Résumé: Recently, the construction industry has been digitizing its production processes, the so-called Construction 4.0, in allusion to the paradigm of the fourth industrial revolution. The application of Deep Learning in computer vision systems has been highlighted in Construction 4.0. Thus, the main contribution of this work is to present a systematic review of Deep Learning for vision systems under Construction 4.0, considering the most cited and most recent journal articles between 2017 and 2021 from Scopus database. For this, a research method selected and analyzed 76 published papers. Six main points were evaluated in the proposed methodology: study area, computer vision applications, Deep Learning methods, hyperparameter tuning, data augmentation, and future work. The following topics stand out as relevant perspectives and directions for continued advancement in this field of research: improving Deep Learning models, increasing the quality of databases, investigating the generality techniques and optimizing processing capacity.

Langue: Anglais