Port Container Detection in Foggy Weather Scenarios Based on YOLOv5

Container detection in foggy weather scenarios is a significant application for logistics, but fog can seriously affect image quality and detection accuracy. This article proposes a deep learning based object detection algorithm, You Only Look Once Version 5 (YOLOv5), for container detection in foggy scenes. This method combines the improved dark channel defogging algorithm (IDCDA) and the Squeeze-and-excitation networks (SE) attention mechanism to improve the defogging effect and detection accuracy. The IDCDA algorithm effectively eliminates the serious impact of fog and image color distortion after defogging by introducing adaptive thresholds and local contrast enhancement. The SE attention mechanism enhances the feature extraction ability and the receptive field range. This article conducted experiments on a self-built foggy container dataset, and the results showed that the proposed method outperformed existing methods in terms of defogging effect and detection accuracy, proving the effectiveness and feasibility of the proposed method.