A multidistance constraint method for three-dimensional reconstruction with coaxial fringe projection measurement system. (September 2020)

• Record Type: Journal Article
• Authors: Ma, Mengchao; Wang, Yuyu; Ling, Xing; Deng, Huaxia; Yao, Pengcheng; Zhang, Jin; Zhong, Xiang
• Is Part Of: Optics and lasers in engineering. Volume 132(2020)
• Publication Date: 2020-09
• Subjects: Three dimensional reconstruction -- Multidistance constriant -- Coaxial system -- Fringe projection; Lasers in engineering -- Periodicals; Optical measurements -- Periodicals; Optics -- Periodicals; Lasers en ingénierie -- Périodiques; Mesures optiques -- Périodiques; Optique -- Périodiques; 621.36605
• DOI: 10.1016/j.optlaseng.2020.106103
• Languages: English
• ISSNs: 0143-8166

Abstract

Highlights: A coaxial measurement system was Established. Combined with the system, the proposed method solved the problem of the occlusion and shadowing in measured scenes. A simple and effective multidistance constraints geometry model was constructed between the absolute phase and 3D coordinates of object in coaxial measurement system. The experiments of measuring object with stepped-height object wit a deep hole, with results of the reconstruction completeness (RC) of MDC method being 98.16% and the RC of triangulation method being 77.77%, illustrates the superiority of MDC method in recovering the 3D shapes of stepped-height objects with deep holes while avoiding shadowing. The experiments of complex-shaped statue further verifies the MDC method is suitable for complex object measurement. Graphical abstract: Abstract: Fringe projection profilometry (FPP) is a popular technique for measuring three-dimensional (3D) objects and is widely used in industrial production, quality detection, and visual guidance, among other applications. However, with traditional FPP systems based on triangulation, significant changes in the surface height of the measured object can lead to occlusion and shadowing in measured scenes. To deal with this problem, a multidistance constraint (MDC) approach based on a coaxial fringe projection system is developed here. This proposed method changes the position of the projector multiple times to increase the number of geometric constraints, and every pixel on the imaging plane of the camera can then obtain more phase information from different-frequency patterns. By constructing a model of the geometric relationship between the phase and the 3D coordinates of object, depth information can be obtained by a least-squares algorithm. Experiments confirm that the MDC method can effectively recover 3D shapes of objects with stepped heights and deep holes while avoiding shadowing.