Monitoring Technique Using a Vision-based Single-Camera System for Reinforced Soil Retaining Wall

Abstract

Reinforced soil retaining walls are widely applied, and their frequency of collapse increases along with their use. Safety inspections are regularly conducted to ensure the structural safety of such walls. However, unexpected collapses occur for different reasons, such as design and construction problems, maintenance issues, and natural disasters including intensive rainfall. In this study, a single-camera system is proposed to evaluate the behavior of a retaining wall based on a single-perspective image. Various feature matching methods were compared to determine the optimal method for monitoring the retaining wall structure. The behaviors of the retaining wall structure were analyzed using the optimal method. The results indicate that the KAZE method provides the best results for monitoring the behaviors of a retaining wall, with errors ranging from 0.03% to 7.37%. The proposed single-camera system is widely used to evaluate the stability of a structure with high accuracy.