This study focuses on traffic accident detection technology based on YOLOv11, aiming to improve road safety levels using advanced computer vision and deep learning methods. By constructing a diverse dataset covering common types of traffic accidents and traffic targets, and using data augmentation techniques such as adding random noise and Gaussian blur, the robustness and generalization ability of the model are enhanced. YOLOv11 has undergone multiple innovations in its model architecture, including the introduction of C3k2 units, SPPF, and C2PSA modules, effectively improving the detection accuracy and processing speed of the model. The experimental results showed that YOLOv11 achieved industry-leading mAP50 and mAP50-95 values in traffic accident detection, demonstrating its superior performance in complex traffic scenarios. Future research directions can focus on multimodal data fusion, real-time performance optimization, and multi task learning to further improve the accuracy and practicality of traffic accident detection systems.
Traffic accident detection based on YOLOv11
2024-12-29
1371604 byte
Conference paper
Electronic Resource
English
Autonomous UAV Detection of Ochotona curzoniae Burrows with Enhanced YOLOv11
DOAJ | 2025
|A Lightweight Anti-Unmanned Aerial Vehicle Detection Method Based on Improved YOLOv11
DOAJ | 2024
|