摘要
公路隧道结构参数、车辆多维扰动对视觉系统产生影响,数字相机安装参数优化和外参数辨识是开展视觉采集和图像处理的前提。首先,分别以多相机安装中心和单相机安装位置为研究对象,基于应用需求构造了优化目标函数,对视觉系统的安装参数优化问题进行数学描述,在此基础上考虑车辆多维扰动对系统的影响,基于多刚体运动变换提出了一种外界扰动作用下的数字相机耦合位姿、物距等外参数辨识方法;其次,根据不同设计时速下隧道结构参数,通过数值方法分析了视觉系统的最优设计结果;最后,根据理论分析研制出公路隧道检测车。工程应用表明,该系统能够快速准确的完成检测任务,为实现我国公路隧道智能化检测与信息化养护提供了一种全新的技术手段。
Road tunnel structureparameters and multi-perturbations have influence on the visualsystem of vehicle. The installation parameters optimizationand extrinsic parameters identification are the precondition for vision acquiring and image processing. Firstly,the install center of multi-camera and location of single-camera are studied. the target function of optimization based on application is formulated. The mathematical description of optimal install parameters of vision system is established. Onthis basis,theimpact of vehicle precondition on system is considered. The extrinsic identification model of coupling pose and object distance of digital camera under vehicle precondition is built based on multi-rigid body motion transformation. Secondly,based on the different parameters of road tunnel structure,the optimal design parameter of vision system is determined by numerical method. Finally,a road tunnel detection vehicle is developed according to the theoretical analysis. Engineering application shows that the system can accomplish detection assignments with rapid speed and high accuracy. It provide some new kind of technology for intelligent detection and informatization maintenance of road tunnel.
作者
刘晓
李洋
薛春明
刘博
段英杰
Liu Xiao;Li Yang;Xue Chunming;Liu Bo;Duan Yingjie(Shanxi Engineering Research Center for Road Intelligent Monitoring,Shanxi Transportation Research Institute,030006,China;Key Lab of highway Construction & Maintenance Technology in Loess Region,Shanxi Transportation Research Institute,Taiyuan 030006,China)
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2018年第5期152-160,共9页
Chinese Journal of Scientific Instrument
基金
国家自然科学基金(51705229)
山西省交通厅科技计划(2017-1-25)项目资助
关键词
公路隧道
隧道检测车
机器视觉
图像处理
参数辨识
road tunnel
tunnel detection vehicle
machine vision
image processing
parameter identification