Two-Layer-Graph Clustering for Real-Time 3D LiDAR Point Cloud Segmentation

Yang, Huayan; Wang, Zhong Lin; Lin, Linglong; Liang, Huawei; Huang, Weixin; Xu, Fengyu

doi:10.3390/app10238534

Cited by 15 publications

(10 citation statements)

References 40 publications

(60 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The two vehicles are tested in a regular urban environment, and the vehicle's trajectory is designed to be on a right turn road section. This paper uses ground segmentation with the original point cloud [15] and then performs clustering processing [16] in the process of point cloud data. Moreover, we use the object tracking algorithm [41] to evaluate the pose estimation algorithm's effect on tracking results.…”

Section: Experimental Results Of Our Experimental Platformmentioning

confidence: 99%

“…In contrast, the adaptability of traditional methods is much better. In the traditional object detection pipeline of the 3D point cloud, it is generally necessary to first perform ground segmentation with the original point cloud [15] and then perform clustering processing with the point cloud data [16]. The object detection task is finally completed after estimating the pose of each obstacle according to the clustering results [17].…”

Section: Of 30mentioning

confidence: 99%

See 1 more Smart Citation

Optimal Vehicle Pose Estimation Network Based on Time Series and Spatial Tightness with 3D LiDARs

et al. 2021

Self Cite

View full text Add to dashboard Cite

Vehicle pose estimation is essential in autonomous vehicle (AV) perception technology. However, due to the different density distributions of the point cloud, it is challenging to achieve sensitive direction extraction based on 3D LiDAR by using the existing pose estimation methods. In this paper, an optimal vehicle pose estimation network based on time series and spatial tightness (TS-OVPE) is proposed. This network uses five pose estimation algorithms proposed as candidate solutions to select each obstacle vehicle's optimal pose estimation result. Among these pose estimation algorithms, we first propose the Basic Line algorithm, which uses the road direction as the prior knowledge. Secondly, we propose improving principal component analysis based on point cloud distribution to conduct rotating principal component analysis (RPCA) and diagonal principal component analysis (DPCA) algorithms. Finally, we propose two global algorithms independent of the prior direction. We provided four evaluation indexes to transform each algorithm into a unified dimension. These evaluation indexes’ results were input into the ensemble learning network to obtain the optimal pose estimation results from the five proposed algorithms. The spatial dimension evaluation indexes reflected the tightness of the bounding box and the time dimension evaluation index reflected the coherence of the direction estimation. Since the network was indirectly trained through the evaluation index, it could be directly used on untrained LiDAR and showed a good pose estimation performance. Our approach was verified on the SemanticKITTI dataset and our urban environment dataset. Compared with the two mainstream algorithms, the polygon intersection over union (P-IoU) average increased by about 5.25% and 9.67%, the average heading error decreased by about 29.49% and 44.11%, and the average speed direction error decreased by about 3.85% and 46.70%. The experiment results showed that the ensemble learning network could effectively select the optimal pose estimation from the five abovementioned algorithms, making pose estimation more accurate.

show abstract

Section: Experimental Results Of Our Experimental Platformmentioning

confidence: 99%

Section: Of 30mentioning

confidence: 99%

Optimal Vehicle Pose Estimation Network Based on Time Series and Spatial Tightness with 3D LiDARs

et al. 2021

Self Cite

View full text Add to dashboard Cite

show abstract

“…Burger et al [14] proposed a mesh structure using loss functions that consist of cluster densities, slopes, distances, and angles; however, its three steps of horizontal, vertical, and fusion updates entail high computation complexity. Yang et al [15] represented a set of horizontally neighboring points as a node and created a set graph. However, Yang et al [15] did not consider the interaction between indexwise non-adjoining nodes, even when, due to occlusion, they can be possible candidates to be clustered geometrically.…”

Section: B Object Segmentationmentioning

confidence: 99%

“…Yang et al [15] represented a set of horizontally neighboring points as a node and created a set graph. However, Yang et al [15] did not consider the interaction between indexwise non-adjoining nodes, even when, due to occlusion, they can be possible candidates to be clustered geometrically. Establishing deficient interplay between scan points, the above segmentation methods are prone to over-segmentation, especially caused by occlusion, and under-segmentation, especially caused by ground segmentation failure.…”

Section: B Object Segmentationmentioning

confidence: 99%

TRAVEL: Traversable Ground and Above-Ground Object Segmentation Using Graph Representation of 3D LiDAR Scans

Jung

Lim

et al. 2022

IEEE Robot. Autom. Lett.

