Detecção de anomalias estruturais em barragens utilizando uma arquitetura autoencoder

Abstract

Safety in dams is an extremely important issue, since accidents involving such constructions can cause great damage to the population and the environment. To ensure the safety of these structures, it is necessary to identify possible cracks and defects, which can lead to ruptures and fractures in the concrete. These defects are known as anomalies and can occur in any environment, from nature to computer chips. In order to guarantee such safety, this study was elaborated: initially, a literature review was carried out, in which the neural networks autoencoders stood out in the solution of such problems. Autoencoders are generative networks capable of reconstructing images, based on a dataset. For training purposes, this architecture was applied, obtaining results similar to those of the works analyzed. Thus, it is clear that such a model could be used to identify cracks, serving as a tool in the prevention of environmental disasters involving dams. [resumo fornecido pelo autor]