Using a 3D Computer Vision System for Inspection of Reinforced Concrete Structures

Abstract

Reinforced concrete (RC) structures need to be frequently inspected. The visual inspection is the practice in the structural engineering field. In the case where the structures are not accessible, using RealSense Camera mounted on a robot plays a vital role to detect external defects of RC members. In this paper, a RealSense Camera is used to inspect two different reinforced concrete beams built with specified surface defects. The First set is control beams by having mostly smooth surface with a small honeycombing area. The second set is RC beams that have excessive honeycombing defects in almost throughout the RC beam’s side. Intel RealSense D435 Depth Camera is employed to scan the side of beams and record the data in X, Y and Z (depth) directions while camera is moving on a robot. Also, the MATLAB toolbox is used to convert the matrix data into image processing technique and the Mesh Plot is exploited to capture the images. The results show that the camera’s images accurately depict the surface damaged areas and provide accurate representation of depths of the surface indentations.