TSU-taught neural network recognizes plankton parameters for environmental monitoring

In Sirius, Tomsk State University researchers and school students created a database of holographic images of plankton, which will enable non-stop automatic environmental monitoring using a submersible digital holographic camera.

TSU First Vice-Rector Victor Dyomin

TSU First Vice-Rector Victor Dyomin explains that traditionally, biologists, limnologists, and environmentalists study plankton “manually” by drawing it from its habitat and analyzing it microscopically. Data is obtained for the entire layer of the aqueous medium, and the plankton is removed from its habitat, which affects its life cycle. A TSU photostimulation method is 10 times more sensitive than the traditional bioindication of plankton, so pollution can be registered at early stages. The researchers are planning to set up underwater observatories for this method.

“Plankton observatories are about 10 times as sensitive as standard biotesting, so ‘alarm signals’ can be picked up even at the slightest pollution. The response to the pollution varies depending on the species of plankton, so it's important to automatically differentiate species, which is what the neural network is doing in real time. As part of the research project-based learning program Great Challenges (Exploration of the Arctic and the Global Ocean track), TSU scientists and school students from all over Russia have recently compiled a large database of holographic images of plankton of the Black Sea in the waters around the learning center Sirius and taught a neural network. Bioindication by plankton is especially important in the waters around high-risk sites such as nuclear power plants, which are usually built on the coast, oil and gas platforms, and gas pipelines. It makes sense to install a few stationary underwater holographic cameras to monitor the plankton's condition. While all the other environment parameters can be monitored, too, the main idea is to pay attention only to changes in the plankton behavior for alarm signals rather than continuously monitor a multitude of parameters, the entire periodic table, and only then look more closely at the causes, which could be radiation, chemical composition, increased salinity, etc.”, says Dyomin.

He adds that installation of the observatories on Kamchatka, Sakhalin, and at Sirius is now being discussed with regional officials. In addition, the method has interested the Research Institute of Civil Defense and Emergencies.

Text: TASS. Photos courtesy of the Laboratory for Radiophysical and Optical Methods of Environmental Research, Faculty of Radiophysics, TSU