The Moving Vehicle Detection Based On Machine Vision

Along with the development of digital image processing,video processing has been used in traffic detection step by step,which is used to improve the current transportation system.Using one or more cameras to collect information of traffic status on driveway,the information of vehicle classification,speed measurement and multiple point vehicle counts etc.are provided by image processing.Furthermore, moving vehicle detection is the basis and prerequisite for vehicle recognition and tracking.Thus,it is essential to detect moving vehicle objects stably and effectively.Through summing up and analyzing the characteristics of existing methods in vehicle detection,this paper pays great attention to investigate the practical techniques for moving objects detection under the condition of a fixed CCD camera,including such techniques as initial background extraction and background update,automatic detection of moving objects,design and implementation of detection system,and so on.The following research has been accomplished:On the basis of traditional background subtract method and inter-frame difference method,a kind of car detection method which uses background subtract method, inter-frame difference method and mathematical morphological is proposed.At first adopted inter-frame difference method to obtain initial information of images,and initial background model is sat up by cumulating and amending.Because the effect of rapid beam change to background is very obvious,the thesis introduces a kind of selective background update method.Then background subtract method complemented with inter-frame differencing method are utilized to detect the moving object area effectively.And the O tsu method is proposed to segment moving car.Finally,do some morphological processing on the detection results to smooth border and eliminate burrs.Design and realize a video car detection system.This system is developed using Visual Basic and NCT Image Studio.This system mainly realizes background extraction, background updating,vehicles detecting and segmenting,morphological transform,etc.The experimental results show that the method proposed in the thesis can get vehicle object integrally.The algorithm above has small amount of count,good real time capability and high detection precision.