Iterative segmentation algorithm for enhancing the accuracy of biomimetic polarized light compass

The primary challenge currently faced by bio-inspired polarized optical compasses lies in their poor environmental adaptability, resulting in low heading angle accuracy under complex weather conditions, such as cloudy and overcast days as well as when obscured by buildings. To enhance environmental adaptability, this paper first theoretically analyzes the impact of major atmospheric interference factors on polarization patterns. An analytical model is established to investigate the influence of multiple scattering on polarization patterns, and simulation data are used to verify the reliability of this model. Based on this model, the causes of errors in existing polarized compasses are analyzed. We propose an iterative symmetric segmentation image processing method that can effectively avoid the aforementioned errors while achieving high-speed computation. Experiments conducted under six different weather conditions, including cloudy days, cirrus, stratus, and building obscuration, demonstrate the high environmental adaptability of the proposed algorithm. The mean error of the heading angle is reduced from 11.7° to 1.4°. Furthermore, compared to more complex algorithms, this algorithm exhibits faster iterative convergence, enabling high-frequency navigation.