Vanishing point detection using local dominant orientation signature for omnidirectional vision-based robots

Compared with monocular cameras, omnidirectional cameras provide a larger field of view and complete information about the surroundings. In this paper, a novel method based on the omnidirectional vision to tackle the tasks related to vanishing point detection is proposed. For detecting vanishing points, the orientation information in the panoramic image is first computed by the Sobel edge detection. The orientation information is next analyzed by local dominant orientation signature. Related experimental results of several scenarios, such as library and corridor, are shown to validate the feasibility of the proposed method.