The Technical Studies Of Rail Detection Location Based On Video

In recent years, with the increasing investment in china’s railway construction and urban rail traffic,the railway plays a more important role in our society as same as the inspection of track geometry. Train positioning is a key point in the inspection of track geometry.The reliable、accurate location information is the base on the train’s safety operation and services.So a research on autonomous train location method is of great theoretical and practical significance.With the rapid development of computer vision technology and image processing technology,non-contact inspection technology gradually become the main research direction of the track inspection.This paper presents a fast response positioning technology which use an adaptable computer vision technology.In order to locate the train,the technology use a video processing method to count the number of the fasteners which is used to database system. The specific work is as follows:According to the need of the positioning system’s image acquisition and processing module,this paper give an analysis and selection on the hardware module of the image acquisition.And this thesis also give a comparison and choice to image pocessing method.The exact location of fastener from original image is an important prerequisite of the train location system.Firstly.the paper compare the existing variety of image pre-processing techniques and choose the median filter which has a inhibition on video noise.Then a method for edge detection based on Laplacian operator is used on binary image, which is compared with Hough line detection algorithm.At last, this thesis use a method of scanning highlights to locate the rail and sleeper. This method set a basis for the extraction of the fasteners in video sequencesThis paper use a method of area tracking to count the number of the fasteners.Tag the fasteners in a video sequences by tracking control table. Entering zone and leaving zone are set to judge the appearance of new fastener and disappearance of a old one.To solve the problem of repeating and missing count.the paper use a method by comparing the disitance between centroids and set up a self-test function.Establish an appropriate motion model of trains.The thesis improve the accuracy and fault tolerance of the location system by using multi-information fusion which is made of pulse sensor and fastener numbers.At last,the paper use the Matlab to verify the above algorithm.The conclusion come out that,the method of extracting the fastener has a good stability, and can quickly locate the area which is including fastener.Tracking method can effectively calculate the number of fasteners.To reduce the positioning error,the fusion model fuse the pulse sensor and the processing of fasteners counting completely.