Improvement of 3D reconstruction based on a new 3D point cloud filtering algorithm

Hazzat, Soulaiman El; Merras, Mostafa

doi:10.1007/s11760-022-02474-y

Cited by 4 publications

(2 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The effect of direct filtering depends on the heuristic threshold [ 37 ]. Gaussian filtering is applicable only to point clouds that follow a normal distribution [ 38 ], bilateral filtering is applicable only to ordered point clouds [ 39 ], and statistical filtering requires a large amount of calculation [ 40 ]. As a result, the voxel filtering method is used in this paper.…”

Section: Methodsmentioning

confidence: 99%

LiDAR Dynamic Target Detection Based on Multidimensional Features

Xu,

Gao,

Sui

et al. 2024

Sensors

View full text Add to dashboard Cite

To address the limitations of LiDAR dynamic target detection methods, which often require heuristic thresholding, indirect computational assistance, supplementary sensor data, or postdetection, we propose an innovative method based on multidimensional features. Using the differences between the positions and geometric structures of point cloud clusters scanned by the same target in adjacent frame point clouds, the motion states of the point cloud clusters are comprehensively evaluated. To enable the automatic precision pairing of point cloud clusters from adjacent frames of the same target, a double registration algorithm is proposed for point cloud cluster centroids. The iterative closest point (ICP) algorithm is employed for approximate interframe pose estimation during coarse registration. The random sample consensus (RANSAC) and four-parameter transformation algorithms are employed to obtain precise interframe pose relations during fine registration. These processes standardize the coordinate systems of adjacent point clouds and facilitate the association of point cloud clusters from the same target. Based on the paired point cloud cluster, a classification feature system is used to construct the XGBoost decision tree. To enhance the XGBoost training efficiency, a Spearman’s rank correlation coefficient-bidirectional search for a dimensionality reduction algorithm is proposed to expedite the optimal classification feature subset construction. After preliminary outcomes are generated by XGBoost, a double Boyer–Moore voting-sliding window algorithm is proposed to refine the final LiDAR dynamic target detection accuracy. To validate the efficacy and efficiency of our method in LiDAR dynamic target detection, an experimental platform is established. Real-world data are collected and pertinent experiments are designed. The experimental results illustrate the soundness of our method. The LiDAR dynamic target correct detection rate is 92.41%, the static target error detection rate is 1.43%, and the detection efficiency is 0.0299 s. Our method exhibits notable advantages over open-source comparative methods, achieving highly efficient and precise LiDAR dynamic target detection.

show abstract

Section: Methodsmentioning

confidence: 99%

LiDAR Dynamic Target Detection Based on Multidimensional Features

Xu,

Gao,

Sui

et al. 2024

Sensors

View full text Add to dashboard Cite

show abstract

“…Stable keypoints have the advantages of deformation invariant and no sensitive with lighting conditions, etc. Keypoints detection techniques have a wide range of practical applications such as face recognition [1], image matching [2,3], visual tracking [4], 3D registration [5], 3D reconstruction and mapping [6][7][8][9]. Therefore, reliable keypoints increase the likelihood of success in subsequent tasks.…”

Section: Introductionmentioning

confidence: 99%

GCNKD-Net:Multi-layer graph convolution networks for point cloud keypoints detection

Jia,

Li,

Zhang

et al. 2023

Preprint

View full text Add to dashboard Cite

Point cloud keypoints detection is an important component for many 3D applications. However, it is still a challenging task due to the irregular topology structure, noisy, low-resolution, and incomplete nature of 3D scan data. In this paper, we propose a novel learning based method to jointly learn a keypoint detector and a 3D feature descriptor, titled GCNKD-Net, based on graph neu-ral networks. Our approach consists of two parts: a keypoint detector to extract reliable keypoints and saliency uncertainties, and a feature descriptor to extract features for each keypoint. We propose a novel graph convolutional network in the two modules by using edge convolution with clustering strategy to extract features of points, which incorporates local neighborhood information and is stacked to learn global topology structures. Moreover, a joint loss function composed of detector loss and descriptor loss is defined for training the proposed GCNKD-Net model to encourage high repeatability and accurate localization of the keypoints. We evaluate our approach on two large LIDAR datasets-the KITTI dataset and the Oxford dataset. Experimental results demonstrate that our method achieves state-of-the-art performances with better repeatability values and matching accuracy of detecting keypoints. Our code is available at https://github.com/Lucky5026/GCNKDNet.git

show abstract

A novel 3D reconstruction method of blast furnace burden surface based on virtual camera array

Huang,

Jiang,

et al. 2025

Measurement

View full text Add to dashboard Cite

Improvement of 3D reconstruction based on a new 3D point cloud filtering algorithm

Cited by 4 publications

References 34 publications

LiDAR Dynamic Target Detection Based on Multidimensional Features

LiDAR Dynamic Target Detection Based on Multidimensional Features

GCNKD-Net:Multi-layer graph convolution networks for point cloud keypoints detection

A novel 3D reconstruction method of blast furnace burden surface based on virtual camera array

Contact Info

Product

Resources

About