Road Vehicle Detection Using Improved YOLOv9 Based on Roadside Cameras

In order to improve the performance of road vehicle recognition based on roadside vision, this paper proposes an improved You Only Look Once v9-c (YOLOv9-c) target detection algorithm to realize road vehicles recognition under adverse weather conditions. Firstly, the YOLOv9-c detection network is used for vehicle detection, and the Convolutional Block Attention Module (CBAM) attention mechanism is introduced into its network to better acquire the target features. Furthermore, to address the problems of poor image quality and low accuracy of vehicle detection under adverse weather, the image quality is improved by combining with the All-in-One Dehazing Network (AOD-Net) de-fogging algorithm to optimize the performance of the target detection algorithm. The experimental results show that under dense fog, the mean average precision of the proposed method is improved by 2.3% and the recall rate is improved by 4.8% compared with the original algorithm.