Simultaneous surface reconstruction and motion estimation for 3D measurements on moving samples

This paper presents a structured light profilometry (SLP) based method for surface reconstruction and simultaneous motion estimation of uniformly moving samples. The reconstruction process of conventional SLP is redesigned by minimizing an objective function that accounts for similarities between the projected patterns and the captured images, recognizing both geometric and textural features, which enables accurate estimation of the sample motion without the need for markers or surface texture. The objective function considers the geometry of the setup, which is characterized beforehand, as well as the topography, texture, and motion of the sample, which are determined optimally in an iterative optimization procedure, enabling surface reconstruction of moving samples with an accuracy of up to 11.6 µm, while simultaneously estimating the sample motion with sub-micrometer accuracy. This qualifies the proposed method for flexible application in industrial production lines, enabling the measurement of samples during transportation on a conveyor belt.