Reconstruction-free fast depth segmentation via temporal orthogonal frequency fringe patterns

Deng, Ji; Li, Jian; Feng, Hao; Yu, Xiaoli; Ding, Shumeng; Han, Wei; Zeng, Zhoumo

doi:10.1016/j.optlastec.2020.106803

Cited by 3 publications

(1 citation statement)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…6,7 With the development of deep learning, neural networks are now widely utilized for 3D reconstruction, just like the depth prediction of objects from life photos. [8][9][10] For the SLP, initially, plenty of studies use networks to carry out some important steps, such as phase demodulation or unwrapping. 11 The measurement accuracy of SLP largely depends on the calibration procedure.…”

Section: Introductionmentioning

confidence: 99%

Single-shot structured light projection profilometry with SwinConvUNet

Wang

Jiaqing

et al. 2022

Opt. Eng.

View full text Add to dashboard Cite

Structured light profilometry (SLP) is now widely utilized in noncontact threedimensional (3D) reconstruction due to its convenience in dynamic measurements. Compared with classic fringe projection profilometries, multiple deep neural networks are proposed to demodulate or unwrap the fringe phase, and these networks utilize convolution layers to extract local features while omitting global characteristics. In this paper, we propose SwinConvUNet, a deep neural network for single-shot SLP that can extract local and global features simultaneously. In the network structure design, convolution layers are applied in shallow layers to extract local features, whereas transformer layers extract global features in deep layers, and an improved loss function by combining gradient-based structural similarity is employed to improve reconstruction details. The experimental results demonstrate that SwinConvUNet is more effective than the U-net model at decreasing learnable parameters while maintaining 3D reconstruction accuracy.

show abstract

Section: Introductionmentioning

confidence: 99%

Single-shot structured light projection profilometry with SwinConvUNet

Wang

Jiaqing

et al. 2022

Opt. Eng.

View full text Add to dashboard Cite

show abstract

Depth segmentation using disparity and self-adaption fringe modulation analysis

Xiao-zhen

Zhou

2021

Optics & Laser Technology

View full text Add to dashboard Cite

Fringe-based depth segmentation via minimum-fringe-period-based singular points extraction

Wu,

Zhang,

Huang

et al. 2024

Opt. Express

View full text Add to dashboard Cite

In the field of machine vision, depth segmentation plays a crucial role in dividing targets into different regions based on abrupt changes in depth. Phase-shifting depth segmentation is a technique that extracts singular points to form segmentation lines by leveraging the phase-shifting invariance of singular points in different wrapped phase maps. This makes it immune to color, texture, and camera exposure. However, current phase-shifting depth segmentation techniques face challenges in the precision of segmentation. To overcome this issue, this paper proposes a singular points extraction technique by constructing a more comprehensive threshold with the help of the minimum period of the phase map. Taking full advantage of the proposed technique, mean-value points and order singular points are accurately filtered out, and the integrity of segmentation lines in high-curvature regions can be guaranteed. During optimization processing, the precision of segmentation is improved by employing a low-cost morphology-based optimization model. Simulation results demonstrate the segmentation accuracy reaches up to 98.58% even in a noisy condition. Experimental results on different objects indicate that the proposed method exhibits good generalization and robustness.

show abstract

Reconstruction-free fast depth segmentation via temporal orthogonal frequency fringe patterns

Cited by 3 publications

References 22 publications

Single-shot structured light projection profilometry with SwinConvUNet

Single-shot structured light projection profilometry with SwinConvUNet

Depth segmentation using disparity and self-adaption fringe modulation analysis

Fringe-based depth segmentation via minimum-fringe-period-based singular points extraction

Contact Info

Product

Resources

About