Multi-angle and full-Stokes polarization multispectral images using quarter-wave plate and tunable filter

• Published in: Scientific data 2024-12, Vol.11 (1), p.1401-13
• Main Authors: Fan, Axin, Xu, Tingfa, Teng, Geer, Wang, Xi, Xu, Chang, Zhang, Yuhan, Li, Jianan
• Format: Article
• Language: English
• Subjects: 639/624/1107/510; 639/624/1107/527; Data Descriptor; Datasets; Humanities and Social Sciences; multidisciplinary; Polarization; Science; Science (multidisciplinary); Simulation
• ISSN: 2052-4463; 2052-4463
• DOI: 10.1038/s41597-024-04233-9

Polarization multispectral imaging has advanced significantly due to its robust information representation capability. Imaging application requires rigorous simulation evaluation and experimental validation using standardized datasets. However, the current full-Stokes polarization multispectral images (FSPMI) dataset, while providing simulation data, is limited by image drift and spectral bands. To overcome these limitations and supplement experimental data, this paper introduces the multi-angle and full-Stokes polarization multispectral images (MAFS-PMI) dataset. The imaging system utilizes a rotatable quarter-wave plate (QWP) and a fixed liquid crystal tunable filter (LCTF) to modulate polarization information. Meanwhile, the LCTF allows switching between multiple spectral bands. The acquired multi-angle polarization multispectral images facilitate the experimental validation of encoding strategies and reconstruction algorithms. Additionally, the derived full-Stokes polarization multispectral images enable the simulation evaluation of imaging methods. The MAFS-PMI dataset involves 73 fast axis angles (0° to 180°), four Stokes parameters, five polarization parameters, 35 spectral bands (520 nm to 690 nm), 400 × 400 pixels, and 12 distinct objects. This dataset offers a valuable resource for developing advanced imaging methods.