Research On Image Processing Of Online Automatic Vehicle Deviation Test System

It's a very common defect of vehicle deviation.When the vehicle moving, this defect may cause the driver to exert a corrective force on the steering wheel,this makes it harder to handle the vehicle,the driver feels exhausted easily,worsen turning around heavily,parts and tyres are wore-out quickly.This problem may cause braking deviation easily,at last,lead to phenomenon of sideslip of vehicle.Especially on high speed road,when the drivers brake the vehicle urgently,it may cause very seriously sideslip,as a result,greatly traffic accidents happened,it causing great harm and injury serious.National standards GB 7258-2004 Safety specifications for motor vehicles operating on roads are compulsory standard,the 4.13th item set that:when motor vehicles driving in a straight line,the line of anterior and posterior axis must be in keeping with the center line of driving track.However,there haven't existed a system that can test deviation of vehicle scientifically and accurately. Vehicles are alignmented usually before delivery. Only after people find vehicles deviation they maintenance them when in practical use. In order to prevent unqualified vehicles inflow into the market,it's necessarily to test all the performances of them before delivery.This thesis put forward a system that can test deviation of vehicle on line.This system connected with product line.When vehicles running pass the portal frame,photo-electric switches are triggered, Charge Couple Device are driven to gather the pictures of vehicle.Image processing is the center of this thesis,so we research mainly on image processing and image sampling based on LabVIEW of this system. In chapter 1,1 introduce the concept and development of image processing and machine vision. In chapter 2,analyze how to choose the right hardware of image processing,it includes:lenses, charge couple device, photo-electric switches,light source of the system,lighting devices and so on.In line with the practical situation,we may choose different devices.After experimented again and again, we come to the conclusion that this system accord with the accuracy of the factory ask for efficient,well-adapted and high automation.In chapter 3,after comparison and study,we choose the most scientific methods of image processing and identification:median filter, Laplacian of Gaussian sharpen operator.After image enhancement,image smoothing,the edges of the pictures become more distinct and intact,then we use template matching,we can get the coordinates of the vehicle,after calculating,the results are got. In chapter 4,based on LabVIEW,we mainly use modularity ideology,apply functions of IMAQ,achieved all the functions above,test them in real time.Different from extracting all the outline of the vehicle in order to locate the it,we put forward a new way to location.We make a mark that can assisted we test,we put it on the vehicle.This method reduce the complexity of the program processing.It can be used with little problem and directly perceived through the senses.It is high accuracy,adapt to the complicated environment. In chapter 5,analysis the accuracy of the system based on a experiment.In the last place,make a simple summarize of the research result and shortcomings..