Computer Processing and Analysis of Scanned Zooplankton Samples: Guidelines

Abstract

Guidelines for computer analysis of images obtained by scanning zooplankton samples using ImageJ / Fiji and ilastik software have been developed. Instructions are given for post-processing of the primary pre-scanned material, which includes segmentation of zooplanker images and creating collections for subsequent recognition in the Object Classifiction / ilastik module. Creating these collections optimizes the work with large high-resolution image files and simplifies constructing a classifier. The method proved effective for studying spatial distribution of size structure in the population of copepod Arctodiaptomus salinus in Lake Shira. The technique was also tested on samples of natural zooplankton from lakes in the northern part of Minusinsk Lowlands (Republic of Khakassia) which have different species compositions and abundance of planktonic organisms. The capacity of bioimage analysis (BIA) for identification of zooplankton species is limited. However, in combination with traditional microscopy, it can be also used for studying the taxonomic structure of zooplankton. Based on the research results, the advantages, limitations and implications of BIA methods for processing zooplankton samples are discussed