Development of an in situ zooplankton identification and counting system based on local auto-correlational masks

The authors propose a versatile and inexpensive technology for the identification and counting of plankton. The combination of local auto-correlational masks (LACMs) and discrimination analysis, which is a two-step feature extraction, is a powerful tool in the extraction of general information from images. The performance of LACMs had been investigated for the purpose of identification of zooplankton. Proof of the principle experiments was performed with images of preserved plankton under a conventional microscope. The accuracy of discrimination between any two taxa was more than 90%. The design and specifications of the submersible microscope are also presented.