A Rust Extraction and Evaluation Method for Navigation Buoys Based on Improved U-Net and Hue, Saturation, and Value

Area of Science:

• Maritime Engineering
• Computer Vision
• Materials Science

Background:

• Navigation buoys are critical for maritime safety but susceptible to degradation like rusting.
• Severe rust can compromise buoy integrity and function, necessitating effective monitoring.

Purpose of the Study:

• To develop a lightweight, automated method for extracting and evaluating rust on navigation buoys using machine vision.
• To improve the accuracy and speed of rust assessment for maritime supervision.

Main Methods:

• An improved U-Net model was employed for precise image segmentation of buoy metal components.
• RGB images were converted to HSV color space to identify optimal thresholds for rust and metal pixel extraction.
• The rust ratio was calculated to quantitatively evaluate the level of corrosion.

Main Results:

• The image segmentation achieved high precision and recall rates exceeding 0.95, with near-perfect accuracy.
• The proposed method demonstrated significantly improved accuracy and processing speed in rust grade evaluation compared to traditional image processing techniques.

Conclusions:

• The developed machine vision method offers an efficient and accurate solution for automatic rust detection and evaluation on navigation buoys.
• This technology can enhance maritime supervision capabilities and contribute to proactive buoy maintenance strategies.