Detection of Steel Structures Degradation through a UAVs and Artificial Intelligence Automated System

Abstract

In recent times, the need for the management and monitoring of steel structures (bridges, but also buildings) has become more and more important; consequently, a new phase has opened up aimed at the surveillance and monitoring of these structural types with the objective of their protection and preservation, also through preventive maintenance activities. Leaving aside the world of large structures (industrial buildings, bridges, etc.), the reality of metal-framed buildings in Italy is not yet strongly established. For this reason, particular attention must be paid to these types of structures. The application of experimental monitoring techniques, however, involves the succession and chaining of various established procedures. Visual inspection is generally the first step to assess any deterioration, but it becomes quite difficult for elements at significant heights. The operational difficulties can be reduced by the UAV drone. Image processing using soft computing techniques also offers the possibility of speeding up the inspection by human operators, who can limit themselves to assessing any damaged parts already selected by artificial intelligence. It is, therefore, necessary to establish appropriate automatic or semi-automatic inspection procedures mainly aimed at providing useful indications to operators on intervention priorities. An automatic monitoring and management procedure is therefore presented, which provides for the detection and evolution of degradation on structural elements and joints of existing steel structures. The implemented methodology follows five main phases: (a) images acquisition by UAVs; (b) 3D creation with geometry and degradation; (c) data processing and defect detection; (d) creation of an "evolutionary" database, able to update the degradation on the basis of the acquisitions made in subsequent inspections by UAVs; (v) implementation of the structure (with its defects) within a structural analysis software FEM (Finite Element Method).