Robust Point Cloud Research Articles

Graph-based robust 3D point cloud map merging approach for large scale

Graph-based robust 3D point cloud map merging approach for large scale

Robust and correspondence-free point cloud registration: an extended approach with multiple hypotheses evaluation

AbstractPoint cloud registration is a fundamental problem in robotics, critical for tasks such as localization and mapping. Most approaches to this problem use feature-based techniques. However, these approaches have issues when dealing with unstructured environments where meaningful features are difficult to extract. Recently, an innovative global point cloud registration algorithm, PHASER, which does not rely on geometric features or point correspondences, has been introduced. It leverages Fourier transforms to identify the optimal rigid transform that maximizes cross-correlation between source and target point clouds. PHASER can also incorporate additional data channels, like LiDAR intensity, to enhance registration results. Because it does not rely on local features and because of its ability to exploit additional data, PHASER is particularly useful when dealing with very noisy point clouds or with many outliers. For this reason, we propose an extension to PHASER that considers multiple plausible rototranslation hypotheses. Our extended approach outperforms the original PHASER algorithm, especially in challenging scenarios where point clouds are widely separated. We validate its effectiveness on the DARPA SubT, and the Newer College datasets, showcasing its potential for improving registration accuracy in complex environments.

Fast Robust Point Cloud Registration Based on Compatibility Graph and Accelerated Guided Sampling

Point cloud registration is a crucial technique in photogrammetry, remote sensing, etc. A generalized 3D point cloud registration framework has been developed to estimate the optimal rigid transformation between two point clouds using 3D key point correspondences. However, challenges arise due to the uncertainty in 3D key point detection techniques and the similarity of local surface features. These factors often lead to feature descriptors establishing correspondences containing significant outliers. Current point cloud registration algorithms are typically hindered by these outliers, affecting both their efficiency and accuracy. In this paper, we propose a fast and robust point cloud registration method based on a compatibility graph and accelerated guided sampling. By constructing a compatible graph with correspondences, a minimum subset sampling method combining compatible edge sampling and compatible vertex sampling is proposed to reduce the influence of outliers on the estimation of the registration parameters. Additionally, an accelerated guided sampling strategy based on preference scores is presented, which effectively utilizes model parameters generated during the iterative process to guide the sampling toward inliers, thereby enhancing computational efficiency and the probability of estimating optimal parameters. Experiments are carried out on both synthetic and real-world data. The experimental results demonstrate that our proposed algorithm achieves a significant balance between registration accuracy and efficiency compared to state-of-the-art registration algorithms such as RANSIC and GROR. Even with up to 2000 initial correspondences and an outlier ratio of 99%, our algorithm achieves a minimum rotation error of 0.737° and a minimum translation error of 0.0201 m, completing the registration process within 1 s.

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.

GeeNet: robust and fast point cloud completion for ground elevation estimation towards autonomous vehicles

GeeNet: robust and fast point cloud completion for ground elevation estimation towards autonomous vehicles

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

ModelNet-O: A large-scale synthetic dataset for occlusion-aware point cloud classification

Recently, 3D point cloud classification has made significant progress with the help of many datasets. However, these datasets do not reflect the incomplete nature of real-world point clouds caused by occlusion, which limits the practical application of current methods. To bridge this gap, we propose ModelNet-O, a large-scale synthetic dataset of 123,041 samples that emulates real-world point clouds with self-occlusion caused by scanning from monocular cameras. ModelNet-O is 10 times larger than existing datasets and offers more challenging cases to evaluate the robustness of existing methods. Our observation on ModelNet-O reveals that well-designed sparse structures can preserve structural information of point clouds under occlusion, motivating us to propose a robust point cloud processing method that leverages a critical point sampling (CPS) strategy in a multi-level manner. We term our method PointMLS. Through extensive experiments, we demonstrate that our PointMLS achieves state-of-the-art results on ModelNet-O and competitive results on regular datasets such as ModelNet40 and ScanObjectNN, and we also demonstrate its robustness and effectiveness. Code available: https://github.com/fanglaosi/ModelNet-O_PointMLS.

Realistic reconstruction of trees from sparse images in volumetric space

The realistic reconstruction of real-world trees is a challenging task in the community of computer graphics because natural trees have complex structures of branches and leaves. Existing terrestrial laser scanning (TLS) system is able to capture dense and precise tree point clouds, yet the TLS system is expensive and not easy to carry around. An alternative low-cost and portable way is the reconstruction of tree point cloud from multiple view images. However, it is usually difficult to reconstruct a complete tree point cloud because of the texture similarity of branches and leaves as well as the lack of a sufficient number of images. Thus, we propose a new approach for reconstructing tree point clouds and geometries from sparse images. We first infer the camera parameters of each image, and then calculate the bounding volume of a tree from the camera parameters. Next, we set the mask of each image and the resolution of voxel, and then project each voxel in 3D space to all the mask images to determine the validity of the voxel. To alleviate the miss deletion of valid voxel, we utilize a boundary threshold and adjust mask resolution for robust point cloud reconstruction. Finally, an efficient tree reconstruction method is proposed to generate plausible tree geometries. We tested 6 different tree species that contain deciduous and evergreen trees, and the results showed that our approach is able to generate complete tree point cloud and realistic tree models even from a few number of images.

