Point Cloud Registration Research Articles

Hierarchical channel-spatial interaction network for partial-to-partial point cloud registration

ABSTRACT Partial-to-partial point cloud registration is a fundamental and challenging task in the three-dimensional computer vision. In this letter, we propose a novel Hierarchical Channel-Spatial Interaction Network (HCSINet), which can hierarchically explore information interaction in terms of the channel and spatial dimensions, enabling well partial point cloud perception for partial-to-partial registration. Specifically, the Channel-wise Feature Interaction (CFI) module is regarded as the basic unit of the encoder sub-network, which achieves feature hierarchical interaction to learn the overlapping scores and intuitively improve the importance of overlapping areas. Furthermore, to improve the discrimination of the overlapping points, the Co-Attention-based Spatial Association (CA-SA) module progressively excavates multi-scale spatial correlation of overlapping points by conjointly computing collaborative attention. Experimental results verify the validity of our HCSINet, which achieves competitive partial-to-partial registration performances compared to existing traditional and deep learning-based methods. The source code is publicly available https://github.com/lxnudt/HCSINet.

Sparse-to-Dense Point Cloud Registration Based on Rotation-Invariant Features

Point cloud registration is a critical problem because it is the basis of many 3D vision tasks. With the popularity of deep learning, many scholars have focused on leveraging deep neural networks to address the point cloud registration problem. However, many of these methods are still sensitive to partial overlap and differences in density distribution. For this reason, herein, we propose a method based on rotation-invariant features and using a sparse-to-dense matching strategy for robust point cloud registration. Firstly, we encode raw points as superpoints with a network combining KPConv and FPN, and their associated features are extracted. Then point pair features of these superpoints are computed and embedded into the transformer to learn the hybrid features, which makes the approach invariant to rigid transformation. Subsequently, a sparse-to-dense matching strategy is designed to address the registration problem. The correspondences of superpoints are obtained via sparse matching and then propagated to local dense points and, further, to global dense points, the byproduct of which is a series of transformation parameters. Finally, the enhanced features based on spatial consistency are repeatedly fed into the sparse-to-dense matching module to rebuild reliable correspondence, and the optimal transformation parameter is re-estimated for final alignment. Our experiments show that, with the proposed method, the inlier ratio and registration recall are effectively improved, and the performance is better than that of other point cloud registration methods on 3DMatch and ModelNet40.

A 3D pickup estimation method for industrial parallel robots based on point cloud simplification and registration

A 3D pickup estimation method for industrial parallel robots based on point cloud simplification and registration

Data processing algorithm of cone-cylinder forgings process based on spectral graph theory and Hungarian matching

This paper presents a novel data processing algorithm. This algorithm is used to solve the problem of incomplete and misaligned of point cloud data due to the complexity of nuclear power containment cone-cylinder forgings and the limitation of laser scanner. Based on spectral graph theory and Hungarian matching, this paper first introduces the lazy random walk, and point cloud state vector is calculated during the walk to judge the local information, thereby eliminate the influence of noise. Then, characteristic edges are extracted using spectral graph theory. Additionally, the feature descriptors are calculated and the cost matrix is constructed using the feature descriptors. The Hungarian algorithm is applied for feature matching, facilitating a coarse registration of the point clouds. Finally, the improved point-to-plane iteration closest point is used for fine registration to ensure accurate alignment between point clouds. The experimental results demonstrate the algorithm's effectiveness in the registration of point clouds for nuclear power containment cone-cylinder forgings.

Accurate Registration of Cross-Modality Geometry via Consistent Clustering.

The registration of unitary-modality geometric data has been successfully explored over past decades. However, existing approaches typically struggle to handle cross-modality data due to the intrinsic difference between different models. To address this problem, in this article, we formulate the cross-modality registration problem as a consistent clustering process. First, we study the structure similarity between different modalities based on an adaptive fuzzy shape clustering, from which a coarse alignment is successfully operated. Then, we optimize the result using fuzzy clustering consistently, in which the source and target models are formulated as clustering memberships and centroids, respectively. This optimization casts new insight into point set registration, and substantially improves the robustness against outliers. Additionally, we investigate the effect of fuzzier in fuzzy clustering on the cross-modality registration problem, from which we theoretically prove that the classical Iterative Closest Point (ICP) algorithm is a special case of our newly defined objective function. Comprehensive experiments and analysis are conducted on both synthetic and real-world cross-modality datasets. Qualitative and quantitative results demonstrate that our method outperforms state-of-the-art approaches with higher accuracy and robustness. Our code is publicly available at https://github.com/zikai1/CrossModReg.

