In road traffic environments, the detection of vehicles and pedestrians is crucial for autonomous driving and intelligent surveillance. However, complex road conditions often introduce significant amounts of noise, and the varied scales and shapes of vehicles and pedestrians, coupled with imbalanced training samples, further challenge the performance of detection models. This paper addresses these issues by enhancing YOLOv8 and proposing a novel vehicle-pedestrian detection algorithm named Road-YOLO. To accommodate the diverse scales and shapes of vehicles and pedestrians, we introduce deformable convolutions in the backbone network. Additionally, we incorporate Convolutional Block Attention Module (CBAM) in the neck of the model to precisely capture critical features in images and suppress irrelevant information from road environments. Last, we adopt the Wise-IoU (WIoU) loss function, focusing on anchor boxes of adequate quality during model training and learning, thereby enhancing the model's generalization capability. We conducted experiments on the M3FD dataset, and the results showed that the average detection accuracy (mAP) of Road-YOLO increased from 90.72% to 93.74%, which proved the effectiveness of Road-YOLO in improving the effectiveness of vehicle-pedestrian detection in complex road traffic environments.
Road-YOLO: Efficient and Robust Detection of Vehicles and Pedestrians on Roads
Lect. Notes Electrical Eng.
International Conference on Computer Engineering and Networks ; 2024 ; Kashi, China October 18, 2024 - October 21, 2024
Proceedings of the 14th International Conference on Computer Engineering and Networks ; Kapitel : 43 ; 470-481
07.05.2025
12 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Detecting Pedestrians and Vehicles on Urban Roads using Computer Vision
Elsevier | 2025
|Detection of Quiet Vehicles by Blind Pedestrians
Online Contents | 2013
|Detection of Quiet Vehicles by Blind Pedestrians
ASCE | 2012
|Detection of Quiet Vehicles by Blind Pedestrians
British Library Online Contents | 2013
|What about Pedestrians and Cyclists on Rural Roads
NTIS | 1986
|