高速公路视频事件检测分析系统的研发及应用

Research and Development and Application of Highway Video Event Detection and Analysis System

为了解决高速公路视频监控主要靠人工巡检发现事件,存在数据错漏、工作量巨大、信息不及时等问题,开展高速公路AI视频事件检测分析与处理关键技术研究开发,重点研发多服务器多GPU训练集群、大规模视频GPU推理并行优化、基于图像的目标检测算法的、多目标跟踪等关键技术,开发出高速公路视频事件检测分析系统,实现高速公路海量视频数据的智能分析,研发成果已在十多条路段应用,大大提高了高速公路隧道运营管理效率。

In order to solve the problems such as data errors and omissions,huge workload,and untimely information and so on with the highway video surveillance mainly relies on manual inspection to find events,this paper carries out the research and development of key technologies for highway AI video event detection analysis and processing,focusing on the research and development of multi-server and multi-GPU training clusters,large-scale video GPU inference parallel optimization,image-based object detection algorithms,multi-target tracking and other key technologies.It develops a highway video event detection and analysis system,achieves intelligent analysis of highway massive video data.The research and development results have been applied in more than ten road sections,which have greatly improved the efficiency of highway tunnel operation and management.