摘要
针对已有汽车轴距差检测方法存在的缺陷,提出一种基于立体视觉的非接触式汽车轴距差动态检测方法。该方法基于双目立体视觉成像原理及轴距差测量模型,应用已标定的摄像机采集车轮图像,采用数字图像处理技术对车轮轮毂图像特征进行提取;建立三维重建模型;为得到轮毂中心坐标,采用最小二乘法对轮毂的50个边界点进行拟合;针对曲线行驶时产生的曲线补偿差值,建立曲线矫正模型,应用Matlab对20个偏转角β进行拟合验证。研究结果表明:通过实车实验及3种不同检测方法对比分析,本系统具有可靠的稳定性,满足车辆检测要求。
Aiming at the defects of the existing detection methods for vehicle wheelbase difference, a dynamic detectionmethod based on the stereo vision for the non-contact vehicle wheelbase difference was proposed. The method was basedon the imaging principle of binocular stereo vision and the measurement model of wheelbase difference, and appliedcalibrated cameras to capture wheels image and using the digital image processing technology to extract wheel hub imagefeature. A 3D reconstruction model was established. In order to get the center coordinates of wheel hub, 50 boundarypoints of the wheel hub were fitted by the least square method. Aiming at the curve compensation value generated duringcurve driving, a curve correction model was established and 20 deflection angle β were fitted by Matlab. The results showthat the stability of the system is verified through testing car in practice and comparative analysis of three differentdetection methods, which can meet the vehicle detection requirements.
出处
《中南大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2017年第11期2960-2966,共7页
Journal of Central South University:Science and Technology
基金
国家自然科学基金资助项目(51205164)
吉林省科技发展计划项目(20150204025GX)~~
关键词
非接触式
动态检测
双目立体视觉
三维重建
曲线矫正
non-contact
dynamic detection
binocular stereo vision
3D reconstruction
curve correction
