Omnidirectional vision scan matching for robot localization in dynamic environments

Menegatti, Emanuele; Pretto, Alberto; Scarpa, Annalisa; Pagello, Enrico

doi:10.1109/tro.2006.875495

Cited by 93 publications

(46 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…a 360 degrees view in one single image). This kind of sensors has a lot of applications, such as video surveillance (Boult et al 2001) or object tracking (Chen et al 2008), and their use has become very common in robot navigation (Menegatti et al 2006) and in autonomous vehicles (Ehlgen et al 2008;Scaramuzza and Siegwart 2008). Interest points and local descriptors-based techniques, such as SIFT, have been applied to omnidirectional images due to their good performance in planar images (Goedeme et al 2005;Tamimi et al 2006;Valgren and Lilienthal 2007;Scaramuzza and Siegwart 2008).…”

Section: State-of-the-artmentioning

confidence: 99%

Scale Invariant Feature Transform on the Sphere: Theory and Applications

et al. 2011

View full text Add to dashboard Cite

A SIFT algorithm in spherical coordinates for omnidirectional images is proposed. This algorithm can generate two types of local descriptors, Local Spherical Descriptors and Local Planar Descriptors. With the first ones, point matching between two omnidirectional images can be performed, and with the second ones, the same matching process can be done but between omnidirectional and planar images. Furthermore, a planar to spherical mapping is introduced and an algorithm for its estimation is given. This mapping allows to extract objects from an omnidirectional image given their SIFT descriptors in a planar image. Several

show abstract

Section: State-of-the-artmentioning

confidence: 99%

Scale Invariant Feature Transform on the Sphere: Theory and Applications

et al. 2011

View full text Add to dashboard Cite

show abstract

“…By assuming the color change on the boundary is apparent [2], [12], the boundary pixels can be detected by an edge detection algorithm [13]. The relative horizontal distance d between a boundary point and the robot is calculated with a distance sensing function d = f (ρ), where ρ is the image distance between the boundary pixel and the image center.…”

Section: Construction Of a 3d Texture Mapmentioning

confidence: 99%

“…with expected measurements. According to the given map, proximity sensors [2] or cameras [3], [8] are used to perceive the environment. A texture map is a metric map including environmental appearance.…”

Section: Introductionmentioning

confidence: 99%

Monte Carlo Localization using 3D texture maps

Fu¹,

Tully²,

Kantor³

et al. 2011

2011 IEEE/RSJ International Conference on Intelligent Robots and Systems

View full text Add to dashboard Cite

Abstract-This paper uses KLD-based (Kullback-Leibler Divergence) Monte Carlo Localization (MCL) to localize a mobile robot in an indoor environment represented by 3D texture maps. A 3D texture map is a simplified model that includes vertical planes with colored texture information associated with each vertical plane. At each time step, a distance measurement and an observed texture from an omnidirectional camera are compared to the expected distance measurement and the expected texture according to each hypothesis of the robot's pose in an MCL framework. Compared to previous implementations of MCL, our proposed approach converges faster than distance-only MCL and localizes the robot more precisely than SIFT-based MCL. We demonstrate this new MCL algorithm for robot localization with experiments in several hallways.

show abstract

“…An alternative approach that predicts 2D range scans using reinforcement learning techniques has been presented by Michels et al [14]. Menegatti et al [15] proposed to simulate range scans from detected color transitions in omnidirectional images and to apply scan-matching and Monte-Carlo methods for localizing a mobile robot. Such color transitions are comparable to our set of edge-based features described in Section 3.3, which form the low-level input to the learning approach described in this paper.…”

Section: Related Workmentioning

confidence: 99%

A nonparametric learning approach to range sensing from omnidirectional vision

Plagemann

Stachniss

Hess

et al. 2010

Robotics and Autonomous Systems

View full text Add to dashboard Cite

We present a novel approach to estimating depth from single omnidirectional camera images by learning the relationship between visual features and range measurements available during a training phase. Our model not only yields the most likely distance to obstacles in all directions, but also the predictive uncertainties for these estimates. This information can be utilized by a mobile robot to build an occupancy grid map of the environment or to avoid obstacles during explorationtasks that typically require dedicated proximity sensors such as laser range finders or sonars. We show in this paper how an omnidirectional camera can be used as an alternative to such range sensors. As the learning engine, we apply Gaussian processes, a nonparametric approach to function regression, as well as a recently developed extension for dealing with input-dependent noise. In practical experiments carried out in different indoor environments with a mobile robot equipped with an omnidirectional camera system, we demonstrate that our system is able to estimate range with an accuracy comparable to that of dedicated sensors based on sonar or infrared light. Key words: omnidirectional vision, learning, range sensing, Gaussian processes Preprint submitted to RAS -Special Issue on Omnidirectional Robot Vision December 18, 2009Figure 1: Our system records intensity images (left) and estimates the distances to nearby obstacles (right) after having learned how visual appearance is related to depth.

show abstract

Omnidirectional vision scan matching for robot localization in dynamic environments

Cited by 93 publications

References 22 publications

Scale Invariant Feature Transform on the Sphere: Theory and Applications

Scale Invariant Feature Transform on the Sphere: Theory and Applications

Monte Carlo Localization using 3D texture maps

A nonparametric learning approach to range sensing from omnidirectional vision

Contact Info

Product

Resources

About