A Vehicle Detection and Alert System Based on YOLOv4: Design, Implementation, and Performance Evaluation
DOI:
https://doi.org/10.53469/wjimt.2026.09(02).02Keywords:
Target detection, Yolov4 algorithm, Pedestrian safety, Deep learningAbstract
As living standards continue to rise, transportation systems have become increasingly convenient; however, associated safety risks have also escalated significantly. Insufficient pedestrian safety awareness has contributed to a rising incidence of traffic accidents. In response, this study leverages the relative maturity of object detection technologies by retraining the YOLOv4 algorithm using the VOC dataset to design and implement a vehicle alert system tailored for unsignalized intersections. The proposed system aims to mitigate collision risks under low-light and poor nighttime visibility conditions by detecting approaching vehicles and providing timely alerts to support safer pedestrian crossing decisions. Evaluation of the model’s performance demonstrates its effectiveness and practical applicability in real-world traffic environments.
References
Fang Su, Qin Yuan. Analysis of Association Rules of Urban Highway Traffic Accidents [J]. Journal of Wuhan University of Technology （Information and Management Engineering Edition）, 2020, 42 （04）: 313-318 + 331.
GIRSHICK R, DONAHUE J, DARRELLT, et al.Rich feature hierarchies for accurate object detection and semantic segmentation [J]. Computer Science, 2013: 580- 587.
Redmon J, Divvala S,Girshick R,et al.You only look once:Unified,real-time object detection [C] // Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.New York:IEEE,2016:779-788.
Neubeck A, Gool L. Efficient Non- Maximum Suppression[C]//International Conference on Pattern Recognition. IEEE Computer Society, 2006.
Alexey Bochkovskiy，Chien-Yao Wang， Hong-Yuan MarkLiao. YOLOv4：Optimal Speed and Accuracy of ObjectDetection［C］//CVPR，Apr， 2020.
Wang H B,Zhang Z D.Text detection algorithm based on improved YOLOv3[C]. Beijing:IEEE the Ninth International Conference on Electronics Information and Emergency Communication,2019.
ZHENG Z,WANG P, LIU W,et al. Distance-IoU loss:faster and better learning for bounding box regression [C] ∥ AAAI,2020:12993-13000.]
Shu Liu, Lu Qi, Haifang Qin, etal. Path aggregation network for instance segmentation [C] // Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition,2018:8759-8768.
Wei Gao, Chen Zhou, Moufa Guo. Research on insulator defect identification based on improved YOLOv4 and SR-GAN [J / OL].1-11 (2021-05-19 08: 34).
RAH MAN MA, WANG Y.Optimizing intersectionover-union in deep neural networks for image segmentation [ J]. International Symposium on Visual Computing 2016．
IAN Goodfellow, YOSHUA Bengio,AARON Courville. Deep learning [M]. Beijing: People's Posts and Telecommunications Publishing House, 2017.