A Correction Method Of Linear CCD Image's Moving Distortion

All kinds of images or video as the carrier of security detection systems,with the development of technological,have been developed.And it widely used in security places,such as airports,government,prisons,bank of reservoir,embassies,military,ports,customs,border,and so on.Vehicle chassis is the key part of the security.The detection system based on computer vision technology is gradually replacing the artificial verifying way in vehicle chassis.In this paper,we not only study the theory which linear CCD compose of vehicle chassis image,and also analyze the moving distortion's^mechanism generated by the factors such as non-uniform vehicle velocity.It provides a correction method for the distorting vehicle chassis image,and eventually achieves implements maximize approximation between standard and the distortion vehicle chassis image.For this paper,we analyze a large amount of image cues extraction methods for the moving distortion of the image characteristics.Thus,a unique cues extraction method,the image's line-cue-atom cues,has been proposed so as to adapt to the particularity of containing moving distorted behaviour's image.This paper was proposed a line-cue-atom matching algorithm,the cross matching algorithm,analyses a large number of cues matching algorithms.Meanwhile,in order to maximum similarity between standard and distortion vehicle chassis's image,we puts forward a cross-matching algorithm by coarse matching.Vehicle chassis images require as source data of other modules.And we are must to identify the effective of the algorithm.Finally,we achieve the algorithm in the environment by VC software.While a large number of experiments were verified the robust and practicality of algorithm for the moving distorted vehicle chassis image.