Large-depth-range 2D and 3D reconstruction based on unidirectional parallel single-pixel imaging.

Traditional structured-light three-dimensional (3D) reconstruction methods often lead to substantial measurement errors when measuring objects outside the depth of field (DOF), due to defocus blur of the system. This drawback restricts their use in applications that require a large working volume. In this study, we propose a unidirectional parallel single-pixel imaging (U-PSI) method to simultaneously achieve two-dimensional (2D) image deblurring and accurate 3D measurement in a large depth range by projecting unidirectional Fourier-basis patterns. The 2D point spread function (PSF) is derived from the one-dimensional (1D) light transport coefficient (LTC) based on the reciprocity between the PSF and LTC for 2D deblurring. Centroid position invariance to defocus of 1D LTC is utilized to achieve accurate and consistent large-depth-range 3D reconstruction. Experiments demonstrate that the proposed method enables simultaneous accurate 2D imaging and 3D measurement within a depth range of 360 mm, which is promising in large-scale industrial inspection, 3D digitization of cultural heritage artifacts, architectural surveying, and other related fields.