Learnable Graph Matching: A Practical Paradigm for Data Association.

Data association is at the core of many computer vision tasks, e.g., multiple object tracking, image matching, and point cloud registration. however, current data association solutions have some defects: they mostly ignore the intra-view context information; besides, they either train deep association models in an end-to-end way and hardly utilize the advantage of optimization-based assignment methods, or only use an off-the-shelf neural network to extract features. In this paper, we propose a general learnable graph matching method to address these issues. Especially, we model the intra-view relationships as an undirected graph. Then data association turns into a general graph matching problem between graphs. Furthermore, to make optimization end-to-end differentiable, we relax the original graph matching problem into continuous quadratic programming and then incorporate training into a deep graph neural network with KKT conditions and implicit function theorem. In MOT task, our method achieves state-of-the-art performance on several MOT datasets. For image matching, our method outperforms state-of-the-art methods on a popular indoor dataset, ScanNet. For point cloud registration, we also achieve competitive results.

Model order reduction by convex displacement interpolation

We present a nonlinear interpolation technique for parametric fields that exploits optimal transportation of coherent structures of the solution to achieve accurate performance. The approach generalizes the nonlinear interpolation procedure introduced in [Iollo, Taddei, J. Comput. Phys., 2022] to multi-dimensional parameter domains and to datasets of several snapshots. Given a library of high-fidelity simulations, we rely on a scalar testing function and on a point set registration method to identify coherent structures of the solution field in the form of sorted point clouds. Given a new parameter value, we exploit a regression method to predict the new point cloud; then, we resort to a boundary-aware registration technique to define bijective mappings that deform the new point cloud into the point clouds of the neighboring elements of the dataset, while preserving the boundary of the domain; finally, we define the estimate as a weighted combination of modes obtained by composing the neighboring snapshots with the previously-built mappings. We present several numerical examples for compressible and incompressible, viscous and inviscid flows to demonstrate the accuracy of the method. Furthermore, we employ the nonlinear interpolation procedure to augment the dataset of simulations for linear-subspace projection-based model reduction: our data augmentation procedure is designed to reduce offline costs — which are dominated by snapshot generation — of model reduction techniques for nonlinear advection-dominated problems.

GTINet: Global Topology-Aware Interactions for Unsupervised Point Cloud Registration

GTINet: Global Topology-Aware Interactions for Unsupervised Point Cloud Registration

Toward Meta-Shape-Based Multi-View 3D Point Cloud Registration: An Evaluation

Toward Meta-Shape-Based Multi-View 3D Point Cloud Registration: An Evaluation

OAAFormer: Robust and Efficient Point Cloud Registration Through Overlapping-Aware Attention in Transformer

OAAFormer: Robust and Efficient Point Cloud Registration Through Overlapping-Aware Attention in Transformer

Robust Point Cloud Registration via Random Network Co-Ensemble

Robust Point Cloud Registration via Random Network Co-Ensemble

SGRTmreg: A Learning-Based Optimization Framework for Multiple Pairwise Registrations.

Point cloud registration is a fundamental task in computer vision and graphics, which is widely used in 3D reconstruction, object tracking, and atlas reconstruction. Learning-based optimization and deep learning methods have been widely developed in pairwise registration due to their own distinctive advantages. Deep learning methods offer greater flexibility and enable registering unseen point clouds that are not trained. Learning-based optimization methods exhibit enhanced robustness and stability when handling registration under various perturbations, such as noise, outliers, and occlusions. To leverage the strengths of both approaches to achieve a less time-consuming, robust, and stable registration for multiple instances, we propose a novel computational framework called SGRTmreg for multiple pairwise registrations in this paper. The SGRTmreg framework utilizes three components-a Searching scheme, a learning-based optimization method called Graph-based Reweighted discriminative optimization (GRDO), and a Transfer module to achieve multi-instance point cloud registration.Given a collection of instances to be matched, a template as a target point cloud, and an instance as a source point cloud, the searching scheme selects one point cloud from the collection that closely resembles the source. GRDO then learns a sequence of regressors by aligning the source to the target, while the transfer module stores and applies the learned regressors to align the selected point cloud to the target and estimate the transformation of the selected point cloud. In short, SGRTmreg harnesses a shared sequence of regressors to register multiple point clouds to a target point cloud. We conduct extensive registration experiments on various datasets to evaluate the proposed framework. The experimental results demonstrate that SGRTmreg achieves multiple pairwise registrations with higher accuracy, robustness, and stability than the state-of-the-art deep learning and traditional registration methods.

