LiDAR Point Cloud Recognition of Overhead Catenary System with Deep Learning

Lin, Shuai; Xu, Cheng; Chen, Lipei; Li, Siqi; Tu, Xiaohan

doi:10.3390/s20082212

Cited by 27 publications

(48 citation statements)

References 27 publications

(28 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The local geometry was captured by KCNet [12] through the convolution layer of the front-end kernel, but there remained the problem of point-by-point loss. Optimization schemes for extracting local features have been successively improved by the latest research [37][38][39][40][41]. However, the extraction and use of local features still face great challenges.…”

Section: B the Methods Of Deep Learningmentioning

confidence: 99%

MLFNet- Point Cloud Semantic Segmentation Convolution Network Based on Multi-Scale Feature Fusion

et al. 2021

View full text Add to dashboard Cite

The project was funded by jinan science and technology bureau and undertaken by QiluUniversity of Technology(Shandong Academy of Sciences),Machine vision-based online intelligentsegmentation of foreign objects in liquids is implemented.2019GXRC067. ABSTRACTIn the semantic segmentation of a point cloud, if the spatial structure correlation between the input features and coordinates are not fully considered, a semantic segmentation error can occur. We propose a method of spatial convolution that makes full use of the characteristics of a multiscale spatial structure by combining local and global features. We call this method MLFNet. We also propose a multiscale feature framework. First, the point cloud is simplified by obtaining the weighted farthest point (by down-sampling combined with farthest-point sampling and the weighted average). The near-near domain of each sampling point is then obtained by a KK octant search (an octant search optimized by the knearest neighbor and a custom threshold), and feature information is obtained. The feature information is added to the subsequent multilayer perceptron, and fusion of local context information is achieved. Finally, the fusion features in multiple directions are maximally pooled. Our method was tested on self-made datasets and other standard basic datasets (ModelNet40, ShapeNet, and Stanford large-scale 3D indoor spaces (S3DIS) data). The accuracy of segmentation was 0.937 in our dataset; two percentage points higher than the latest deep learning method. Also, our method obtained a mean intersection over a union of 0.867 in ShapeNet, which was 0.3 percentage points higher than the latest PointGrid. The accuracy on S3DIS was 0.8153, which was three percentage points higher than the latest spatial aggregation net. The results of semantic segmentation verified the superiority of the proposed method.

show abstract

Section: B the Methods Of Deep Learningmentioning

confidence: 99%

MLFNet- Point Cloud Semantic Segmentation Convolution Network Based on Multi-Scale Feature Fusion

et al. 2021

View full text Add to dashboard Cite

show abstract

“…Although the conventional manual monitoring method shown in Figure 1 b is more flexible, there are many shortcomings in artificial detection such as low efficiency, high risk, and human factors such as personnel experience, which can affect detection results. Therefore, it has been unable to meet the existing detection needs [ 5 , 6 , 7 , 8 ]. At present, determining the detection method with high efficiency, high precision, and low risk has become a research hotspot.…”

Section: Introductionmentioning

confidence: 99%

“…Gutiérrez-Fernández et al [ 23 ] presented a new method for the automatic extraction of power cable locations in railways using surface LiDAR systems. Lin et al and [ 7 ] presented a method based on deep learning to recognize point clouds of OCS components. In general, the mean extraction accuracy of these methods can reach 92–98%.…”

