Automated Analysis of Zooplankton Size and Taxonomic Composition Using the VideoPlankton Recorder

摘要

The goal of this project was to develop methods for automated analysis ofzooplankton sire and taxonomic composition using images from the Video Plankton Recorder (VPR). Our three main objectives are to: (1) use our existing hardware to develop methods for detecting, in real time (60 fields per second, fps), the presence of in-focus organisms on the video and storing the images to disk, (2) develop pattern recognition software for classification of organisms into major taxonomic groupings (copepods, chaetognaths, doliolids, etc.), and (3) develop specifications for transferring the software routines to hardware to enable real-time sorting of zooplankton according to major taxa. The first objective was accomplished in year 1 and significant progress was made on the second two objectives in year 2. Our new 3-year grant which begins 1/1/95 will allow us to complete objectives 2 and 3 and develop a system for real-time visualization of planktonic taxa distributions while at sea.