Semi-autonomous inspection for concrete structures using digital models and a hybrid approach based on deep learning and photogrammetry

Abstract Bridge inspections are relied heavily on visual inspection, and usually conducted within limited time windows, typically at night, to minimize their impact traffic. This makes it difficult inspect every meter of the structure, especially for large-scale bridges with hard-to-access areas, which creates a risk missing serious defects or even safety hazards. paper presents new technique semi-automated damage detection in tunnel linings using hybrid approach based photogrammetry deep learning. The first involves reconstruct 3D model. It is shown that model sub-centimeter accuracy can be obtained after noise removal. However, removal also reduces point cloud density, making unsuitable quantification small-scale damages such as fine cracks. Therefore, captured images analyzed convolutional neural network (CNN) models enable crack segmentation. For this aim, second approach, generated by output CNN localization digital These two approaches were evaluated separate case studies, showing proposed could valuable tool assist human inspectors detecting, localizing, quantifying concrete structures.