Scholarlink

To address various challenges, including delayed traffic accident detection response, poor localization accuracy, and inadequate rescue efficiency, this study presents a YOLOv10-based intelligent detection and classification system. Featuring a three-tier “perception-decision-action” framework, the system uses an enhanced YOLOv10 model for precise accident detection and implements multiscale feature fusion to improve the recognition of diverse accident scenarios. The classification module employs dynamic weighting to assess accidents as mild, moderate, or severe, coupled with a tri-level response protocol. For route optimization, the path planning component uses a refined A* algorithm with a four-dimensional cost function. Tests show that the system delivers 98.2% detection accuracy with subsecond response times, maintains 93.5% reliability in harsh weather, and enhances rescue operations by 23.7% in efficiency. This solution offers a robust technical approach for smart traffic management systems.

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

Design of a Road Traffic Classification Accident Detection and Identification System Based on YOLOv10

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Sifan Wang,
  • Yanqing Wang,
  • Shengli Wang,
  • Yufan Mei

摘要

To address various challenges, including delayed traffic accident detection response, poor localization accuracy, and inadequate rescue efficiency, this study presents a YOLOv10-based intelligent detection and classification system. Featuring a three-tier “perception-decision-action” framework, the system uses an enhanced YOLOv10 model for precise accident detection and implements multiscale feature fusion to improve the recognition of diverse accident scenarios. The classification module employs dynamic weighting to assess accidents as mild, moderate, or severe, coupled with a tri-level response protocol. For route optimization, the path planning component uses a refined A* algorithm with a four-dimensional cost function. Tests show that the system delivers 98.2% detection accuracy with subsecond response times, maintains 93.5% reliability in harsh weather, and enhances rescue operations by 23.7% in efficiency. This solution offers a robust technical approach for smart traffic management systems.