Acta Optica Sinica, Volume. 45, Issue 1, 0123002(2025)
Deep Learning Optimized Liquid Crystal Microlens Array Design for Hyperspectral Reconstruction Systems
Fig. 4. Results of LCDE-IDN. (a) Comparison of reconstructed spectral curves; (b) loss function change curves
Fig. 5. Schematic diagram of structure of LC-MLA and related experimental results. (a) Structural diagram of LC-MLA and actual sample diagram; (b) comparison of transmittance of empirical LC-MLA designed based on empirical method and optimized LC-MLA designed based on LCDE-IDN method at 3.0 V and 8.0 V, with wavelength range of 400 nm to 900 nm; (c) PSF comparison results of empirical LC-MLA and optimized LC-MLA at voltage of 4.8 V; (d)(e) light field imaging results based on empirical LC-MLA and optimized LC-MLA, respectively
Fig. 6. Comparison of spectral reconstruction results of optimized LC-MLA designed by LCDE-IDN method and empirical LC-MLA designed by empirical method (insets Ⅰ‒Ⅲ are comparison of true values and spectral reconstruction curves extracted from red, orange, and blue boxes on scene, respectively)
Get Citation
Copy Citation Text
Shiqi Li, Hui Li, Chuan Qiao, Ting Zhu, Yuntao Wu. Deep Learning Optimized Liquid Crystal Microlens Array Design for Hyperspectral Reconstruction Systems[J]. Acta Optica Sinica, 2025, 45(1): 0123002
Category: Optical Devices
Received: Aug. 29, 2024
Accepted: Oct. 14, 2024
Published Online: Jan. 22, 2025
The Author Email: Li Hui (lihui00317@163.com)
CSTR:32393.14.AOS241493