Automated enumeration and size distribution analysis of Microcystis aeruginosa via fluorescence imaging
AbstractDue to climate change, toxic cyanobacteria and algae blooms and the associated exposure risk to humans has become a global issue. As a result, routine monitoring to evaluate cell concentrations is increasingly required to ensure safe water supplies. Current methods for cyanobacteria and algae cells enumeration are time consuming and cost-intensive due to the need for manual labor, which prevents their widespread adoption for routine water monitoring.. Automated enumeration with computer-assisted image analysis has strong potential to become a viable solution for continuous routine monitoring; however, the design of such automated systems is challenging due to: a) poor contrast between the target cells and the background, b) presence of confounding cells and abiotic particles and b) image quality variability depending on factors such as the underlying microscopy system in use and the sample condition. In this study, we introduce a novel integrated imaging-based method for automated enumeration and size distribution of Microcystis aeruginosa, a species of freshwater cyanobacteria that can originate harmful blooms. The target cells were excited using a 546nm light source and the resulting fluorescent imaging signal was acquired. A probabilistic unsupervised classification approach was taken to detect Microcystis cells from the surrounding background based on the fluorescent signal. A Gaussian mixture model was learned from the fluorescent imaging signal. The detected Microcystis cells were then enumerated and statistics regarding their size distribution automatically computed. When compared to the manual enumeration data using an hemacytometer, the developed method achieved higher accuracy using much less time and resources, without cell staining. These preliminary results demonstrate the potential of the proposed method as a powerful and robust tool for water quality monitoring and safe water quality control when used alongside gold standard methods.
How to Cite
Jin, C., Mesquita, M., Emelko, M., & Wong, A. (2016). Automated enumeration and size distribution analysis of Microcystis aeruginosa via fluorescence imaging. Journal of Computational Vision and Imaging Systems, 2(1). https://doi.org/10.15353/vsnl.v2i1.87