Image Dataset of Microfauna in Activated Sludge

Area of Science:

• Environmental Science
• Ecology
• Computer Science

Background:

• Microfauna are vital for activated sludge system function and monitoring.
• Current methods rely on manual observation, lacking standardized data for automated analysis.
• Developing automated recognition models requires robust, labeled image datasets.

Purpose of the Study:

• To introduce the Microfauna Dataset in Activated Sludge (MD-AS-2025).
• To provide a standardized, publicly available image dataset for computer vision tasks.
• To support the development of automated microfauna recognition in activated sludge.

Main Methods:

• The dataset includes two components: annotated full-frame images (Dataset A) and single-object cropped images (Dataset B).
• Dataset A contains 4,000 labeled microscopic images across five microfaunal groups.
• Dataset B comprises 14,257 single-object images derived from automated detection and manual verification.
• Images were captured using a high-throughput imaging system under realistic conditions.

Main Results:

• The MD-AS-2025 dataset is now publicly available.
• The dataset is structured to reflect practical functional groupings in activated sludge analysis.
• It supports object detection, classification, and other computer vision applications.

Conclusions:

• MD-AS-2025 is a reusable and extensible resource for intelligent microfauna analysis.
• The dataset facilitates the development of advanced tools for wastewater treatment monitoring.
• It establishes a scalable framework for creating aquatic environment datasets.