YOLOv8-C2f-Faster-EMA: An Improved Underwater Trash Detection Model Based on YOLOv8

Area of Science:

• Environmental Science
• Robotics
• Computer Vision

Background:

• Anthropogenic waste in aquatic environments degrades water quality, impacting ecosystems and human health.
• Underwater robotic technologies offer solutions for identifying and removing submerged litter.
• Existing detection algorithms face challenges with high miss and false detection rates for small debris in aquatic settings.

Purpose of the Study:

• To develop a refined algorithm for enhanced detection of small-scale underwater debris.
• To mitigate high miss and false detection rates in aquatic litter detection.
• To improve the accuracy and computational efficiency of underwater debris identification.

Main Methods:

• Introduction of the YOLOv8-C2f-Faster-EMA algorithm, optimizing backbone, neck layer, and C2f module for underwater conditions.
• Incorporation of an effective attention mechanism to enhance detection capabilities.
• Empirical comparison against the conventional YOLOv8n framework.

Main Results:

• The proposed YOLOv8-C2f-Faster-EMA algorithm demonstrated superior performance compared to YOLOv8n.
• Achieved a 6.7% increase in precision (P), a 4.1% surge in recall (R), and a 5% enhancement in mean average precision (mAP).
• The algorithm improved underwater litter detection accuracy while simplifying the computational model.

Conclusions:

• The refined YOLOv8-C2f-Faster-EMA algorithm significantly enhances underwater debris detection.
• This methodology offers a promising tool for marine conservation and pollution monitoring.
• Potential applications extend to remote sensing, improving localized surveillance precision.