Identifying Algae Sample In Freshwater Using Yolov8

Abstract

Automated detection of freshwater algae is essential for sustainable water quality management and the early recognition of harmful blooms. This study evaluates the performance of the object detection models YOLOv8 using exclusively secondary, open-access microscopy datasets. A total of 6,000 labelled images, encompassing 8,250 annotated instances across Chlorella, Microcystis, and Anabaena, were compiled from publicly available ecological repositories representing diverse conditions. The dataset was partitioned into training (80%) and testing (20%) subsets, with multiple cross-validation applied for robust comparison. Model performance was assessed using mean Average Precision (mAP), Precision, Recall, and F1-score metrics. The model achieved the detection accuracy of mAP = 0.94; F1 = 0.94.  Overall, the findings demonstrate that secondary data-driven machine learning frameworks can provide cost-effective, scalable, and reproducible solutions for freshwater algae monitoring. Future work should incorporate multi-source imagery, temporal dynamics, and transfer learning to improve predictive accuracy and enable real-time environmental management applications.