Vison for robotics - A tool for model-based object tracking

Vincze, Markus; Schlemmer, Matthias J.; Gemeiner, Peter; Ayromlou, Minu

doi:10.1109/mra.2005.1577040

Cited by 19 publications

(16 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Other methods, (e.g., Dickmanns and Graefe 1988;Lu et al 2000;Marchand et al 2001;Drummond and Cipolla 2002;Comport et al 2003;Kim and Kweon 2003;Kyrki and Schmock 2005;Vincze et al 2005), track local features such as corners. This would limit the system to use a small subset of the available visible data (e.g., corners or line-segments).…”

Section: Visual Object Trackingmentioning

confidence: 99%

“…Both Kyrki and Schmock (2005) and Vincze et al (2005) use model-based techniques that rely on tracking features such as points and lines. Whilst a variety of feature-based tracking techniques have been proposed in the literature, they are not well suited to tracking objects with complex surface features and use only a small subset of the available data.…”

Section: Object Tracking For Autonomous Navigationmentioning

confidence: 99%

“…One key advantage of feature-based methods is that, due to using only a subset of the data, it is easier to achieve real-time performance when compared to using image-based techniques. Indeed, of all the realtime pose trackers noted above (Jurie and Dhome 2002;Dickmanns and Graefe 1988;Lu et al 2000;Marchand et al 2001;Drummond and Cipolla 2002;Comport et al 2003;Kim and Kweon 2003;Kyrki and Schmock 2005;Vincze et al 2005), all except (Jurie and Dhome 2002) use feature-based methods. Since image-based techniques are, generally, more computationally intensive, achieving realtime performance is difficult.…”

Section: Object Tracking For Autonomous Navigationmentioning

confidence: 99%

“…The key differences lie in the type of model and method of matching the model to input data. In (Dickmanns and Graefe 1988;Marchand et al 2001;Drummond and Cipolla 2002;Comport et al 2003;Kim and Kweon 2003;Kyrki and Schmock 2005;Vincze et al 2005), the model is a set of known features on the target's 3D surface and these features are matched to features detected in the input images. In the proposed technique, a texture-mapped visual model is used, coupled with optical flow, allowing all surface details to be used for tracking.…”

Section: Object Tracking For Autonomous Navigationmentioning

confidence: 99%

“…All of the trackers in (Marchand et al 2001;Drummond and Cipolla 2002;Comport et al 2003;Kim and Kweon 2003;Kyrki and Schmock 2005;Vincze et al 2005) track with pixel accuracy or better. The proposed methodology tracks with sub-pixel accuracy.…”

Section: Comparison With Feature-based Trackersmentioning

confidence: 99%

See 4 more Smart Citations

Visual-model-based, real-time 3D pose tracking for autonomous navigation: methodology and experiments

Ruiter

Benhabib

2008

Auton Robot

View full text Add to dashboard Cite

This paper presents a novel 3D-model-based computer-vision method for tracking the full six degreeof-freedom (dof) pose (position and orientation) of a rigid body, in real-time. The methodology has been targeted for autonomous navigation tasks, such as interception of or rendezvous with mobile targets. Tracking an object's complete six-dof pose makes the proposed algorithm useful even when targets are not restricted to planar motion (e.g., flying or rough-terrain navigation). Tracking is achieved via a combination of textured model projection and optical flow. The main contribution of our work is the novel combination of optical flow with z-buffer depth information that is produced during model projection. This allows us to achieve six-dof tracking with a single camera.A localized illumination normalization filter also has been developed in order to improve robustness to shading. Real-time operation is achieved using GPU-based filters and a new data-reduction algorithm based on colour-gradient redundancy, which was developed within the framework of our project. Colour-gradient redundancy is an important property of colour images, namely, that the gradients of all colour channels are generally aligned. Exploiting this property provides a threefold increase in speed. A processing rate of approximately 80 to 100 fps has been obtained in our work when utilizing synthetic and real target-motion sequences. Sub-pixel accuracies were obtained in tests performed under different lighting conditions.

show abstract