Development of an image-based automatic reinforcement modeling and inspection (I-ARMI) technique

Inspecting the accuracy of reinforcing bar placement in reinforced concrete (RC) structures is a critical procedure for ensuring the safety of structures. However, the current inspection methods are time-consuming and labor-intensive. Recent studies have been exploring the use of 3D scanners and depth cameras as inspection tools for reinforcing bars. Although there has been extensive research on bar length and spacing, limited numbers of studies have been carried on diameter detection, and they showed relatively low accuracy, especially for smaller bar diameters (D10 and D13). This study presents a photogrammetry-based automatic reinforcing bar information detection technique. By using readily accessible smartphones, images are acquired to generate 3D point clouds, and then automatic inspection procedures for bar diameter, length, and spacing are conducted. The efficacy of the proposed technique is experimentally investigated through the verification on laboratory-scale specimens (grid-type bar assembly and steel cage). It showed an accuracy of 97% in length and spacing the estimation of steel bars. Furthermore, it can effectively differentiate the diameter of D10 and D13 bars. This technique is expected to be utilized for accurate and rapid reinforcing bar placement inspections on construction sites.