Phytoplankton Identification Using Deep Learning YOLOv8 Algorithm and Its Implementation on Websites

This paper was not presented at the conference.

Abstract

This paper was not presented at the conference.

Phytoplankton has an important role in the ecosystem and a bioindicator of water quality. The genera Bacteriastrum, Chaetoceros, and Thalassiothrix are the dominant diatoms because of their morphology which can live in polluted environments. Identification of phytoplankton is crucial for analysis and preventing potential ecosystem damage. However, conventional methods require time and observation skills due to morphological similarities, so an easy and accurate identification method is needed. This research introduces a new approach using deep learning with the YOLOv8 algorithm to identify phytoplankton. The use of deep learning is usually only for reporting and model performance and is not applied practically, so this research aims to implement deep learning on websites so that website users can identify phytoplankton automatically, also contributing as citizen science. The performance of the model with 700 datasets and 3000 epochs produces a high accuracy of 97.14% which can properly identify three phytoplankton genera and is implemented on a website using Flask.