A novel algorithm for estimating the relative rotation angle of solar azimuth through single-pixel rings from polar coordinate transformation for imaging polarization navigation sensors

Abstract

Polarized skylight provides navigation compass for many polarization sensitive birds and insects. In this paper, a method based on polar coordinate transformation (PCTA) is proposed to use all the polarization information of the polarization angle image. First, a polar coordinate system is set up on the image of angle of polarization (AOP), and then the single pixel ring corresponding to each pole diameter is extracted. After that, the corresponding point of the same pixel value is found from the two consecutive single pixel rings, and the relative rotation angle of the solar meridian is calculated according to the trigonometric relationship. Finally, simulation and outdoor experiments are performed. The algorithm is preliminarily tested and the accuracy of the simulation is in −0.004°∼+0.045°. The outdoor static experiment is carried out on a roof of a building on campus. The experiment shows that the outdoor static experimental precision is within 0.63°. In order to be applied to the estimation of the actual carrier rotation angle, the dynamic application research experiment is carried out. The center region is chosen as the true value of AOP. The error is about −0.7°∼+0.5°. It proves that the proposed algorithm is effective and accurate.