Related Experiment Videos
Research on Tracking and Detecting Algorithm for Road Signs Based on SCMCg
Feng Wang1, Ruining Jiang1, Zhirui Tang1
1The School of Automobile and Traffic Engineering, Wuhan University of Science and Technology, Wuhan 430081, China.
Sensors (Basel, Switzerland)
|August 13, 2026
Summary
This study introduces YOLO-DeepSort with Spatial Multivariate Clustering Algorithm (SCMCg) for accurate road sign detection, significantly improving precision and reducing spatial errors in highway maintenance.
Area of Science:
- Computer Vision
- Artificial Intelligence
- Intelligent Transportation Systems
Background:
- Road sign detection is vital for highway maintenance.
- Current methods struggle with sign loss, occlusion, and spatial misjudgments.
- Existing systems lack robustness in complex road environments.
Purpose of the Study:
- To develop an advanced road sign detection and tracking framework.
- To enhance accuracy and spatial correctness in automated highway maintenance.
- To mitigate challenges like occlusion and mapping errors.
Main Methods:
- Proposed YOLO-DeepSort framework integrating Spatial Multivariate Clustering Algorithm with GPS (SCMCg).
- Enhanced YOLOv9 detector with Mixed Local Channel Attention (MLCA) and DualConv modules.
- DeepSort tracking refined with SCMCg, Delaunay triangulation, and hierarchical GPS constraints.
Main Results:
- Achieved 97.8% precision, 91.2% mAP, and 97.3% tracking success rate with 21.4M parameters.
- Demonstrated significant improvements over baseline YOLOv8-DeepSort.
- Effectively reduced occlusion and spatial tracking errors.
Conclusions:
- The proposed YOLO-DeepSort with SCMCg offers a robust and lightweight solution for automated road sign condition assessment.
- The framework enhances the reliability of intelligent transportation infrastructure monitoring.
- This methodology addresses key limitations in current road sign detection systems.