Subpocket Similarity-Based Hit Identification for Challenging Targets: Application to the WDR Domain of LRRK2.

We herewith applied a priori a generic hit identification method (POEM) for difficult targets of known three-dimensional structure, relying on the simple knowledge of physicochemical and topological properties of a user-selected cavity. Searching for local similarity to a set of fragment-bound protein microenvironments of known structure, a point cloud registration algorithm is first applied to align known subpockets to the target cavity. The resulting alignment then permits us to directly pose the corresponding seed fragments in a target cavity space not typically amenable to classical docking approaches. Last, linking potentially connectable atoms by a deep generative linker enables full ligand enumeration. When applied to the WD40 repeat (WDR) central cavity of leucine-rich repeat kinase 2 (LRRK2), an unprecedented binding site, POEM was able to quickly propose 94 potential hits, five of which were subsequently confirmed to bind in vitro to LRRK2-WDR.

Point cloud registration via sampling-based evolutionary multitasking

Point cloud registration (PCR) is an important task in both computer vision and pattern recognition. However, since the success of most existing two-stage PCR methods largely depends on the implicit assumption that the first coarse registration stage can provide an approximately optimal initial transformation for the second fine stage, these methods suffer from low success rate when the assumption is not satisfied. Evolutionary multitasking can find better solutions for multiple tasks simultaneously by transferring knowledge among multiple related tasks. Inspired by evolutionary multitasking, this study proposes a sampling-based multitask optimization method for PCR, which solves the problem of low registration success rate by sharing knowledge between two related tasks generated by a dataset. Specifically, one task is for roughly aligning the sampled point cloud, and the other is for the original unsampled point cloud. The main purpose of the first task is to continuously provide useful knowledge to the second task so that it can achieve precise alignment. Then, a new knowledge transfer mechanism guided by historical information is proposed, which can effectively adjust the search behavior of the original unsampled task. In addition, we develop a dynamic task complement strategy to complement the accuracy that may be lost by downsampling. The results on five different datasets show that the proposed method can achieve higher registration accuracy with higher success rate compared to other state-of-the-art PCR methods.

Porównanie komercyjnego i darmowego oprogramowania do przetwarzania chmur punktów

Terrestrial laser scanning technology is becoming increasingly common for automated spatial data acquisition and digitization in the fields of surveying, civil engineering and architecture. The data from measurements made with terrestrial laser scanners are a huge array of points in space, called a point cloud, which describes the captured surface of the object under study. The point cloud processing is performed in specialized software products for handling measurements from laser scanners, which provide different possibilities for manipulating the point cloud and forming different results. The software available on the market differs according to its data processing capabilities and functionalities, application areas, methods used, manufacturer and cost. To be able to perform spatial data processing and analysis correctly and with high quality, it is important to understand the available functionalities of the different software products and their advantages and disadvantages compared to others. A comparison is made for three software packages for point cloud processing - Autodesk ReCap Pro, CloudCompare and Trimble RealWorks. The different functionalities available in the products are described and presented on small building measurements along with their performance accuracy and efficiency. The strengths and weaknesses of the different software products are identified through the comparison performed. The first section describes the basic principles of the terrestrial laser scanning method. In section two, the different point cloud processing software products on the market are presented, together with a description of the different file formats for data exchange and a theoretical section on point cloud registration, filtering, and modelling. The third section contains a presentation of the main functions and processing capabilities in Autodesk ReCap Pro, CloudCompare and Trimble RealWorks software. The fourth section describes the data used for the study, the measurements performed, their processing and results in the three software, together with an assessment of accuracy by control measurements. Section five contains conclusions and implications.

A Point Cloud Registration Algorithm Based on Weighting Strategy for 3D Indoor Spaces

Point cloud registration is a technology that aligns point cloud data from different viewpoints by computing coordinate transformations to integrate them into a specified coordinate system. Many cutting-edge fields, including autonomous driving, industrial automation, and augmented reality, require the registration of point cloud data generated by millimeter-wave radar for map reconstruction and path planning in 3D environments. This paper proposes a novel point cloud registration algorithm based on a weighting strategy to enhance the accuracy and efficiency of point cloud registration in 3D environments. This method combines a statistical weighting strategy with a point cloud registration algorithm, which can improve registration accuracy while also increasing computational efficiency. First, in 3D indoor spaces, we apply PointNet to the semantic segmentation of the point cloud. We then propose an objective weighting strategy to assign different weights to the segmented parts of the point cloud. The Iterative Closest Point (ICP) algorithm uses these weights as reference values to register the entire 3D indoor space’s point cloud. We also show a new way to perform nonlinear calculations that yield exact closed-form answers for the ICP algorithm in generalized 3D measurements. We test the proposed algorithm’s accuracy and efficiency by registering point clouds on public datasets of 3D indoor spaces. The results show that it works better in both qualitative and quantitative assessments.