Robust Point Cloud Registration for Aircraft Engine Pipeline Systems.

Aircraft engine systems are composed of numerous pipelines. It is crucial to regularly inspect these pipelines to detect any damages or failures that could potentially lead to serious accidents. The inspection process typically involves capturing complete 3D point clouds of the pipelines using 3D scanning techniques from multiple viewpoints. To obtain a complete and accurate representation of the aircraft pipeline system, it is necessary to register and align the individual point clouds acquired from different views. However, the structures of aircraft pipelines often appear similar from different viewpoints, and the scanning process is prone to occlusions, resulting in incomplete point cloud data. The occlusions pose a challenge for existing registration methods, as they can lead to missing or wrong correspondences. To this end, we present a novel registration framework specifically designed for aircraft pipeline scenes. The proposed framework consists of two main steps. First, we extract the point feature structure of the pipeline axis by leveraging the cylindrical characteristics observed between adjacent blocks. Then, we design a new 3D descriptor called PL-PPFs (Point Line-Point Pair Features), which combines information from both the pipeline features and the engine assembly line features within the aircraft pipeline point cloud. By incorporating these relevant features, our descriptor enables accurate identification of the structure of the engine's piping system. Experimental results demonstrate the effectiveness of our approach on aircraft engine pipeline point cloud data.

Robust point cloud registration for map-based autonomous robot navigation

Robust point cloud registration for map-based autonomous robot navigation

R-PointNet: Robust 3D Object Recognition Network for Real-World Point Clouds Corruption

Point clouds obtained with 3D scanners in realistic scenes inevitably contain corruption, including noise and outliers. Traditional algorithms for cleaning point cloud corruption require the selection of appropriate parameters based on the characteristics of the scene, data, and algorithm, which means that their performance is highly dependent on the experience and adaptation of the algorithm itself to the application. Three-dimensional object recognition networks for real-world recognition tasks can take the raw point cloud as input and output the recognition results directly. Current 3D object recognition networks generally acquire uniform sampling points by farthest point sampling (FPS) to extract features. However, sampled defective points from FPS lower the recognition accuracy by affecting the aggregated global feature. To deal with this issue, we design a compensation module, named offset-adjustment (OA). It can adaptively adjust the coordinates of sampled defective points based on neighbors and improve local feature extraction to enhance network robustness. Furthermore, we employ the OA module to build an end-to-end network based on PointNet++ framework for robust point cloud recognition, named R-PointNet. Experiments show that R-PointNet reaches state-of-the-art performance by 92.5% of recognition accuracy on ModelNet40, and significantly outperforms previous networks by 3–7.7% on the corruption dataset ModelNet40-C for robustness benchmark.

PCKRF: Point Cloud Completion and Keypoint Refinement With Fusion Data for 6D Pose Estimation.

Some robust point cloud registration approaches with controllable pose refinement magnitude, such as ICP and its variants, are commonly used to improve 6D pose estimation accuracy. However, the effectiveness of these methods gradually diminishes with the advancement of deep learning techniques and the enhancement of initial pose accuracy, primarily due to their lack of specific design for pose refinement. In this paper, we propose Point Cloud Completion and Keypoint Refinement with Fusion Data (PCKRF), a new pose refinement pipeline for 6D pose estimation. The pipeline consists of two steps. First, it completes the input point clouds via a novel pose-sensitive point completion network. The network uses both local and global features with pose information during point completion. Then, it registers the completed object point cloud with the corresponding target point cloud by our proposed Color supported Iterative KeyPoint (CIKP) method. The CIKP method introduces color information into registration and registers a point cloud around each keypoint to increase stability. The PCKRF pipeline can be integrated with existing popular 6D pose estimation methods, such as the full flow bidirectional fusion network, to further improve their pose estimation accuracy. Experiments demonstrate that our method exhibits superior stability compared to existing approaches when optimizing initial poses with relatively high precision. Notably, the results indicate that our method effectively complements most existing pose estimation techniques, leading to improved performance in most cases. Furthermore, our method achieves promising results even in challenging scenarios involving textureless and symmetrical objects. Our source code is available at https://github.com/zhanhz/KRF.

Depth-Guided Robust and Fast Point Cloud Fusion NeRF for Sparse Input Views

Novel-view synthesis with sparse input views is important for real-world applications like AR/VR and autonomous driving. Recent methods have integrated depth information into NeRFs for sparse input synthesis, leveraging depth prior for geometric and spatial understanding. However, most existing works tend to overlook inaccuracies within depth maps and have low time efficiency. To address these issues, we propose a depth-guided robust and fast point cloud fusion NeRF for sparse inputs. We perceive radiance fields as an explicit voxel grid of features. A point cloud is constructed for each input view, characterized within the voxel grid using matrices and vectors. We accumulate the point cloud of each input view to construct the fused point cloud of the entire scene. Each voxel determines its density and appearance by referring to the point cloud of the entire scene. Through point cloud fusion and voxel grid fine-tuning, inaccuracies in depth values are refined or substituted by those from other views. Moreover, our method can achieve faster reconstruction and greater compactness through effective vector-matrix decomposition. Experimental results underline the superior performance and time efficiency of our approach compared to state-of-the-art baselines.

