Deep Learning-Based Polarization Image Fusion for Sparse Aperture Optical Systems

Xiyu Liu; Jun Wang; Quanying Wu; Junliu Fan; Baohua Chen; Zhixiang Li; An Xu

doi:10.3788/AOS250738

Acta Optica Sinica, Volume. 45, Issue 15, 1510004(2025)

Deep Learning-Based Polarization Image Fusion for Sparse Aperture Optical Systems

Xiyu Liu^1,3, Jun Wang^1、*, Quanying Wu^1、**, Junliu Fan¹, Baohua Chen¹, Zhixiang Li^1,2, and An Xu^1,2

¹Key Laboratory of Efficient Low-Carbon Energy Conversion and Utilization of Jiangsu Provincial Higher Education Institutions, School of Physical Science and Technology, Suzhou University of Science and Technology, Suzhou 215009, Jiangsu , China

²Jiangsu Province Graduate Workstation, Suzhou FOIF Co., Ltd., Suzhou 215006, Jiangsu , China

³Jiangsu Province Graduate Workstation, Zhangjiagang Optical Instrument Co., Ltd., Suzhou 215616, Jiangsu , China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(16)

Fig. 1. Golay 3 optical system pupil map and PSF. (a) Pupil diagram; (b) 2D distribution of PSF

Download full size

Fig. 2. Golay 3 optical system MTF. (a) 2D distribution of MTF; (b) MTF curve

Download full size

Fig. 3. Experimental flowchart

Download full size

Fig. 4. PSAFNet encoder

Download full size

Fig. 5. Fusion module structure

Download full size

Fig. 6. Experimental schematic diagram

Download full size

Fig. 7. Image processing flowchart. (a) Images under different polarization directions; (b) Stokes parameter images; (c) DoLP and AoP images; (d) fused image

Download full size

Fig. 8. Mid-frequency information and edge feature extraction. (a) SA image; (b) SA mid-frequency image; (c) SA mid-frequency edge; (d) PSAFNet image; (e) PSAFNet mid-frequency image; (f) PSAFNet mid-frequency edge

Download full size

Fig. 9. Quantitative comparison of mid-frequency edge characteristics

Download full size

Fig. 10. Indoor scene experimental results. (a) WT; (b) CVT; (c) RP; (d) PCNN; (e) PFNet; (f) PSAFNet; (g) FA

Download full size

Fig. 11. Image processing flowchart. (a) Images under different polarization directions; (b) Stokes parameter images; (c) DoLP and AoP images; (d) fused image

Download full size

Fig. 12. Mid-frequency information and edge feature extraction. (a) SA image; (b) SA mid-frequency image; (c) SA mid-frequency edge; (d) PSAFNet image; (e) PSAFNet mid-frequency image; (f) PSAFNet mid-frequency edge

Download full size

Fig. 13. Quantitative comparison of mid-frequency edge characteristics

Download full size

Fig. 14. Outdoor scene experimental results. (a) WT; (b) CVT; (c) RP; (d) PCNN; (e) PFNet; (f) PSAFNet; (g) FA

Download full size

Table 1. Indoor scene image fusion evaluation metrics
View table
Table 1. Indoor scene image fusion evaluation metrics
SA WT CVT RP PCNN PFNet PSAFNet
EN 6.397 7.233 6.668 6.825 4.068 6.895 7.043
SD 48.752 43.395 28.842 43.561 49.118 54.026 79.703
MS-SSIM 0.985 0.773 0.780 0.917 0.784 0.787 0.959

Table 2. Outdoor Scene Image Fusion Evaluation Metrics
View table
Table 2. Outdoor Scene Image Fusion Evaluation Metrics
SA WT CVT RP PCNN PFNet PSAFNet
EN 7.208 7.147 6.825 7.267 5.348 6.928 7.464
SD 52.568 49.818 36.726 52.347 67.689 47.915 68.525
MS-SSIM 0.879 0.739 0.658 0.838 0.888 0.673 0.925

Tools

Get Citation

Copy Citation Text

Xiyu Liu, Jun Wang, Quanying Wu, Junliu Fan, Baohua Chen, Zhixiang Li, An Xu. Deep Learning-Based Polarization Image Fusion for Sparse Aperture Optical Systems[J]. Acta Optica Sinica, 2025, 45(15): 1510004

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites