Defect Detection and Recognition of Crane Girder Structure Based on 3D Reconstruction of UAV

The traditional detection process of deformation or crack of crane main girder structure is tedious, inefficient and difficult to control safety risk. In this work, combining unmanned aerial vehicle (UAV) images with photogrammetry technology, an efficient and high-precision method for crane main girder defect detection and identification based on UAV 3D reconstruction is proposed. Firstly, the flight path, suitable for the chamfered rectangular section is selected, according to the structural characteristics of the main girder and the flight requirements of the UAV photography, and the calculation equation of flight strategy parameters is deduced. Secondly, based on 3D reconstruction methods and crucial techniques, the influence of image number, marker points and structure size effects on the accuracy of the 3D reconstruction model are investigated, separately, thus comprehensively verifying the accuracy of the 3D reconstructed model. Finally, the proposed method is tested and evaluated by the experiments for deformation detection of crane main girder structure and crack identification of scaled carton, obtaining millimeter-level accurate results, and the high precision deformation detection and crack identification using non-contact 3D reconstruction are realized. The results demonstrate that the combination of UAV and photogrammetry show potential in the deformation detection of crane main girder structure, it also has a prospect in accurately evaluating the geometry of cracks and helping structural engineers to assess the health status of structures.