Vanishing point detection based on slope trend for omnidirectional vision-based robot localization

This paper proposes one novel method to detect vanishing points in indoor environment. The original environmental data are from a standard omnidirectional camera equipped on one mobile robot. The orientation information in the panoramic image captured by the omnidirectional camera is computed by any standard edge detection, and further analyzed by the proposed slope trend approach. Related experimental studies in different scenarios are performed to demonstrate the feasibility of the approach.