Monocular visual Odometry using vertical lines in urban area

Abstract

When a mobile robot travels in an urban area, Global Positional System (GPS) signals might be obstructed by high buildings. Visual Odometry becomes a choice. Since vertical lines, such as the vertical edges of buildings, are easy to extract, insensitive to lighting conditions/shadows, and sensitive to robot movements on the ground, they are robust features if compared with regular point features or line features. In this paper, we develop a monocular vision-based odometry system which utilizes the vertical lines from the scene to estimate the robot ego-motion. The main difference from the related works is that the proposed method can work well even the ground is not flat. Furthermore, 5 degree-of-freedom of the mobile robot, including two translations and three rotations, can be obtained. Finally, some physical experiments are carried out to validate the proposed method.