A Coarse-to-Fine Generalized-ICP Algorithm With Trimmed Strategy

Wang, Xin; Li, Yun; Peng, Yaxin; Ying, Shihui

doi:10.1109/access.2020.2976132

Cited by 17 publications

(14 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…PERCH 2.0 [21] uses GPU acceleration to significantly improve computational speed by employing GICP [37] and using parallel computing for rendering. However, ICP-based methods are susceptible to local minima [38], especially in cases of heavy occlusion. To overcome this problem, PERCH 2.0 generated a large set of initial pose hypotheses around the detected region of the target object.…”

Section: B Non-learning-based Approachesmentioning

confidence: 99%

6D Object Pose Estimation Using a Particle Filter With Better Initialization

Lee

Kim

et al. 2023

IEEE Access

View full text Add to dashboard Cite

Estimation of 6D object poses is a key issue in robotic grasping tasks. Recently, many high-performance learning-based methods have been introduced using robust deep learning techniques; however, applying these methods to real robot environments requires many ground truth 6D pose annotations for training. To address this problem, we propose a template matching-based particle filter approach for 6D pose estimation; the proposed method does not require ground truth 6D poses. Although particle filter approaches can stochastically avoid local optima, they require adequate initial pose hypotheses for estimating an accurate 6D object pose. Therefore, we estimated an initial translation of the target object for accurately initializing a particle filter by developing a new deep network. Once the proposed centroid prediction network (CPN) is trained with a specific dataset, no additional training is required for new objects not in the dataset. We evaluated the performance of the CPN and the proposed 6D pose estimation method on benchmark datasets, which demonstrated that the CPN can predict the centroid for any object, including those not in the training data, and that our 6D pose estimation method outperforms existing methods for partially occluded objects. Finally, we tested a grasping task based on our proposed method using a real robot platform to demonstrate an application of our method to a downstream task. This experiment shows that our method can be applied to part assembly, bin picking, and object manipulation without large training datasets with 6D pose annotations. The code and models are available at: https://github.com/oorrppp2/Particle_filter_approach_6D_pose_estimation.

show abstract

Section: B Non-learning-based Approachesmentioning

confidence: 99%

6D Object Pose Estimation Using a Particle Filter With Better Initialization

Lee

Kim

et al. 2023

IEEE Access

View full text Add to dashboard Cite

show abstract

“…We compare our coarse-to-fine registration algorithm with the Super4PCS coarse algorithm [33], the robust trimmed ICP [35], [36], and the coarse-to-fine adaptive generalized-ICP algorithm (AGICP) [37] We measure the error by the average of point-to-point distances, marked as TSD, on the overlap area of P and Q. Meanwhile, the stopping values of the trimmed ICP, AGICP and our fine registration method are: 1) The maximum iteration number τ = 200; 2) The TSD error ε η < e −10 ; and 3) |ε η − ε η−1 | < e −10 , respectively.…”

Section: A Experiments Designmentioning

confidence: 99%

“…3) our coarse trimmed module (named our coarse method), 4) the trimmed ICP [35], 5) AGICP [37], 6) our coarse-to-fine registration without denoising (named our method w/o denoising), 7) our coarse-to-fine registration method. In Figure 13, the first row is the data contaminated by noise N (0, 0.0003), and the second row is contaminated by N (0, 0.0007).…”

Section: Perturbation Noisementioning

confidence: 99%

“…Hence, 4PCS based algorithms are relatively robust to partial registration. On the other hand, the trimmed ICP method is introduced via overlap ratio estimation [34]- [37], which can trim outliers and makes the algorithm suitable for the relatively low overlap ratio (under 50%). However, the initial transformation and additional coarse registration procedure are still necessary, when the overlap ratio is low or the noise level is high.…”

Section: Introductionmentioning

confidence: 99%

“…Han et al proposed a hierarchical searching scheme for multi-resolution data to improve the robustness with respect to the local minimum [40]. black Recently, a coarse-to-fine GICP algorithm combines the planeto-plane and point-to-point trimmed ICP, which balances the stability and accuracy by changing the neighborhood search range from wide-base to narrow-base in each iteration [37]. However, this adaptive strategy still fails under the low overlap ratio.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

An Accelerated and Robust Partial Registration Algorithm for Point Clouds

et al. 2020

Self Cite

View full text Add to dashboard Cite

Partial registration for point clouds plays an important role in various fields such as 3D mapping reconstruction, remote sensing, unmanned driving, and cultural heritage protection. Unfortunately, partial registration is challenging due to difficulties such as the low overlap ratio of two point clouds and the perturbation in the orderless and sparse 3D point clouds. Thus, a variety of the 3D shape context descriptors are introduced for finding the optimal matching. However, extracting geometric features and descriptors are time consuming and easily degenerated by noise. To overcome these problems, we introduce a parallel coarse-to-fine partial registration method. Our contributions can be summarized as: Firstly, a robust coarse trimmed method is proposed to estimate the coarse overlap area and the initial transformation via fast bilateral denoising and parallel point feature histogram (PPFH) descriptor aligning. Secondly, an accelerated fine registration procedure is conducted by a parallel trimmed iterative closest point (PTrICP) method. Moreover, most parts of our coarse-to-fine workflow are accelerated under the Graphics Processing Unit (GPU) parallel execution mode for efficiency. Thirdly, we extend our method from the rigid registration to the isotropic scaling registration, which improves its applicability. Experiments have demonstrated that our method is feasible and robust in various situations, including the low overlap ratio, outlier, noise and scaling.

show abstract