To address the generalization issues of train forward obstacle detection methods based on video imagery under factors such as lighting conditions and target distances, a LiDAR-based detection method is proposed. Firstly, focusing on the accuracy loss in the VoxelNeXt voxelization process, dynamic voxelization technology is incorporated to optimize this process, thereby minimizing information loss. Secondly, considering the spatial distribution characteristics of the forward operating environment of the train, an L-shaped residual sparse convolution module is designed, so as to effectively capture the deep semantic features of point cloud data in the forward operating environment of the train. Finally, a cross dimensional automatic encoding module is proposed, which integrates with the backbone feature extraction network to form a cross-dimensional automatic encoding network, further enhancing the expression capability of the network's output features. The results show that the average accuracy of the proposed method can reach 72.38%, and the average recall rate can reach 76.59%, demonstrating significant performance advantages compared to other methods. This method fulfills the requirements for high-precision, long-distance, and fast detection of obstacles in the forward direction of trains, providing effective technical support for active train safety assurance.
QINY, CAOZ W, SUNY F, et al. Research on Active Safety Methodologies for Intelligent Railway Systems [J]. Engineering, 2023, 27: 266-279.
[2]
CAOZ W, QINY, JIAL M, et al. Railway Intrusion Detection Based on Machine Vision: a Survey, Challenges, and Perspectives [J]. IEEE Transactions on Intelligent Transportation Systems, 2024, 25 (7): 6427-6448.
[3]
YET, RENC, ZHANGX, et al. Application of Lightweight Railway Transit Object Detector [J]. IEEE Transactions on Industrial Electronics, 2021, 68 (10): 10269-10280.
JIAOYueli, SUChao, HUANGShenyue. Study on the Active Obstacle Detecting System for Rail Vehicles [J]. Intelligent Rail Transit, 2023, 6 (12): 12-15. in Chinese
[6]
CAOZ W, QINY, JIAL M, et al. Haze Removal of Railway Monitoring Images Using Multi-Scale Residual Network [J]. IEEE Transactions on Intelligent Transportation Systems, 2021, 22 (12): 7460-7473.
HUHao, SHITianyun, YANGWen. Railway Perimeter Intrusion Detection Method Based on Improved FairMOT [J]. China Railway Science, 2023, 44 (5): 222- 232. in Chinese
CHANGYaohui, CHENNiansheng, RAOLei, et al. Lidar Point Cloud Descriptor with Rotation and Translation Invariance in Dynamic Environment [J]. Acta Optica Sinica, 2022, 42 (24): 2401007. in Chinese
SHENTuo, QIANYanzuo, XIELanxin, et al. Obstacle Detection Algorithm of Fully Automatic Train Considering Reflection Intensity [J]. Journal of Tongji University (Natural Science), 2022, 50 (1): 6-12. in Chinese
[13]
WANGZ Y, YUG Z, WUX K, et al. A Camera and LiDAR Data Fusion Method for Railway Object Detection [J]. IEEE Sensors Journal, 2021, 21 (12): 13442-13454.
[14]
GUOY L, WANGH Y, HUQ Y, et al. Deep Learning for 3D Point Clouds: a Survey [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021, 43 (12): 4338-4364.
[15]
XIAOA R, HUANGJ X, GUAND Y, et al. Unsupervised Point Cloud Representation Learning with Deep Neural Networks: a Survey [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023, 45 (9): 11321-11339.
[16]
HAZERA, YILDIRIMR. Deep Learning Based Point Cloud Processing Techniques [J]. IEEE Access, 2022, 10: 127237-127283.
[17]
LANGA H, VORAS, CAESARH, et al. PointPillars: Fast Encoders for Object Detection from Point Clouds [C]// 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Long Beach, CA, USA. New York: IEEE Press, 2019: 12689-12697.
[18]
YINT W, ZHOUX Y, KRÄHENBÜHLP. Center-Based 3D Object Detection and Tracking [C]// 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Nashville, TN, USA. New York: IEEE Press, 2021: 11779-11788.
[19]
CHENY K, LIUJ H, ZHANGX Y, et al. VoxelNeXt: Fully Sparse VoxelNet for 3D Object Detection and Tracking [C]// 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Vancouver, BC, Canada. New York: IEEE Press, 2023: 21674-21683.
[20]
ZHOUY, SUNP, ZHANGY, et al. End-to-End Multi-View Fusion for 3D Object Detection in LiDAR Point Clouds [EB/OL]. 2019: arXiv: 1910.06528.
ZHUY W, ZHENGC R, YUANC J, et al. CamVox: a Low-Cost and Accurate LiDAR-Assisted Visual SLAM System [EB/OL]. 2020: arXiv: 2011.11357.
[23]
SHIG S, LIR F, MAC. PillarNet: Real-Time and High-Performance Pillar-Based 3D Object Detection [M]// Computer Vision – ECCV 2022. Lecture Notes in Computer Science, vol 13670. Cham: Springer, 2022.
[24]
SHIS S, WANGZ, SHIJ P, et al. From Points to Parts: 3D Object Detection from Point Cloud With Part-Aware and Part-Aggregation Network [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021, 43 (8): 2647-2664.