Absolute Orientation and Localization Estimation from an Omnidirectional Image

摘要

In this paper, we present a novel method for instantly and autonomously determining global-localization and pose in a wide-area outdoor environment based on a single panoramic image and a 2.5D city model. In contrast to existing method, our approach is not entirely dependent on prior GPS data and inclined to use omnidirectional visual information to provide a precise city-localization in urban scene. We estimate the orientation and localization of camera based on spherical panoramic imaging model. We evaluate the proposed method on a challenging dataset. The experiments indicate that pose precision from our method is obviously superior to that from the consumer sensors and we remain unbeatable in terms of time cost compared to previous methods.