Algae classification and identification based on snapshot spectral microscopy imaging technology.

Area of Science:

• Environmental Science
• Microscopy
• Computational Biology

Background:

• Algae serve as crucial indicator species for ecosystem health and water quality.
• Accurate algae identification is vital for environmental monitoring and ecological research.
• Current methods may lack efficiency and speed for large-scale assessments.

Purpose of the Study:

• To develop and validate a rapid algae classification method using spectral microscopy and neural networks.
• To compare the performance of spectral imaging versus grayscale imaging for algae identification.
• To assess the potential of snapshot spectral microscopy for environmental applications.

Main Methods:

• Acquired spectral images of five algae species: *Anabaena azotica*, *Uronema elongatum*, *Chlorella*, *Dunaliella salina*, and *Haematococcus pluvialis*.
• Analyzed algae absorption spectral characteristics.
• Constructed ResNet50 convolutional neural network models using both grayscale and spectral images.

Main Results:

• The classification model using spectral images achieved 97.14% accuracy and 96.67% recall.
• Spectral image-based classification significantly outperformed grayscale image-based classification.
• Demonstrated the effectiveness of spectral characteristics for differentiating algae species.

Conclusions:

• Snapshot spectral microscopy integrated with neural networks offers a highly accurate method for algae identification.
• This approach provides a simple, rapid, and low-cost solution for algae classification in environmental monitoring.
• Highlights the potential of spectral imaging for ecological research and water quality assessment.