A Lightweight Multi-Scale Feature Fusion Signal Detection Model for Metro Computer Interlocking Systems
Source abstract
Abstract
Metro Computer Interlocking Systems are crucial for ensuring the safe and efficient operation of rail transit. However, existing vision-based signal detection methods face challenges including small target sizes, high target density, low image resolution, and the need for deployment on resource-constrained devices. To address these issues, this paper proposes a two-stage lightweight signal detection framework for Metro Computer Interlocking Systems. First, based on YOLOv8, a small object detection layer together with feature fusion modules is introduced to form the YOLOv8-SFF architecture. A Scale Sequence Feature Fusion (SSFF) module is added to adjust the resolution of feature maps and retain critical fine-grained information, enhancing the detection of small visual signals. A Triple Feature Encoding (TFE) module is designed to enhance the recognition of dense small signals while replacing some traditional feature concatenation and upsampling operations, yielding a more compact network. Second, to enable practical deployment, a joint optimization strategy combining Layer-Adaptive Magnitude-based Pruning (LAMP) and Channel-wise Knowledge Distillation (CWD) is applied, in which the unpruned YOLOv8-SFF serves as the teacher and the pruned model serves as the student. In addition, an automatic annotation subsystem based on digital image processing is developed, leveraging color and morphological features to generate high-quality labels. Experimental results show that YOLOv8-SFF achieves a mean average precision (mAP@50) of 98.7%, improving mAP@50 by 11.3 percentage points and recall by 22.1 percentage points over the original YOLOv8. After joint pruning and distillation, the final compact model retains 98.0% mAP while reducing the parameter count by 86.2% and the model size to 1.1 MB, making it well suited for real-time deployment in resource-constrained metro dispatching systems.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Advanced Neural Network Applications · Infrastructure Maintenance and Monitoring · Advanced Data and IoT Technologies
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
The College of Information, Mechanical and Electrical Engineering, Shanghai Normal University, Shanghai 201418, China
Nanjing Kangni Mechanical & Electrical Co., Ltd., Nanjing 210008, China
Nanjing Ruijie Intelligent Transportation Technology Research Institute Co., Ltd., Nanjing 210004, China
School of Automation, Nanjing Institute of Technology, Nanjing 211167, China
References (34)
Show source-provided references
1. Huang, L. (2020, January 11–13). The past, present and future of railway interlocking system. Proceedings of the 2020 IEEE 5th International Conference on Intelligent Transportation Engineering (ICITE), Beijing, China.10.1109/ICITE50838.2020.9231438 ↗
2. Image robust recognition based on feature-entropy-oriented differential fusion capsule network10.1007/s10489-020-01873-3 ↗
3. Reliability and safety analysis on railway signal regional computer interlocking system
4. Shopa, P., Sumitha, N., and Patra, P.S.K. (2014, January 27–28). Traffic sign detection and recognition using OpenCV. Proceedings of the International Conference on Information Communication and Embedded Systems (ICICES2014), Chennai, India.10.1109/ICICES.2014.7033810 ↗
5. Traffic sign detection and recognition model using support vector machine and histogram of oriented gradient
6. Girshick, R., Donahue, J., Darrell, T., and Malik, J. (2014, January 23–28). Rich feature hierarchies for accurate object detection and semantic segmentation. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Columbus, OH, USA.10.1109/CVPR.2014.81 ↗
7. Redmon, J., Divvala, S., Girshick, R., and Farhadi, A. (2016, January 27–30). You Only Look Once: Unified, Real-Time Object Detection. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA.10.1109/CVPR.2016.91 ↗
8. Liu, W., Anguelov, D., Erhan, D., Szegedy, C., Reed, S., Fu, C.Y., and Berg, A.C. (2016, January 11–14). SSD: Single Shot MultiBox Detector. Proceedings of the European Conference on Computer Vision (ECCV), Amsterdam, The Netherlands.10.1007/978-3-319-46448-0_2 ↗
9. A topic-based multi-channel attention model under hybrid mode for image caption10.1007/s00521-021-06557-8 ↗
10. Redmon, J., and Farhadi, A. (2018). YOLOv3: An Incremental Improvement. arXiv.
11. Jocher, G. (2024, September 01). YOLO by Ultralytics (Version 5.7.0). GitHub. Available online: https://github.com/ultralytics/yolov5.
12. Bochkovskiy, A., Wang, C.Y., and Liao, H.Y.M. (2020). YOLOv4: Optimal speed and accuracy of object detection. arXiv.
13. Jocher, G., Chaurasia, A., and Qiu, J. (2024, September 01). YOLO by Ultralytics (Version 8.0.0). GitHub. Available online: https://github.com/ultralytics/ultralytics.
