P. triestinum and a smaller phytoplankton.