摘要
首先分析了既有隧道衬砌裂缝检测系统存在的问题,然后应用最新技术发展成果,提出基于图像识别技术的隧道衬砌裂缝检测系统。该系统能对隧道衬砌图像予以高速采集和存储,其后端采用基于深度学习的裂缝识别算法对海量的隧道衬砌图像数据进行快速识别,并提取裂缝特征参数。将该系统安装在现有轨道车上进行了试验,结果表明该系统可以50 km/h的速度对1 mm以上衬砌裂缝无遗漏采集。
Firstly,the existing problems of tunnel lining crack detection system were analyzed.Secondly,the tunnel lining crack detection system based on image recognition technology was put forward by using the latest technology development results,which could collect and store the tunnel lining image at high speed. The back end of system adopted the crack recognition algorithm based on deep learning to quickly recognize the mass of tunnel lining image data and extract the characteristic parameters of the cracks.The system was installed on the existing rail vehicle and tested.The results show that the system can collect more than 1 mm lining cracks at 50 km/h speed.
出处
《铁道建筑》
北大核心
2018年第1期20-24,共5页
Railway Engineering
基金
中国铁路总公司科技研究开发计划(2016G006-B)
中国铁道科学研究院基金(2016YJ029)
关键词
铁路隧道
裂缝识别
试验研究
图像识别技术
深度学习
Railway tunnel
Crack recognition
Experimental research
Image recognition technology
Deep learning
