Spherical Camera Localization in Man-made Environment Using 3D-2D Matching of Line Information

In this paper, we propose a novel method for global six degree of freedoms (DoF) localization of a spherical camera in a man-made environment. Specifically, a 3D-2D matching method based on line information of a known 3D model of the environment is proposed. There are two challenging points. First is to design a unique representation of 2D line information from the image and 3D line information from the 3D environment model. Second is to evaluate similarity of the line information extracted from both a real spherical camera image taken in the environment and arbitrary 6 DoF poses in the 3D environment model in order to localize the camera. To deal with the former, a novel descriptor is designed based on a Hough space for the line information. Then, earth mover’s distance (EMD) is calculated to evaluate similarity between the descriptors. We evaluated the proposed method in a real environment with its 3D model. The results demonstrated that our proposed method can effectively estimate the 6 DoF pose of a spherical camera using a single image.