Deformation and Crack Monitoring of RC Beams Based on Computer Vision

Deformation and cracks are valuable clues for evaluating the condition of existing concrete structures. In previous studies, computer vision technology has been proven to be an efficient and accurate means for detailed deformation and crack information acquisition. But there is still a gap in determining the functional level of structures directly from the extracted information. In this study, six concrete beams with different reinforcements were loaded to failure, and the loading processes were recorded by cameras. The dynamic evolution of deformation and cracks under different load steps was analyzed by our designed image processing pipeline. It was found deformation and crack evolution were strongly related to failure mode. On this basis, a crack characteristic parameter system used for structure monitoring was established, which is composed of the initial position, width, length, direction, and occurrence time of the fracture, corresponding characterization and extraction methods also were proposed. The established parameter system can be used to solve reverse problems such as failure mode pre-diction and load estimating.