An efficient iterative pose estimation algorithm

Or, Siu-Hang; Luk, Wai-Shing; Wong, Kevin; King, Irwin

doi:10.1007/3-540-63931-4_262

Cited by 6 publications

(2 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The model-based pose tracking of the target using known 3D geometrical shape of the object and the 3D locations of the features in every frame leads to a least squares minimization problem in order to find motion parameters [25]. The state-of-the art technique which is commonly adopted in 6 DoF 3D model-based tracking is the ICP algorithm [26].…”

Section: Autonomous Rendezvous Using 3d Depth Datamentioning

confidence: 99%

Close Range Tracking of an Uncooperative Target in a Sequence of Photonic Mixer Device (PMD) Images

et al. 2018

View full text Add to dashboard Cite

This paper presents a pose estimation routine for tracking attitude and position of an uncooperative tumbling spacecraft during close range rendezvous. The key innovation is the usage of a Photonic Mixer Device (PMD) sensor for the first time during space proximity for tracking the pose of the uncooperative target. This sensor requires lower power consumption and higher resolution if compared with existing flash Light Identification Detection and Ranging (LiDAR) sensors. In addition, the PMD sensor provides two different measurements at the same time: depth information (point cloud) and amplitude of the reflected signal, which generates a grayscale image. In this paper, a hybrid model-based navigation technique that employs both measurements is proposed. The principal pose estimation technique is the iterative closed point algorithm with reverse calibration, which relies on the depth image. The second technique is an image processing pipeline that generates a set of 2D-to-3D feature correspondences between amplitude image and spacecraft model followed by the Efficient Perspective-n-Points (EPnP) algorithm for pose estimation. In this way, we gain a redundant estimation of the target's current state in real-time without hardware redundancy. The proposed navigation methodology is tested in the German Aerospace Center (DLR)'s European Proximity Operations Simulator. The hybrid navigation technique shows the capability to ensure robust pose estimation of an uncooperative tumbling target under severe illumination conditions. In fact, the EPnP-based technique allows to overcome the limitations of the primary technique when harsh illumination conditions arise.

show abstract

Section: Autonomous Rendezvous Using 3d Depth Datamentioning

confidence: 99%

Close Range Tracking of an Uncooperative Target in a Sequence of Photonic Mixer Device (PMD) Images

et al. 2018

View full text Add to dashboard Cite

show abstract

“…In this paper, we are to estimate object pose and recover its 3D structure from 2D-to-3D feature correspondences. This problem is also known as the perspective n-point problem or PnP problem [2], extrinsic camera calibration problem [3], camera localization problem [4], model-based pose estimation problem [5] or exterior orientation problem [6].…”

Section: Introductionmentioning

confidence: 99%

Pose Estimation and Structure Recovery from Point Pairs

Zhong

Yi²,

Zhao

Proceedings of the 2005 IEEE International Conference on Robotics and Automation

View full text Add to dashboard Cite

This paper presents a new feature point pairs based technique for object pose estimation and structure recovery from a single view. It first estimates rotational matrix independently, then computes translation vector and recovers the 3D structure of the object directly. Linear and nonlinear strategies are presented to estimate the rotational matrix. One is for small rotational motion and the other is used to estimate large rotational parameters. When the nonlinear technique is applied, its initial guesses are given automatically by the proposed linear estimation method. On the other hand, the presented structure recovery method is not sensitive to the rotational matrix estimation results. The proposed method is applicable to three, four or more feature points and has no constraints, such as collinear or coplanar, on their relative positions. As the number of feature points increases, the estimation results are improved while the computation cost is almost unchanged. Many experiments are performed on synthetic data and real images to demonstrate the presented technique.

show abstract