Non-Line-of-Sight Imaging Based on Detection-Data Confocalization and Convolutional Approximation

Wenbo Wang; Qi Zhang; Yue Zheng

doi:10.3788/CJL241083

Chinese Journal of Lasers, Volume. 52, Issue 7, 0710002(2025)

Non-Line-of-Sight Imaging Based on Detection-Data Confocalization and Convolutional Approximation

Wenbo Wang, Qi Zhang, and Yue Zheng^*

Author Affiliations

School of Instrument Science and Opto-Electronics Engineering, Beihang University, Beijing 100191, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(9)

Fig. 1. Non-line-of-sight imaging process

Download full size

Fig. 2. Schematic diagram of quick positioning

Download full size

Fig. 3. Schematic of ellipsoidal interpolation module

Download full size

Fig. 4. Experimental scene

Download full size

Fig. 5. Experimental equipment and hidden objects. (a) Experimental equipment; (b) hidden object “E”; (c) hidden object “H”

Download full size

Table 1. Comparison of imaging algorithms

View table

Table 1. Comparison of imaging algorithms

Method

Formulation

Speed

Scanning pattern

Image restoration quality

FBP

L A^{T} τ

Fast

No requirement

Bad

LCT

a r g m i n_{ρ} ‖τ - k_{c} * ρ‖

Very fast

Confocal

Good

Convolutional

approximation

a r g m i n_{ρ} ‖A^{T} τ - k * ρ‖

Fast

No requirement

Very good

Table 2. Simulation conditions

View table

Table 2. Simulation conditions

Parameter

Time resolution /ps

Pixel number

Object size /（m×m）

Field of view after confocalization /mm

Distance from objectto relay wall /mm

Setting 1

60 （non-confocal）

30 （confocalized）

32×32

1×1

1.5

Setting 2

60 （non-confocal）

30 （confocalized）

32×32

0.5×0.5

0.8

Table 3. Front views of recovered images and root mean square errors in simulation

View table

Table 3. Front views of recovered images and root mean square errors in simulation

Imaging algorithm	FBP	LCT	Convolutional approximation
Image of object with size of 1.0 m×1.0 m
RMSE	0.2890	0.2733	0.2490
SSIM	0.3174	0.4756	0.5324
PSNR	8.0124	8.5632	10.2561
Image of object with size of 0.5 m×0.5 m
RMSE	0.3336	0.3903	0.3209
SSIM	0.2682	0.3825	0.4520
PSNR	7.0235	7.4556	9.2356

Table 4. Front views of hidden “E” and “H” objects and image recovery quality in experiments

View table

Table 4. Front views of hidden “E” and “H” objects and image recovery quality in experiments

Imaging algorithm	FBP	LCT	Convolutional approximation
Recovered image of “E”
RMSE	0.3770	0.3566	0.3314
SSIM	0.1460	0.3290	0.3958
PSNR	7.2601	6.9823	8.2796
Recovered image of “H”
RMSE	0.3060	0.2986	0.2730
SSIM	0.1896	0.4256	0.4590
PSNR	6.9852	7.9232	8.6234

Tools

Get Citation

Copy Citation Text

Wenbo Wang, Qi Zhang, Yue Zheng. Non-Line-of-Sight Imaging Based on Detection-Data Confocalization and Convolutional Approximation[J]. Chinese Journal of Lasers, 2025, 52(7): 0710002

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: remote sensing and sensor

Received: Jul. 23, 2024

Accepted: Nov. 14, 2024

Published Online: Apr. 15, 2025

The Author Email: Yue Zheng (zhengyue@buaa.edu.cn)

DOI:10.3788/CJL241083

CSTR:32183.14.CJL241083

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology