利用双目视觉方法的车身振动响应识别与分析  

Recognition and Analysis of Vehicle Body Vibration Response Using Stereo Vision Method

作　　者：林文梁 王晓佳 李国兴 吴通 贠星辰 LIN Wenliang;WANG Xiaojia;LI Guoxing;WU Tong;YUN Xingchen(Taiyuan University of Technology,College of Mechanical and Vehicle Engineering,Taiyuan 030024,China)

机构地区：[1]太原理工大学机械与运载工程学院,太原030024

出　　处：《重庆理工大学学报（自然科学）》2021年第11期10-17,共8页Journal of Chongqing University of Technology：Natural Science

基　　金：国家自然科学基金青年基金项目(51805335);山西省自然科学基金青年基金项目(201801D221222)。

摘　　要：车辆状态检测和评估对安全行驶非常重要。为此提出一种基于双目视觉测量车身振动的方法,用于测量车辆受激励下的振动。在特制减速带激励下,车辆的姿态会有较大幅度的俯仰,通过惯性测量单元(IMU)实时监测车身角度、加速度等信息,固定于车身的双目摄像头通过识别圆形标志点反推车辆俯仰运动。试验结果表明:双目视觉方法对车身受激下振动的测量准确性良好,与IMU数据贴合度较好。鉴于车身刚体运动的主要频率分布在0.25~15 Hz,证明双目视觉测量车辆振动的方法可以在不损失信息的前提下满足车身振动测量对准确性的需求,同时能够改善IMU测振动的漂移问题。In the process of vehicle use,periodic and real-time vehicle status detection is very important,which is related to the safe driving of the vehicle.The health status of a vehicle suspension can be detected by its response when it is excited.Under different conditions,the response of the excited vibration will be different in the time domain and frequency domain.Traditional vibration detection is realized by accelerometer,but its inherent drift problem affects the accuracy and reliability of its measurement results.In order to achieve more accurate and reliable detection,a method based on stereo vision to measure vehicle body vibration was proposed.Taking advantage of the non-drift characteristic of visual measurement,the method was applied to measure the vibration information of the vehicle under excitation.In the test,under the excitation of a special speed bump,the vehicle’s posture will have a relatively large pitch,the inertial measurement unit(IMU)was used to monitor the body angle,acceleration and other information in real time.A circle marker with a known circle center spacing was positioned in the front of the vehicle,and a stereo camera fixed to the car body took sequence pictures.Through image recognition algorithms such as edge detection and circle center fitting,the effects of different algorithms were compared,and the best solution was selected to obtain the pixel coordinates of the center of the circle marker.At the same time,based on the principle of stereo ranging,a two-point ranging method was proposed as a constraint,which improved the accuracy of depth information and ensured reliable ranging results.On the basis of obtaining accurate pixel displacement and mark point depth information,a motion decoupling method was proposed,which simplified the vehicle motion model,and a two-degree-of-freedom model of the vehicle was proposed that analyzed the movement relationship between the front wheel and the rear wheel when they were excited.The pixel displacement and depth information of the two marker

关 键 词：双目视觉 车身振动 高帧视频 运动解耦 

分 类 号：U467.4[机械工程—车辆工程]