摘要
为解决目前路面破损状态自动检测采用专用摄影检测车检测费用高、检测结果需要专业软件分析不易普及应用的问题,提出一种基于行车记录仪的路面破损状况识别方法。以行车记录仪采集的视频图片、GPS数据为基础,建立基于行车记录仪图片的高速公路路面状况巡查及报警系统,实现路面破损状况识别及报警。系统利用行车记录仪采集的图片,基于卷积神经网络的深度学习算法,实现路面的破损自动识别;利用行车记录仪GPS定位数据以及GIS电子地图,实现行车记录仪图片的GIS地图实时显示及报警。
In order to solve the problems of high cost of auto-detection of pavement damage status using special photographic inspection vehicle and the difficulty of popularization and application of professional software analysis for detection results, a method of pavement damage status identification based on traffic recorder is proposed. Based on the video pictures and GPS data collected by the traffic recorder, a highway pavement condition inspection and alarm system based on the traffic recorder pictures is established to realize the recognition and alarm of pavement damage. The system uses the images collected by the traffic recorder and the deep learning algorithm based on convolution neural network to realize the automatic recognition of road surface damage. The system uses the GPS positioning data of the traffic recorder and the electronic map of GIS to realize the real-time display and alarm of the GIS map of the traffic recorder pictures.
作者
张月
ZHANG Yue(Hebei Qu Gang Expressway Development Co.,Ltd.,Dingzhou 073000,China)
出处
《交通与运输》
2019年第4期54-57,共4页
Traffic & Transportation
基金
曲港高速公路智慧运营及管养辅助决策支持系统关键技术研究项目资助