View full text Add to dashboard Cite

Perception of traversable regions and objects of interest from a 3D point cloud is one of the critical tasks in autonomous navigation. A ground vehicle needs to look for traversable terrains that are explorable by wheels. Then, to make safe navigation decisions, the segmentation of objects positioned on those terrains has to be followed up. However, over-segmentation and under-segmentation can negatively influence such navigation decisions. To that end, we propose TRAVEL, which performs traversable ground detection and object clustering simultaneously using the graph representation of a 3D point cloud. To segment the traversable ground, a point cloud is encoded into a graph structure, tri-grid field, which treats each tri-grid as a node. Then, the traversable regions are searched and redefined by examining local convexity and concavity of edges that connect nodes. On the other hand, our above-ground object segmentation employs a graph structure by representing a group of horizontally neighboring 3D points in a spherical-projection space as a node and vertical/horizontal relationship between nodes as an edge. Fully leveraging the node-edge structure, the above-ground segmentation ensures real-time operation and mitigates over-segmentation. Through experiments using simulations, urban scenes, and our own datasets, we have demonstrated that our proposed traversable ground segmentation algorithm outperforms other state-ofthe-art methods in terms of the conventional metrics and that our newly proposed evaluation metrics are meaningful for assessing the above-ground segmentation. We will make the code and our own dataset available to public at https://github.com/url-kaist/TRAVEL.

show abstract

“…Sampath and Shan first calculated the probability that each point in the point cloud can be used as the clustering center and then used fuzzy k-means to realize the point cloud segmentation. Yang et al designed an improved breadth first search algorithm to update the clustering of point clouds so as to realize the segmentation of point clouds.…”

Section: Introductionmentioning

confidence: 99%

Coal Wall and Roof Segmentation in the Coal Mine Working Face Based on Dynamic Graph Convolution Neural Networks

Zhao

Guo

et al. 2021

ACS Omega

View full text Add to dashboard Cite

The intersection line information of the point cloud between the coal wall and the roof can not only accurately reflect the direction information of the scraper conveyor but also provide a preliminary basis for realizing the intelligent coal mine. However, the indirect method of using deep learning to segment the point cloud of coal mine working face cannot make full use of the rich information provided by the point cloud data. The direct method of using deep learning to segment the point cloud ignores the local feature relationship between points. Therefore, we propose to use dynamic graph convolution neural networks (DGCNNs) to segment the point cloud of the coal wall and roof so as to obtain the intersection line between them. First, in view of the characteristics of heavy dust and strong electromagnetic interference in the environment of the coal mine working face, we have installed an underground inspection robot so that we use light detection and ranging to obtain the point cloud of the coal mine working face. At the same time, we put forward a fast labeling method of the point cloud of the coal mine working face and an efficient training method of the depth neural network. Second, on the basis of edge convolution, being the greatest innovation of DGCNNs, we analyze the influence of the number of layers, K value, and output feature dimension of edge convolution on the effect of DGCNNs segmenting the point cloud of the coal mine working face and obtaining the intersection line of the coal wall and roof. Finally, we compare DGCNNs with PointNet and PointNet++. The results show that the DGCNN exhibits the best performance. What is more, the results provide a research foundation for the application of DGCNNs in the field of energy. Last but not least, the research results provide a direct and key basis for the adjustment of the scraper conveyor, which is of great significance for an intelligent coal mine working face and accurate construction of a geological information model.

show abstract

Two-Layer-Graph Clustering for Real-Time 3D LiDAR Point Cloud Segmentation

Cited by 15 publications

References 40 publications

Optimal Vehicle Pose Estimation Network Based on Time Series and Spatial Tightness with 3D LiDARs

Optimal Vehicle Pose Estimation Network Based on Time Series and Spatial Tightness with 3D LiDARs

TRAVEL: Traversable Ground and Above-Ground Object Segmentation Using Graph Representation of 3D LiDAR Scans

Coal Wall and Roof Segmentation in the Coal Mine Working Face Based on Dynamic Graph Convolution Neural Networks

Contact Info

Product

Resources

About