Open AccessFringe projection profilometry by conducting deep learning from its digital twin
Author(s) -
Yi Zheng,
Shaodong Wang,
Qing Li,
Beiwen Li
Publication year - 2020
Publication title -
optics express
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.394
H-Index - 271
ISSN - 1094-4087
DOI - 10.1364/oe.410428
Subject(s) - structured light 3d scanner , computer science , profilometer , artificial intelligence , projection (relational algebra) , computer vision , computer graphics (images) , computer graphics , one shot , deep learning , optics , graphics , algorithm , materials science , physics , surface finish , mechanical engineering , scanner , engineering , composite material
High-accuracy and high-speed three-dimensional (3D) fringe projection profilometry (FPP) has been widely applied in many fields. Recently, researchers discovered that deep learning can significantly improve fringe analysis. However, deep learning requires numerous objects to be scanned for training data. In this paper, we propose to build the digital twin of an FPP system and perform virtual scanning using computer graphics, which can significantly save cost and labor. The proposed method extracts 3D geometry directly from a single-shot fringe image, and real-world experiments have demonstrated the success of the virtually trained model. Our virtual scanning method can automatically generate 7,200 fringe images and 800 corresponding 3D scenes within 1.5 hours.