Low-Cost Hyperspectral Imaging in Macroalgae Monitoring.

Area of Science:

• Marine Ecology
• Remote Sensing
• Artificial Intelligence

Background:

• Kelp beds are crucial marine habitats requiring continuous monitoring due to environmental changes.
• Traditional RGB imaging struggles to differentiate macroalgae species with similar spectral profiles.
• High costs of existing hyperspectral imaging (HSI) systems limit widespread in situ ecological monitoring.

Purpose of the Study:

• To develop a cost-effective hyperspectral imaging (HSI) system for macroalgae monitoring.
• To assess the system's capability in differentiating macroalgae species using artificial intelligence (AI).
• To enable large-scale and automated ecological surveillance of marine environments.

Main Methods:

• Developed a low-cost HSI system using a GoPro camera and a linear variable spectral bandpass filter.
• Collected spectral data from brown algae (Fucus serratus, Fucus versiculosus) and red algae (Ceramium sp., Vertebrata byssoides) in a controlled aquatic setting.
• Utilized a one-dimensional convolutional neural network (CNN) for macroalgae classification.

Main Results:

• The custom HSI system successfully captured unique spectral fingerprints of the target macroalgae species.
• AI-powered analysis achieved high classification metrics: 99.9% precision, 89.5% recall, and 94.4% F1-score.
• Demonstrated effective differentiation of morphologically and spectrally similar macroalgae, outperforming RGB imaging.

Conclusions:

• The developed low-cost HSI system is effective for macroalgae identification and monitoring.
• This approach significantly reduces the financial barrier for deploying advanced HSI technology in marine research.
• The study supports the potential for large-scale, automated ecological monitoring of vital marine habitats.