摘要
文章对盾构隧道表面图像采集系统进行设计,提出一种新颖的隧道裂缝图像智能识别方法。图像采集系统包括面阵相机及光源一体化模块、相机支架、工控机和供电系统,该系统可实现隧道表面图像多目相机同步采集,图像辨识精度达到0.2 mm。裂缝图像智能识别方法主要包括:基于边缘检测的深度卷积网络和基于图像后处理的裂缝识别算法,该图像识别算法可实现像素级的裂缝纹理检测以及精细化提取。通过现场实验和数据分析,文章提出的图像采集系统和智能识别方法可为实际隧道裂缝检测提供较好的技术支持。
This study designs a shield tunnel surface image acquisition system and proposes a novel intelligent identification method of metro tunnel cracks.The image acquisition system includes the area scan camera and light source integration module,camera bracket,industrial control computer and power supply system.This system can realize synchronous acquisition of tunnel surface image by multi-vision camera,and the image recognition resolution can reach 0.2 mm.The intelligent recognition methods of crack image mainly include:depth convolution network based on edge detection and crack recognition algorithm based on image post-processing.This image recognition algorithm can achieve pixel level crack surface detection and accuracy pick-up.Through field experiments and data analysis,the image acquisition system and intelligent recognition method proposed in this paper can provide better technical support for actual tunnel crack detection.
出处
《现代城市轨道交通》
2022年第11期1-6,共6页
Modern Urban Transit
基金
科技部国家重点研发计划资助项目(2016YFB1200402)
北京交通大学2011协同创新中心项目(M22JBXT00020)。
关键词
地铁
机器视觉
图像采集
图像处理
边缘检测
隧道裂缝
metro
machine vision
image acquisition
image processing
edge detection
tunnel crack