Section: Introductionmentioning

confidence: 99%

Railway Overhead Contact System Point Cloud Classification

Xiao

Chen

Liu

et al. 2021

Sensors

View full text Add to dashboard Cite

As the railway overhead contact system (OCS) is the key component along the high-speed railway, it is crucial to detect the quality of the OCS. Compared with conventional manual OCS detection, the vehicle-mounted Light Detection and Ranging (LiDAR) technology has advantages such as high efficiency and precision, which can solve the problems of OCS detection difficulty, low efficiency, and high risk. Aiming at the contact cables, return current cables, and catenary cables in the railway vehicle-mounted LiDAR OCS point cloud, this paper used a scale adaptive feature classification algorithm and the DBSCAN (density-based spatial clustering of applications with noise) algorithm considering OCS characteristics to classify the OCS point cloud. Finally, the return current cables, catenary cables, and contact cables in the OCS were accurately classified and extracted. To verify the accuracy of the method presented in this paper, we compared the experimental results of this article with the classification results of TerraSolid, and the classification results were evaluated in terms of four accuracy indicators. According to statistics, the average accuracy of using this method to extract two sets of OCS point clouds is 99.83% and 99.89%, respectively; the average precision is 100% and 99.97%, respectively; the average recall is 99.16% and 99.42%, respectively; and the average overall accuracy is 99.58% and 99.69% respectively, which is overall better than TerraSolid. The experimental results showed that this approach could accurately and quickly extract the complete OCS from the point cloud. It provides a new method for processing railway OCS point clouds and has high engineering application value in railway component detection.

show abstract

“…Other methods [4,5,7] relied on images to inspect an OCS with DNNs, but high-resolution images allowed a larger amount of calculation than point clouds. To further improve OCS recognition, the approach in [14] leveraged the k-nearest neighbors algorithm (KNN) [15] and DNNs to process point clouds of an OCS in real high-speed scenarios. As the point density and distribution pattern influence point cloud segmentation, Lin et al [14] adopted data from a fixed scan area.…”

mentioning

confidence: 99%

“…To further improve OCS recognition, the approach in [14] leveraged the k-nearest neighbors algorithm (KNN) [15] and DNNs to process point clouds of an OCS in real high-speed scenarios. As the point density and distribution pattern influence point cloud segmentation, Lin et al [14] adopted data from a fixed scan area. They summed the previous and next frames and obtained the 3D coordinate values of the current frame.…”

mentioning

confidence: 99%

An Optimized Deep Neural Network for Overhead Contact System Recognition from LiDAR Point Clouds

Liu

et al. 2021

Remote Sensing

Self Cite

View full text Add to dashboard Cite

As vital infrastructures, high-speed railways support the development of transportation. To maintain the punctuality and safety of railway systems, researchers have employed manual and computer vision methods to monitor overhead contact systems (OCSs), but they have low efficiency. Investigators have also used light detection and ranging (LiDAR) to generate point clouds by emitting laser beams. The point cloud is segmented for automatic OCS recognition, which improves recognition efficiency. However, existing LiDAR point cloud segmentation methods have high computational/model complexity and latency. In addition, they cannot adapt to embedded devices with different architectures. To overcome these issues, this article presents a lightweight neural network EffNet consisting of three modules: ExtractA, AttenA, and AttenB. ExtractA extracts the features from the disordered and irregular point clouds of an OCS. AttenA keeps information flowing in EffNet while extracting useful features. AttenB uses channel and spatialwise statistics to enhance important features and suppress unimportant ones efficiently. To further speed up EffNet and match it with diverse architectures, we optimized it with a generation framework of tensor programs and deployed it on embedded systems with different architectures. Extensive experiments demonstrated that EffNet has at least a 0.57% higher mean accuracy, but with 25.00% and 9.30% lower computational and model complexity for OCS recognition than others, respectively. The optimized EffNet can be adapted to different architectures. Its latency decreased by 51.97%, 56.47%, 63.63%, 82.58%, 85.85%, and 91.97% on the NVIDIA Nano CPU, TX2 CPU, UP Board CPU, Nano GPU, TX2 GPU, and RTX 2,080 Ti GPU, respectively.

show abstract

LiDAR Point Cloud Recognition of Overhead Catenary System with Deep Learning

Cited by 27 publications

References 27 publications

MLFNet- Point Cloud Semantic Segmentation Convolution Network Based on Multi-Scale Feature Fusion

MLFNet- Point Cloud Semantic Segmentation Convolution Network Based on Multi-Scale Feature Fusion

Railway Overhead Contact System Point Cloud Classification

An Optimized Deep Neural Network for Overhead Contact System Recognition from LiDAR Point Clouds

Contact Info

Product

Resources

About