Spatial Resolving Power Determination of Polarization Imaging Sensors: A Multi-target Laboratory Study
摘要
Recent technological advances have enabled miniaturized polarization imaging sensors capable of simultaneous multi-state polarization capture, making them particularly suitable for drone-based aerial imaging and subsequent engineering applications. While ground sampling distance (GSD) provides a baseline metric for spatial resolution, ground resolving distance (GRD) offers a more precise characterization of true spatial resolution. This study presents the first systematic investigation of GRD in polarization imaging sensors through controlled laboratory experimentation using state-of-the-art spatial resolution calibration targets: USAF-1951 with novel sub-pixel analysis capabilities, Slanted-edge, and Siemens star patterns. For a system with 0.39 mm GSD, method-dependent variations in GRD measurements were observed: the slanted-edge method demonstrated values ranging from 1.30 to 1.33 mm across polarization angles (0°, 45°, 90°, 135°); the Siemens star method yielded 0.96–1.00 mm; and the USAF target showed 1.38–1.42 mm. Analysis of Variance (ANOVA) confirmed significant inter-method differences (F(2,9) = 595.27, p value < 0.001). The average GRD-to-GSD ratio was calculated as 3.12, with method-specific ratios of 3.54, 2.48, and 3.34 for USAF, Siemens, and slanted-edge respectively. Testing on materials with varying polarization properties revealed material-dependent resolution differences. Spatial analysis revealed minimal GRD variation across the sensor’s field of view, with measurements of 1.32 mm, 1.30 mm, and 1.39 mm for central, upper, and lower image regions respectively. Investigation of GRD characteristics in polarization-derived features—specifically the angle of linear polarization (AoLP) and degree of linear polarization (DoLP)—revealed limitations in conventional targets for these derived parameters. This study advances the understanding of spatial resolution characteristics in polarization imaging systems and provides insights for optimizing measurement methodologies in practical applications.