14. Lee, J., Park, S., Mo, S., Ahn, S., and Shin, J. (2021, January 3–7). Layer-adaptive sparsity for the magnitude-based pruning. Proceedings of the International Conference on Learning Representations (ICLR), Virtual.
15. Shu, C., Liu, Y., Gao, J., Yan, Z., and Shen, C. (2021, January 10–17). Channel-wise knowledge distillation for dense prediction. Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), Montreal, QC, Canada.10.1109/ICCV48922.2021.00526 ↗
16. ASF-YOLO: A novel YOLO model with attentional scale sequence fusion for cell instance segmentation10.1016/j.imavis.2024.105057 ↗
17. Optimal generation of test sequences for EMU and ATP on-board equipment interface test based on deep learning and genetic algorithm
18. Yan, X. (2019). Research on Simulation Test Method of Computer Interlocking Software. [Master’s Thesis, Beijing Jiaotong University].
19. Real-time small traffic sign detection with revised Faster-RCNN10.1007/s11042-018-6428-0 ↗
20. Choodowicz, E., Lisiecki, P., and Lech, P. (2020). Hybrid algorithm for the detection and recognition of railway signs. Progress in Computer Recognition Systems 11, Springer.10.1007/978-3-030-19738-4_34 ↗
21. YOLOv4 for advanced traffic sign recognition with synthetic training data generated by various GAN10.1109/ACCESS.2021.3094201 ↗
22. Zhang, Y., Guo, Z., Wu, J., Tian, Y., Tang, H., and Guo, X. (2022). Real-time vehicle detection based on improved YOLOv5. Sustainability, 14.10.3390/su141912274 ↗
23. Research on subway pedestrian detection algorithms based on SSD model10.1049/iet-its.2019.0806 ↗
24. Lin, M., Li, C., Bu, X., Sun, M., Lin, C., Yan, J., Ouyang, W., and Deng, Z. (2020). DETR for crowd pedestrian detection. arXiv.
25. Han, S., Pool, J., Tran, J., and Dally, W.J. (2015, January 7–12). Learning both weights and connections for efficient neural networks. Proceedings of the Advances in Neural Information Processing Systems (NeurIPS), Montreal, QC, Canada.
26. Liu, Z., Li, J., Shen, Z., Huang, G., Yan, S., and Zhang, C. (2017, January 22–29). Learning efficient convolutional networks through network slimming. Proceedings of the IEEE International Conference on Computer Vision (ICCV), Venice, Italy.10.1109/ICCV.2017.298 ↗
27. Li, H., Kadav, A., Durdanovic, I., Samet, H., and Graf, H.P. (2017, January 24–26). Pruning filters for efficient ConvNets. Proceedings of the International Conference on Learning Representations (ICLR), Toulon, France.
28. Hinton, G., Vinyals, O., and Dean, J. (2015). Distilling the knowledge in a neural network. arXiv.
29. Chen, X., and Li, Q. (2024, January 24–26). Image recognition and analysis of computer interlock interface based on OpenCV technology. Proceedings of the 2024 IEEE 6th Advanced Information Management, Communications, Electronic and Automation Control Conference (IMCEC), Chongqing, China.10.1109/IMCEC59810.2024.10575357 ↗
30. Zhang, H., and Kuang, W. (2024, January 24–26). Computer interlocking host computer recognition based on OpenCV and neural network. Proceedings of the 2024 IEEE 6th Advanced Information Management, Communications, Electronic and Automation Control Conference (IMCEC), Chongqing, China.10.1109/IMCEC59810.2024.10574961 ↗
31. Cheng, H., He, T., and Tian, R. (2023, January 5–7). Research on the application of YOLOv5 in station interlocking test. Proceedings of the Third International Conference on Image Processing and Intelligent Control (IPIC 2023), Kuala Lumpur, Malaysia.10.1117/12.3003775 ↗
32. Rezatofighi, H., Tsoi, N., Gwak, J., Sadeghian, A., Reid, I., and Savarese, S. (2019, January 15–20). Generalized intersection over union: A metric and a loss for bounding box regression. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Long Beach, CA, USA.10.1109/CVPR.2019.00075 ↗
33. Distance-IoU loss: Faster and better learning for bounding box regression
34. Focal and efficient IOU loss for accurate bounding box regression10.1016/j.neucom.2022.07.042 ↗