Neural Network-Based Noise Suppression Algorithm for Star Images Captured During Daylight Hours

Yuchen Liu; Chunhui Zhao; Qing Xu

doi:10.3788/AOS201939.0610003

Journals >Acta Optica Sinica >Volume 39 >Issue 6 >Page 0610003 > Article

Acta Optica Sinica
Vol. 39, Issue 6, 0610003 (2019)

Neural Network-Based Noise Suppression Algorithm for Star Images Captured During Daylight Hours

Yuchen Liu^*, Chunhui Zhao, and Qing Xu

Author Affiliations

Beijing Institute of Control Engineering, China Academy of Space Technology, Beijing 100190, China

show less

DOI: 10.3788/AOS201939.0610003 Cite this Article Set citation alerts

Yuchen Liu, Chunhui Zhao, Qing Xu. Neural Network-Based Noise Suppression Algorithm for Star Images Captured During Daylight Hours[J]. Acta Optica Sinica, 2019, 39(6): 0610003 Copy Citation Text

show less

Fig. 1. Relationship between pixel response and different input parameters. (a) Solar azimuth; (b) altitude angle of observation

Download full size

Fig. 2. Schematic of network structure

Download full size

Fig. 3. Structural diagrams of different network shortcut connections. (a) Overall shortcut; (b) local shortcut

Download full size

Fig. 4. Convergence curves of network with different types of shortcut connection

Download full size

Fig. 5. Convergence curves of network with different functional structures

Download full size

Fig. 6. Processing results of different algorithms. (a) Original star image; (b) K-SVD; (c) BM3D; (d) DnCNN; (e) proposed method

Download full size

Fig. 7. Renderings of real star image after denoising and subtracting background by different algorithms. (a) Original image; (b) algorithm in Ref. [7]; (c) algorithm in Ref. [8]; (d) proposed algorithm

Download full size

Latitude season	Mid-latitude summer	Latitude season	Mid-latitude summer
Terrain	Forest	Altitude angle of observation ${μ_{1}}_{}$ /(°)	70
Weather	Sunny	Solar azimuth ${μ_{2}}_{}$ /(°)	90
Altitude H /km	8	Solar elevation θ /(°)	70

Table 1. Example of input parameters for ModTran software

Parameter	Value	Parameter	Value
F /(°)	2	N_x× ${N_{y}}_{}$ /(pixel×pixel)	512×512
${S_{pix}}_{}$ /μm	11	f /mm	161.1
D /mm	41	τ	0.80
ζ	0.75	Q	0.40
d_w	120000e	$σ_{PRNU}^{2}$	0.0001
t /ms	10	n	3
${λ_{\min}}_{}$ /nm	800	I_min	400
${λ_{\max}}_{}$ /nm	1100	I_max	3600

Table 2. Optical-system parameters of detector

Latitude season	Mid-latitude summer	Latitude season	Mid-latitude summer
Terrain	Forest	Altitude angle of observation ${μ_{1}}_{}$ /(°)	50-90
Weather	Sunny	Solar azimuth ${μ_{2}}_{}$ /(°)	55-150
Altitude H /km	8	Solar elevation θ /(°)	70

Table 3. Input parameters of ModTran software

Types of convolutional network	Dilated Conv	Downsampling Conv	Plain Conv
${t_{GPU}}_{}$ /ms	16.21	16.11	16.84
${t_{CPU}}_{}$ /s	0.87	0.22	0.86

Table 4. Running time of network with different functional structures

Algorithm	PSNR /dB
Original star image	22.36
K-SVD	25.43
BM3D	26.15
DnCNN	37.42
Proposed algorithm	37.43

Table 5. PSNR of different algorithms

Algorithm	K-SVD	BM3D	DnCNN	Proposed
${t_{GPU}}_{}$ /ms	-	-	45.43	16.84
${t_{CPU}}_{}$ /s	1.64	1.84	12.12	0.21

Table 6. Running time of different algorithms

Algorithm	${R_{SN}}_{}$ /dB
Original image	5.82
Algorithm in Ref. [8]	7.95
Algorithm in Ref. [7]	7.56
Proposed algorithm	195

Table 7. Statistical RSN of star points after denoising by different algorithms

Yuchen Liu, Chunhui Zhao, Qing Xu. Neural Network-Based Noise Suppression Algorithm for Star Images Captured During Daylight Hours[J]. Acta Optica Sinica, 2019, 39(6): 0610003

Download Citation

Set citation alerts for the article

Tools

Set citation alerts for the article

Save the article for my favorites

Paper Information