An indoor positioning framework based on panoramic visual odometry for visually impaired people

Positioning is one of the most urgent problems for assisting visually impaired people, especially in indoor environments where GPS signals are weak. In this paper, we present a positioning framework based on panoramic visual odometry for assisting visually impaired people. We introduce a panoramic annular lens to visual odometry and use the Taylor camera model to describe its projection rules. Some critical techniques in visual odometry, including direct image alignment and stereo matching, are extended to fit this camera model. Besides, an easy-to-maintain coordinate alignment method based on multi-pinhole image rectification and marker detection is also proposed to unify the results of visual odometry in the world coordinate system. We evaluate our system on both synthetic and real-world datasets in a comparative set of experiments and compare against state-of-the-art algorithms. The experiment results show that the robustness of positioning has been significantly improved by the proposed visual odometry algorithm with a panoramic annular lens and the system has the ability to provide reliable positioning results in indoor environments.