Automated acquisition of phytoplankton microscopic images using bright-field and fluorescence tomographic imaging

The automated phytoplankton analysis relies on high-quality microscopic images. However, conventional microscopy in thick counting chambers is limited to a single focal plane. This, together with impurity interference, causes cells distributed at other depths to be missed, leading to substantial quantification errors. To address the problem, we employ a dual-optical-path imaging system combined with tomographic scanning, utilizing fluorescence signals for localization to extract in-focus phytoplankton images from bright-field sequences via the image clarity evaluation method. The comparative analysis demonstrates high consistency with manual acquisition, maintaining a relative counting error below 10%. In high-impurity environments, the method sustains a missed detection rate of ≤10%, significantly outperforming manual counting. This work provides a reliable, interference-resistant strategy for image acquisition in complex conditions and provides reliable technical support for microscopic imaging of the phytoplankton analysis.