LAUSR

The division of focal plane (DoFP) imaging polarimeter operates by integrating a focal plane array (FPA) with a micro-polarizer array (MPA). While the instantaneous field of view (IFoV) error is an inherent problem in such imaging sensor, which negatively affects the imaging accuracy. The influence of IFoV error can be effectively mitigated by optimizing the MPA configuration. In this paper, we propose an MPA configuration design scheme with four basic polarization angles 0°, 45°, 90° and 135° for DoFP imaging polarimeter. First, we analyze the frequency spectrum characteristics of MPA-filtered images and polarization frequency structure is proposed to record the frequency components and their positions. Then we analyze the frequency components and the crosstalk between them. Finally, we propose our design method that formulates the MPA design as a multi-objective optimization problem. And we design MPAs with better reconstruction performance of various sizes by optimizing their spectrum distribution. Besides, the reconstruction performance of many existing MPA designs were tested on statistically estimated scenes or a single image. This paper builds a database of 500 groups of infrared images of four polarization angles to test the performance of various MPA designs. Experiment results demonstrate that the proposed MPA designs outperform the state-of-the-art MPA designs in both objective and visual evaluations. Thus, the effectiveness of our proposed design method is further confirmed.