Localization in Urban Environments Using a Panoramic Gist Descriptor

Abstract

Abstract-Vision-based topological localization and mapping for autonomous robotic systems have received increased research interest in recent years. The need to map larger environments requires models at different levels of abstraction and additional abilities to deal with large amounts of data efficiently. Most successful approaches for appearance-based localization and mapping with large datasets typically represent locations using local image features. We study the feasibility of performing these tasks in urban environments using global descriptors instead and taking advantage of the increasingly common panoramic datasets. This paper describes how to represent a panorama using the global gist descriptor, while maintaining desirable invariance properties for location recognition and loop detection. We propose different gist similarity measures and algorithms for appearance-based localization and an online loop-closure detection method, where the probability of loop closure is determined in a Bayesian filtering framework using the proposed image representation. The extensive experimental validation in this paper shows that their performance in urban environments is comparable with local-feature-based approaches when using wide field-of-view images.

Improving Robot Mobility by Combining Downward-Looking and Frontal Cameras

“…The most typical ways to implement global descriptors are using textons [21] and Gist [22,23]. Both descriptors are based on the same idea, that is, applying a bank of Gabor filters at various locations, scales, and orientations to a given image.…”

Understanding how camera configuration and environmental conditions affect appearance-based localization

“…Vision-based odometry or localization algorithms are usually evaluated using either front-facing cameras [1], [11], [16], [19] or cameras pointed to the side [7], [20], [21] and using average focal length lenses. Some researchers have proposed using panoramic [3], [18] or omni-directional imagery [15], which could potentially improve localization by using a much wider field of view (FOV). However, our analysis shows that a wider FOV does not necessarily improve localization accuracy.…”

Development of Height Indicators using Omnidirectional Images and Global Appearance Descriptors

“…On the one hand, some relevant landmarks or outstanding points or regions (either natural or artificial) can be extracted and described using any local descriptor that captures the appearance of the landmarks' neighbourhood trying to get invariance to position, scale and rotation [9][10][11][12]. On the other hand, each scene can be represented through a unique global appearance descriptor that contains information on the whole scene [13][14][15][16].…”