In order to solve the problem of missing target detection in traffic road target detection task due to excessive change of target scale in complex environment. An improved object detection method based on YOLOv5 is proposed, which introduces Coordinate Attention (CA), it fuses the coordinate attention mechanism into the last layer of the backbone network C3 module and the Bottleneck module in the C3 module to enhance the feature extraction ability of the target. The experimental results show that the vehicle detection accuracy of the fusion model reaches 96.5% on KITTI data set. Compared with other models, it can effectively improve the problem that the target scale of traffic road changes greatly.
Multi-scale object detection based on attention mechanism
International Conference on Smart Transportation and City Engineering (STCE 2023) ; 2023 ; Chongqing, China
Proc. SPIE ; 13018
14.02.2024
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Traffic prediction method based on multi-scale space-time attention mechanism
Europäisches Patentamt | 2025
|