PointCVaR: Risk-Optimized Outlier Removal for Robust 3D Point Cloud Classification

With the growth of 3D sensing technology, the deep learning system for 3D point clouds has become increasingly important, especially in applications such as autonomous vehicles where safety is a primary concern. However, there are growing concerns about the reliability of these systems when they encounter noisy point clouds, either occurring naturally or introduced with malicious intent. This paper highlights the challenges of point cloud classification posed by various forms of noise, from simple background noise to malicious adversarial/backdoor attacks that can intentionally skew model predictions. While there's an urgent need for optimized point cloud denoising, current point outlier removal approaches, an essential step for denoising, rely heavily on handcrafted strategies and are not adapted for higher-level tasks, such as classification. To address this issue, we introduce an innovative point outlier cleansing method that harnesses the power of downstream classification models. Using gradient-based attribution analysis, we define a novel concept: point risk. Drawing inspiration from tail risk minimization in finance, we recast the outlier removal process as an optimization problem, named PointCVaR. Extensive experiments show that our proposed technique not only robustly filters diverse point cloud outliers but also consistently and significantly enhances existing robust methods for point cloud classification. A notable feature of our approach is its effectiveness in defending against the latest threat of backdoor attacks in point clouds.

Deep Correspondence Matching-Based Robust Point Cloud Registration of Profiled Parts

Due to ability to estimate the spatial transformation of coordinate frames, point cloud registration is a fundamental technique in manufacturing. Previous methods prone to converge to wrong local minima, in the cases of large initialization, noise, outliers, and partiality. This study presents a new learning-based robust point cloud registration approach to predict a rigid transformation in a one-shot way. Our network aims to determine a matchability matrix to yield an accurate registration result. Each element of the matchability matrix refers to similarity of learned per-point embeddings and represents the probability of a potential correspondence. Two major blocks are developed to guide the matchability matrix to represent correct correspondences: 1) an attention block is introduced to enhance the discriminativeness of learned per-point embeddings; 2) a zero-mean Gaussian based annealing layer and a differentiable Sinkhorn normalization layer are designed to enforce a permutation matchability matrix. With the matchability matrix, an intuitive solution is integrated to obtain the relative transformation of the source and target point clouds. Different from existing work, our network can handle partially overlapped point-cloud pairs effectively. Experimental results demonstrate the superiority of the proposed approach over the state-of-the-art registration approaches in terms of accuracy and robustness.

Non-Repetitive Scanning LiDAR Sensor for Robust 3D Point Cloud Registration in Localization and Mapping Applications.

Three-dimensional point cloud registration is a fundamental task for localization and mapping in autonomous navigation applications. Over the years, registration algorithms have evolved; nevertheless, several challenges still remain. Recently, non-repetitive scanning LiDAR sensors have emerged as a promising 3D data acquisition tool. However, the feasibility of this type of sensor to leverage robust point cloud registration still needs to be ascertained. In this paper, we explore the feasibility of one such LiDAR sensor with a Spirograph-type non-repetitive scanning pattern for robust 3D point cloud registration. We first characterize the data of this unique sensor; then, utilizing these results, we propose a new 3D point cloud registration method that exploits the unique scanning pattern of the sensor to register successive 3D scans. The characteristic equations of the unique scanning pattern, determined during the characterization phase, are used to reconstruct a perfect scan at the target distance. The real scan is then compared with this reconstructed scan to extract objects in the scene. The displacement of these extracted objects with respect to the center of the unique scanning pattern is compared in successive scans to determine the transformations that are then used to register these scans. The proposed method is evaluated on two real and different datasets and compared with other state-of-the-art registration methods. After analysis, the performance (localization and mapping results) of the proposed method is further improved by adding constraints like loop closure and employing a Curve Fitting Derivative Filter (CFDT) to better estimate the trajectory. The results clearly demonstrate the suitability of the sensor for such applications. The proposed method is found to be comparable with other methods in terms of accuracy but surpasses them in performance in terms of processing time.

SUCOFT: Robust Point Cloud Registration Based on Guaranteed Supercore Maximization and Flexible Thresholding

SUCOFT: Robust Point Cloud Registration Based on Guaranteed Supercore Maximization and Flexible Thresholding

PointDifformer: Robust Point Cloud Registration With Neural Diffusion and Transformer

PointDifformer: Robust Point Cloud Registration With Neural Diffusion and Transformer

Robust Point Cloud Registration in Robotic Inspection With Locally Consistent Gaussian Mixture Model

Robust Point Cloud Registration in Robotic Inspection With Locally Consistent Gaussian Mixture Model