3D imaging using multi-frequency fringe projection profilometry

Abstract

Fringe projection profilometry has been gaining popularity due to its noncontact, noninvasive nature and wide range of applications including manufacturing, and computer graphics. In particular, multi-frequency phase-shifting profilometry (PSP), that makes use of multi-frequency sinusoidal fringes projected onto a 3D scene, has been attracting interest because it is easy to implement and retrieves more accurate information on highly complex 3D objects while minimizing the amount of fringe patterns needed. Here, we analyze two multi-frequency PSP methods: Dual-frequency, which contains both high and low frequencies in one pattern, and Bi-frequency, which separates the frequencies, and compare them with the standard using a virtual set-up performed in Blender, a free open-source computer graphics software, and a real-world set-up that makes use of a makeshift projector.