LAUSR

Abstract Polarization imaging is a promising technology with great potential. Besides light intensity, it provides information in another dimension, polarization. When imaging through water, two frames of the target at two orthogonally polarized states, Imax and Imin, are needed. In most experiments, they are determined by searching for the darkest and brightest moments during polarizer rotation. But in cases where no noticeable difference can be observed, the final detection may fail. To cope with this problem, this study exploits optical correlation to assist to determine the optimal image pair. It provides reference information by comparing the similarity between two images and quantify it with parameter PCE (Peak-to-correlation energy). Final determination is based upon the fact that the optimal image pair should be in the least similarity. Experiments demonstrate that the location of optimal images can be accurately indicated through the method. Final recovery with the selected images presents the best contrast and image quality. The established strategy will enhance performance of underwater polarization imaging with better vision, higher resolution and accuracy.