Automated region extraction and displacement detection for paving blocks adjacent to deep excavation using photogrammetry

Abstract

Construction projects in urban environments often involve deep excavations, leading to ground deformations, such as settlement or uplift, which can harm nearby infrastructure. Monitoring is crucial for ensuring stability and safety because the displacement of paving blocks can indicate subsurface deformation. Traditional methods, like using settlement markers and leveling devices, only measure specific points rather than the entire surface deformation. Recent advancements in terrestrial photogrammetry offer point cloud data (PCD) to track sidewalk displacement but still require manual definition of the monitoring zone and displacement assessments. This paper focuses on automating Region Extraction and Displacement Detection (RED2) using PCD from photogrammetry. It describes the development of an algorithm that involves extracting the regions of interest and refining the surfaces. Displacement detection was then performed by identifying displacements and removing false positives. The proposed method provides an automated solution for monitoring ground deformations and enhancing safety measures for infrastructure adjacent to excavation.