基于AI的路面病害自动识别系统研发与应用  

Development and Application of Automatic Identification System for Pavement Diseases Based on AI

作　　者：翟晓成 丁攀 苟晓锋 刘俊杰 刘士宽 费昀 ZHAI Xiaocheng;DING Pan;GOU Xiaofeng;LIU Junjie;LIU Shikuan;FEI Yun(Highway Development Center in Lanzhou Gansu,Lanzhou 730000,China;China Trans Geomatics Co.,Ltd.,Beijing 100094,China)

机构地区：[1]甘肃省兰州公路事业发展中心,甘肃兰州730000 [2]中交宇科(北京)空间信息技术有限公司,北京100094

出　　处：《地理空间信息》2024年第9期84-88,共5页Geospatial Information

基　　金：2022年甘肃省交通厅科研课题。

摘　　要：在日常养护巡查路面病害识别方面,大多采用肉眼观测实际路面或人工判读路面影像的方式,而人工识别病害存在漏检现象,准确率和工作效率较低。为解决该问题,以AI代替人眼进行路面病害识别,并进行路面病害的可视化管理。基于AI的路面病害自动识别系统可显著提升路面病害识别效率,为路面病害自动化识别技术的发展做出贡献,提高养护管理科技水平。In terms of pavement daily inspection,the modes of visual observation of actual pavement or artificial interpretation of pavement image are frequently adopted.Regardless,the artificial recognition method would cause missed detections,low accuracy and low work efficiency.On the basis of this,we instead artificial intelligence(AI)of human eyes to identify pavement diseases,and provides visual management of pavement diseases.The automatic identification system for pavement diseases based on AI has improved the efficiency of pavement disease identification significantly.The achievement has made contributions to the development of automatic identification technology for pavement disease,and improves the technical level of maintenance management.

关 键 词：日常养护巡查 AI 路面病害自动识别 病害可视化 

分 类 号：P208[天文地球—地图制图学与地理信息工程]