Scholarlink

This study proposes the Intelligent over Emergency (IE) - you only look once (YOLO) object detection method based on YOLOv7 to address the low detection accuracy, often missed detections, and false detections of traffic signs in complex road scenes using existing object detection methods. Firstly, the ML-SPPF module was designed to enhance and fuse multi-scale features. Then, a loss function was proposed to accelerate the convergence speed of the model and improve the detection accuracy. Finally, the YOLOv7 network model was improved by proposing a cross-level fusion mechanism, and the detection head was redesigned to make object localization more accurate. Meanwhile, in response to the current lack of traffic sign datasets under extreme conditions, a traffic sign detection dataset under extreme conditions, traffic sign dataset benchmarks (TSDB), is made, which can meet the requirements of common traffic sign detection tasks. The experimental results indicate that the IE-YOLO method performs well, and the mAP@0.5 evaluation index is 93.8%, which is 5.2% higher than the baseline method. At the same time, IE-YOLO’s reasoning speed meets the real-time requirements of the auto drive system, and its comprehensive performance is better than the SOTA object detection method, which can more efficiently complete the task of traffic object detection in complex driving scenarios.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Efficient Real-Time Traffic Sign Detection for Autonomous Driving in Adverse Weather Using Deep Learning Models

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Yang Liu,
  • Hongyu Sun,
  • Weiqin Li

摘要

This study proposes the Intelligent over Emergency (IE) - you only look once (YOLO) object detection method based on YOLOv7 to address the low detection accuracy, often missed detections, and false detections of traffic signs in complex road scenes using existing object detection methods. Firstly, the ML-SPPF module was designed to enhance and fuse multi-scale features. Then, a loss function was proposed to accelerate the convergence speed of the model and improve the detection accuracy. Finally, the YOLOv7 network model was improved by proposing a cross-level fusion mechanism, and the detection head was redesigned to make object localization more accurate. Meanwhile, in response to the current lack of traffic sign datasets under extreme conditions, a traffic sign detection dataset under extreme conditions, traffic sign dataset benchmarks (TSDB), is made, which can meet the requirements of common traffic sign detection tasks. The experimental results indicate that the IE-YOLO method performs well, and the mAP@0.5 evaluation index is 93.8%, which is 5.2% higher than the baseline method. At the same time, IE-YOLO’s reasoning speed meets the real-time requirements of the auto drive system, and its comprehensive performance is better than the SOTA object detection method, which can more efficiently complete the task of traffic object detection in complex driving scenarios.