Color ghost imaging via sparsity constraint and non-local self-similarity

Pengwei Wang; Chenglong Wang; Cuiping Yu; Shuai Yue; Wenlin Gong; Shensheng Han

doi:10.3788/COL202119.021102

Journals >Chinese Optics Letters >Volume 19 >Issue 2 >Page 021102 > Article

Chinese Optics Letters
Vol. 19, Issue 2, 021102 (2021)

Color ghost imaging via sparsity constraint and non-local self-similarity

Pengwei Wang^1、2, Chenglong Wang¹, Cuiping Yu³, Shuai Yue³, Wenlin Gong^1、2、*, and Shensheng Han^1、2、4

Author Affiliations

¹Key Laboratory of Quantum Optics, Shanghai Institute of Optics and Fine Mechanics, Chinese Academy of Sciences, Shanghai 201800, China

²Center of Materials Science and Optoelectronics Engineering, University of Chinese Academy of Sciences, Beijing 100049, China

³Wuhan Optics Valley Aerospace Sanjiang Laser Industrial Technology Research Institute Co., Ltd., Wuhan 430075, China

⁴Hangzhou Institute for Advanced Study, University of Chinese Academy of Sciences, Hangzhou 310024, China

show less

DOI: 10.3788/COL202119.021102 Cite this Article Set citation alerts

Pengwei Wang, Chenglong Wang, Cuiping Yu, Shuai Yue, Wenlin Gong, Shensheng Han. Color ghost imaging via sparsity constraint and non-local self-similarity[J]. Chinese Optics Letters, 2021, 19(2): 021102 Copy Citation Text

EndNote(RIS)

BibTex

Plain Text

show less

Fig. 1. (a) Principle scheme of proposed color ghost imaging and (b) its sampling model.

Download full size | View in the Article

Illustration of non-local self-similarity in the spatial domain for a single-wavelength image and in the spectral domain between two wavelength images.

Fig. 2. Illustration of non-local self-similarity in the spatial domain for a single-wavelength image and in the spectral domain between two wavelength images.

Download full size | View in the Article

Fig. 3. Simulation results of three reconstruction algorithms in the case of detection

SNR = 30 dB

and k

= 10, 000

. The first row is the object and their corresponding RGB images. The second row is the results of GISC_R. The third row is the results of GISC. The fourth row is the results of GISCNL. (c) The PSNR of reconstruction results in different reconstruction algorithms. The last column is the average processing time.

Download full size | View in the Article

Fig. 4. Effect of measurement number k on the reconstruction quality when the detection

SNR = 30 dB

. The first row is the results of GISC_R. The second row is the results of GISC. The third row is the results of GISCNL. (a)–(f) Corresponding reconstruction results when the measurement number k is 4000, 6000, 8000, 10,000, 12,000, and 14,000, respectively. (

g

) The curve of PSNR versus k.

Download full size | View in the Article

Fig. 5. Influence of detection SNR on the reconstruction quality when k

= 10, 000

. The first row is the results of GISC_R. The second row is the results of GISC. The third row is the results of GISCNL. (a)–(f) Corresponding reconstruction results when the detection SNR is 15 dB, 20 dB, 25 dB, 30 dB, 35 dB, and 40 dB, respectively. (g) The curve of PSNR versus detection SNR.

Download full size | View in the Article