Scholarlink

Pedestrian and cyclist safety is a crucial concern in densely populated urban areas, where traffic congestion and human vulnerability intersect. This research paper presents a computer vision and machine learning-based system for real-time pedestrian and cyclist detection, employing the YOLOv9 (You Only Look Once) model to enhance road safety. The World Health Organization reports over 270,000 pedestrian fatalities annually, highlighting the need for advanced detection systems in urban settings. By utilizing a dataset consisting of varied urban environments, including different lighting conditions, traffic densities, and pedestrian/cyclist behavior, the YOLOv9 model was trained and evaluated for performance. The model achieved high detection accuracy and demonstrated the ability to perform real-time predictions. Despite the model's strengths in well-lit and uncrowded environments, challenges remained in detecting obscured objects and handling complex urban scenarios. The study concludes with recommendations for improving detection robustness and the potential for integrating predictive models to further enhance urban road safety.

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

Enhancing Safety Through Advanced Pedestrian and Cyclist Detection and Movement Prediction Using Computer Vision and Machine Learning

  • 章节

  • 发表日期：

    2026年

查看全文
  • Shaurya Agarwal,
  • Reetu Jain

摘要

Pedestrian and cyclist safety is a crucial concern in densely populated urban areas, where traffic congestion and human vulnerability intersect. This research paper presents a computer vision and machine learning-based system for real-time pedestrian and cyclist detection, employing the YOLOv9 (You Only Look Once) model to enhance road safety. The World Health Organization reports over 270,000 pedestrian fatalities annually, highlighting the need for advanced detection systems in urban settings. By utilizing a dataset consisting of varied urban environments, including different lighting conditions, traffic densities, and pedestrian/cyclist behavior, the YOLOv9 model was trained and evaluated for performance. The model achieved high detection accuracy and demonstrated the ability to perform real-time predictions. Despite the model's strengths in well-lit and uncrowded environments, challenges remained in detecting obscured objects and handling complex urban scenarios. The study concludes with recommendations for improving detection robustness and the potential for integrating predictive models to further enhance urban road safety.