Deep learning in fringe projection: A review

Abstract

Fringe projection is widely recognized as a prominent technique for 3D measurement, owing to its non-contact nature, high precision, and exceptional spatial resolution. However, it faces challenges in achieving a delicate equilibrium between speed and accuracy, conducting measurements on intricate optical surfaces, and capturing objects requiring a high dynamic range. Deep learning, with its robust capacity to comprehensively learn and automatically extract features, holds great promise in effectively addressing the challenges encountered in fringe projection. This article offers a comprehensive analysis, discussion, and summary of the historical development, current state, and emerging trends in the applications of deep learning in fringe projection. It covers various applications, network structures, datasets, challenges, and potential future directions of utilizing deep learning in fringe projection. The aim of this paper is to provide researchers in this field with a comprehensive understanding of the latest advancements, research trends, and practical applications of this technology, thus fostering its further development and application.