Qualitative route scene description using autonomous landmark detection

Coresets for visual summarization with applications to loop closure

“…Studied intensely between the vision [26], [25], [15], [24] and the robotics [11], [4], [16] communities, attempts have been made to allow faster loop-closure processing of larger datasets [5] and to utilize 3D information in order to increase specificity [21]. Maddern et al [18] propose a way of minimizing the effects of life-long collection of images in the mapping phase, especially with respect to location model creation.…”

Generating Dynamic Projection Images for Scene Representation and Understanding

“…In summary, if we represent a smooth moving trace of a camera in a plane as a connection of circular paths with different radius and linear paths, we can obtain a dynamic projection image that is a smooth connection of images from the consecutive circular paths. One such example is the generalized panoramic view proposed for route recognition by mobile robots [15,16]. In Fig.…”

Introduction: Machine vision research at Osaka University

“…Both types of panoramic views are first constructed to memorize a route, then used in navigating the same route by matching a newly obtained panoramic view with the previous one by means of landmarks, such as strong vertical edges or distinct regions in color or pattern. A qualitative scene description along a route, such as a red rectangle between a white tall pole and a green triangle, is constructed by automatically detecting these landmarks [36].…”