YOLOX-DG robotic detection systems for large-scale underwater concrete structures

Area of Science:

• Marine Engineering
• Robotics
• Artificial Intelligence

Background:

• Underwater concrete structures face damage, with manual inspection being inefficient and inaccurate.
• The harsh marine environment presents significant challenges for robotic applications in ocean engineering.

Purpose of the Study:

• To develop a controllable, manta ray-inspired underwater robot for collecting damage datasets of concrete structures.
• To enhance the accuracy of underwater concrete damage detection using a novel algorithm.
• To integrate the detection model into a robotic system for real-world ocean applications.

Main Methods:

• Development of a manta ray-inspired underwater robot with precise controllability.
• Proposal and implementation of the YOLOX-DG algorithm for improved damage detection accuracy.
• Integration of the YOLOX-DG algorithm into the robotic system for automated damage assessment.

Main Results:

• Successful creation of damage datasets for underwater concrete structures.
• Significant improvement in the accuracy of underwater concrete damage detection.
• Satisfactory performance of the integrated robotic detection system during ocean testing in marine harbors.

Conclusions:

• The manta ray-inspired robotic detection system provides an accurate and efficient solution for monitoring underwater concrete structures.
• The YOLOX-DG algorithm effectively enhances damage detection capabilities in challenging marine environments.
• This technology holds promise for improving the safety and maintenance of critical ocean engineering infrastructure.