Digital reconstruction of substation equipment and facility layout via LiDAR point cloud registration

Generating as-built Building information models (BIMs) is promising in power substation construction projects because they can reflect the actual conditions of facilities. However, traditional manual-designed BIMs are different from real-world scenarios due to reality gaps. In this paper, we present a new method of reconstructing the layout of power equipment and facilities in substations using LIDAR point clouds. The proposed method extracts electric equipment and facilities via object segmentation and model retrieval. In particular, we investigate PFH, FPFH and SHOT descriptors for the 3D-SIFT keypoints in the 3D shape retrieval of complex electric equipment and facilities. After the best-match model is retrieved from a model library, the layout of typical electric equipment and facilities is reconstructed by aligning the model to the scene point cloud via point cloud registration. Experimental results validate the effectiveness of the proposed method. The proposed method enhances the efficiency of generating 3D models of power substations.

The Application Characteristics and Creation of Digital Sculpture Based on Parameter Model

Recently, digital sculpture based on parametric models has emerged both domestically and internationally. To improve the accuracy and efficiency of constructing digital sculpture parameter models, this study proposes a method for constructing digital sculpture parameter models based on improved point cloud registration algorithms by iterating the nearest point algorithm and multi-scale matrix descriptors to increase the iteration efficiency and accuracy of point cloud registration algorithms. The results show that under the Bunny model, the number of three scale feature points is 2146, and the coarse registration error is 0.925 * 10-6 m; As the noise increases, the efficiency of the algorithm proposed in the study increases by an average of 10.1%. The average relative rotation error is reduced by 1-2 orders of magnitude, and the registration time is shortened by 62.7% on average. For a 30dB noise point cloud, it is reduced by one order of magnitude, and as the noise increases, the efficiency of the algorithm proposed in the study increases by an average of 10.1%. In the Dragon model, the number of three scale feature points is 2235, the coarse registration time is only 26 seconds, and the registration error is 0.285 * 10-5m; The average relative rotation error is reduced by 1-2 orders of magnitude, and the registration time is shortened by 62.7% on average. The digital sculpture creation method on the ground of parameter models proposed in this study has improved the registration accuracy and efficiency of parameter models.

RCFI-Net: A reliable correspondences evaluation and feature interaction network for fast and accurate point cloud registration

3D point cloud registration is a crucial task in computer vision, robotics, and safeguarding cultural artifacts in the digital realm. However, maintaining a balance between efficiency and accuracy during the 3D point cloud registration process continues to be a challenge. To solve this issue, a Reliable Correspondences Evaluation and Feature Interaction Network called RCFI-Net has been proposed for fast and precise registration. Our model use a two-stage approach to improve registration efficiency. In the first stage, a Transformer with position encoding network is employed to reinforce point features. Overlapping masks are learned based on the attention mechanism to identify overlapping areas, followed by the sampling of interest points that have high scores in these areas to accelerate registration. In the second stage, non-distinguishing points are eliminated, and a triangulated descriptor is introduced to further differentiate inliers and outliers and find reliable correspondences. Our proposed method has been evaluated on both public datasets and real Terracotta Warriors data, with results showing that it outperforms traditional and feature-learning methods in terms of accuracy, efficiency, and robustness.

Coherent point drift with Skewed Distribution for accurate point cloud registration

Point cloud registration methods based on Gaussian Mixture Models (GMMs) exhibit high robustness. However, GMM cannot precisely depict point clouds, because the Gaussian distribution is spatially symmetric and local surfaces of point clouds are typically non-symmetric. In this paper, we propose a novel method for rigid point cloud registration, termed coherent point drift with Skewed Distribution (Skewed CPD). Our method employs an asymmetric distribution constructed from the local surface normals and curvature radii. Compared to the Gaussian distribution, this skewed distribution provides a more accurate spatial description of points on local surfaces. Additionally, we integrate an adaptive multiplier to the covariance, which reallocates the weight of the covariance for different components in the probabilistic mixture model. We employ the EM algorithm to address this maximum likelihood estimation (MLE) issue and leverage GPU acceleration. In the M-step, we adopt an unconstrained optimization technique rooted in a Lie group and Lie algebra to attain the optimal transformation. Experimental results indicate that our method outperforms state-of-the-art methods in both accuracy and robustness. Remarkably, even without loop closure detection, the cumulative error of our approach remains minimal.