Hybrid system model for wind abrasion segmentation using semi-automatic classification of remote sensing...

Area of Science:

• Environmental Science
• Materials Science
• Geospatial Analysis

Background:

• Wind abrasion, driven by wind-transported particles, degrades structures, a growing concern for infrastructure in exposed locations, particularly renewable energy facilities.
• Metallic structures are highly susceptible to windblown sand particle degradation.
• Characterizing remote, vulnerable sites is challenging, necessitating remote sensing and satellite data.

Purpose of the Study:

• To identify and delineate geographic areas susceptible to wind abrasion.
• To develop a hybrid model integrating historical data and satellite image classification for vulnerability assessment.
• To contribute to the understanding of material wear and optimize resilient structure design.

Main Methods:

• Creation of a robust scientific database.
• Acquisition and management of long-term satellite and historical data.
• Semi-automatic classification of multispectral satellite images and supervised learning techniques for modeling.

Main Results:

• A hybrid model was developed and validated, demonstrating high accuracy (R² = 0.9922).
• The methodology effectively identifies and delineates wind abrasion vulnerable zones.
• The approach is presented as a generalizable method for assessing material degradation risks.

Conclusions:

• The developed hybrid model is a valuable tool for identifying wind abrasion hotspots.
• This research enhances knowledge of particle-induced material wear.
• The findings support the optimization of resilient infrastructure design in susceptible regions.