A low-cost methodology based on artificial intelligence for contamination detection in microalgae production systems

In our study, we introduce an innovative, low-cost artificial intelligence approach for detecting contamination in microalgae production systems. By developing a neural network that classifies microalgae genera using spectral data (300–750 nm range) and analyzing the softmax layer output, we achieve highly accurate contamination detection. Trained on pure samples of four microalgae genera—Spirulina, Chlorella, Synechococcus, and Scenedesmus—the model demonstrated a macro F1 score of 98.64% during validation. Further testing in different photobioreactors confirmed its reliability for real-world applications, offering a practical solution for continuous monitoring without the need for costly equipment or specialized personnel. This approach enhances reactor maintenance by enabling early contamination detection, supporting more efficient and sustainable microalgae production.