Point cloud denoising review: from classical to deep learning-based approaches

Zhou, Lang; Sun, Guoxing; Li, Yong; Li, Weiqing; Su, Zhiyong

doi:10.1016/j.gmod.2022.101140

Cited by 39 publications

(9 citation statements)

References 72 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Essentially, it is about removing redundant points from a large-scale point set. The functional model for point cloud denoising can be defined as follows [ 17 ]:…”

Section: Methodsmentioning

confidence: 99%

“…Despite significant achievements in the field of point cloud denoising by related research [ 17 ], particularly in recent years with the infiltration of deep learning technology in the domain of point clouds, algorithms for point cloud denoising based on the deep learning approach have seen rapid development [ 18 , 19 , 20 , 21 ]. However, due to the inherent unstructured nature of point clouds and the fundamentally ill-posed nature of the point cloud denoising task, existing denoising algorithms still exhibit varying degrees of limitation [ 22 ].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Point Cloud Denoising and Feature Preservation: An Adaptive Kernel Approach Based on Local Density and Global Statistics

Wang,

Chen,

Song

et al. 2024

Sensors

View full text Add to dashboard Cite

Noise removal is a critical stage in the preprocessing of point clouds, exerting a significant impact on subsequent processes such as point cloud classification, segmentation, feature extraction, and 3D reconstruction. The exploration of methods capable of adapting to and effectively handling the noise in point clouds from real-world outdoor scenes remains an open and practically significant issue. Addressing this issue, this study proposes an adaptive kernel approach based on local density and global statistics (AKA-LDGS). This method constructs the overall framework for point cloud denoising using Bayesian estimation theory. It dynamically sets the prior probabilities of real and noise points according to the spatial function relationship, which varies with the distance from the points to the center of the LiDAR. The probability density function (PDF) for real points is constructed using a multivariate Gaussian distribution, while the PDF for noise points is established using a data-driven, non-parametric adaptive kernel density estimation (KDE) approach. Experimental results demonstrate that this method can effectively remove noise from point clouds in real-world outdoor scenes while maintaining the overall structural features of the point cloud.

show abstract

“…Essentially, it is about removing redundant points from a large-scale point set. The functional model for point cloud denoising can be defined as follows [ 17 ]:…”

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Point Cloud Denoising and Feature Preservation: An Adaptive Kernel Approach Based on Local Density and Global Statistics

Wang,

Chen,

Song

et al. 2024

Sensors

View full text Add to dashboard Cite

show abstract

“…The goal of these techniques is to automatically and directly learn latent representations for denoising from the noisy point cloud. Its overall performance on noise in the actual world is still restricted though [ 186 ]. Figure 24 shows the architecture of this model.…”

Section: Augmentationmentioning

confidence: 99%

Deep Learning for 3D Reconstruction, Augmentation, and Registration: A Review Paper

Vinodkumar,

Karabulut,

Avots

et al. 2024

Entropy

View full text Add to dashboard Cite

The research groups in computer vision, graphics, and machine learning have dedicated a substantial amount of attention to the areas of 3D object reconstruction, augmentation, and registration. Deep learning is the predominant method used in artificial intelligence for addressing computer vision challenges. However, deep learning on three-dimensional data presents distinct obstacles and is now in its nascent phase. There have been significant advancements in deep learning specifically for three-dimensional data, offering a range of ways to address these issues. This study offers a comprehensive examination of the latest advancements in deep learning methodologies. We examine many benchmark models for the tasks of 3D object registration, augmentation, and reconstruction. We thoroughly analyse their architectures, advantages, and constraints. In summary, this report provides a comprehensive overview of recent advancements in three-dimensional deep learning and highlights unresolved research areas that will need to be addressed in the future.

show abstract

“…Considering the mentioned deficiencies of mmw radar point cloud, one of mainstream research topics on mmw radar point cloud is data denoising, which involves removing unwanted data points and artifacts from the data. At present, there are various techniques available for point cloud denoising, including spatial filtering, feature-based methods, and deep learning-based approaches [6].…”

Section: Introductionmentioning

confidence: 99%

QDTree: Quasi-Density-Tree Accelerates Free Space Detection with MMW Radar Point Cloud

Li,

Yu,

et al. 2024

Advances in Transdisciplinary Engineering

View full text Add to dashboard Cite

In autonomous driving, free space detection based on 3+1D radar point cloud severely suffers from the existence of noises. To tackle this problem, we propose the novel non-data-driven Quasi-Density-Tree (QDTree) method, where no distance metrics are employed, and within the QDTree we use voting tool to analyse its nodes in terms of point cloud density or SNR, resulting noise data effectively filtered. We further split left point clouds into static and moving types according to Doppler to detect free space and Gaussian Mixture Method was used to generate safety confidence. The QDTree structure and point cloud dichotomy allow us to apply parallel computation in free space detection. Furthermore, the proposed free space detection method outperforms those state-of-the-art algorithms in terms of computation speed and result accuracy. Specifically, the proposed free space detection method consumes less than 30ms (with parallel computation, it is within 18ms) when processing 5-frame accumulated 4-corner-radar point cloud (∼5000 points) and hence it is adequate to meet real-time requirements.

show abstract

Point cloud denoising review: from classical to deep learning-based approaches

Cited by 39 publications

References 72 publications

Point Cloud Denoising and Feature Preservation: An Adaptive Kernel Approach Based on Local Density and Global Statistics

Point Cloud Denoising and Feature Preservation: An Adaptive Kernel Approach Based on Local Density and Global Statistics

Deep Learning for 3D Reconstruction, Augmentation, and Registration: A Review Paper

QDTree: Quasi-Density-Tree Accelerates Free Space Detection with MMW Radar Point Cloud

Contact Info

Product

Resources

About