Pose estimation using polarimetric imaging in low-light environment

Visual pose estimation is of great significance in the field of precision measurement and navigation and positioning. However, under low-light illumination conditions, the feature extraction of traditional visual images is easy to fail, resulting in the failure of visual pose estimation. This paper proposes a pose estimation method based on polarimetric imaging under low-light illumination conditions. This method calculates the polarization information of the target environment and uses the polarization image to calculate the pose. Since the degree of polarization can highlight the contour of the target, this method combines the advantages of the polarization characteristics of the target environment to estimate the pose. Through environmental experiments, we found that the grayscale distribution of the polarization image is more uniform in the low-light environment, and the grayscale does not change significantly with the illumination. We verified the feasibility of the proposed posture estimation method based on polarimetric imaging. This method provides a technical reference for special scenarios (tunnels, underground parking lots).