Recursive Bayesian pose and shape estimation of 3D objects using transformed plane curves

Faion, Florian; Zea, Antonio; Steinbring, Jannik; Baum, Marcus; Hanebeck, Uwe D.

doi:10.1109/sdf.2015.7347698

Cited by 14 publications

(14 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Recently, an extension for 3D GPs has been published [13]. Furthermore, [14] estimates the 3D objects by using transformed planes, whereas [15] estimates the 3D vehicle contour in LiDAR measurements.…”

Section: Related Workmentioning

confidence: 99%

Fast 3D Extended Target Tracking using NURBS Surfaces

Naujoks

Burger

Wuensche

2019

2019 IEEE Intelligent Transportation Systems Conference (ITSC)

View full text Add to dashboard Cite

This paper proposes fast and novel methods to jointly estimate the target's unknown 3D shape and dynamics. Measurements are noisy and sparsely distributed 3D points from a light detection and ranging (LiDAR) sensor. The methods utilize non-uniform rational B-splines (NURBS) surfaces to approximate the target's shape. One method estimates Cartesian scaling parameters of a NURBS surface, whereas the second method estimates the corresponding NURBS weights, too. Major advantages are the capability of estimating a fully 3D shape as well as the fast processing time. Real-world evaluations with a static and dynamic vehicle show promising results compared to state-of-the-art 3D extended target tracking algorithms.

show abstract

Section: Related Workmentioning

confidence: 99%

Fast 3D Extended Target Tracking using NURBS Surfaces

Naujoks

Burger

Wuensche

2019

2019 IEEE Intelligent Transportation Systems Conference (ITSC)

View full text Add to dashboard Cite

show abstract

“…While F f and Q f matrices, which model the evolution of the extent, are specified in (11), the details of the motion dynamics designated by F and Q will be exposed in Section VIII. Finally, the initial covariance of the extent, P f 0 , is determined by the underlying GP model as presented in (9).…”

Section: B Process Modelmentioning

confidence: 99%

“…A specific adaptation of RHM to tackle people tracking using depth data is presented in [10], and therein 3D shape is approximated as a cylinder. From a similar perspective, a more general tracking framework based on the assumption that 3D object surface can be constructed by some transformations, e.g., translation, rotation, of a plane curve is proposed in [11]. However, this approach necessitates a special formulation of the recursive estimator in accordance with the particular transformation considered.…”

Section: Introductionmentioning

confidence: 99%

Three-Dimensional Extended Object Tracking and Shape Learning Using Gaussian Processes

Kumru,

Özkan

2019

Preprint

View full text Add to dashboard Cite

In this study, we investigate the problem of tracking objects with unknown shapes using three-dimensional (3D) point cloud data. We propose a Gaussian process-based model to jointly estimate object kinematics, including position, velocity and orientation, together with the shape of the object in an online fashion. We describe the unknown shape by a radial function in 3D, and induce a correlation structure via a Gaussian process. Furthermore, we propose an efficient algorithm to reduce the computational complexity of working with 3D data. This is accomplished by casting the tracking problem into projection planes which are attached to the object's local frame. The proposed methods provide an analytical expression for the object shape together with confidence intervals. The confidence intervals, which quantify the uncertainty in the shape estimate, can later be used for solving the gating and association problems inherent in object tracking. The performance of the methods is demonstrated both on simulated and real data. The results are compared with an existing random matrix model, which is commonly used for extended object tracking in the literature.

show abstract

“…As we derived a SOEKF for a closed-form measurement update in Section 4, we would like to evaluate its performance compared to Monte Carlo sampling for the moment matching in (15), (16), and (17) using 10000 samples. In order to focus on the measurement update, we first consider a stationary object.…”

Section: Stationary Ellipsementioning

confidence: 99%

“…The random hypersurface (RH) model [1,3,4,10,11,12,13,14,15] reduces the extended object tracking problem to a curve fitting problem by means of scaling the shape contours. This idea can be used for basic geometric shapes such as ellipses but also for general star-convex shapes and three-dimensional objects.…”

Section: Introductionmentioning

confidence: 99%

Extended Kalman filter for extended object tracking

Yang

Baum

2017

2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

Self Cite

View full text Add to dashboard Cite

In this paper, we propose a novel method for estimating an elliptic shape approximation of a moving extended object that gives rise to multiple scattered measurements per frame. For this purpose, we parameterize the elliptic shape with its orientation and the lengths of the semi-axes. We relate an individual measurement with the ellipse parameters by means of a multiplicative noise model and derive a second-order extended Kalman filter for a closed-form recursive measurement update. The benefits of the new method are discussed by means of Monte Carlo simulations for both static and dynamic scenarios.

show abstract

Recursive Bayesian pose and shape estimation of 3D objects using transformed plane curves

Cited by 14 publications

References 8 publications

Fast 3D Extended Target Tracking using NURBS Surfaces

Fast 3D Extended Target Tracking using NURBS Surfaces

Three-Dimensional Extended Object Tracking and Shape Learning Using Gaussian Processes

Extended Kalman filter for extended object tracking

Contact Info

Product

Resources

About