Research On The Detection Of Bullet Train Shaft End Bolt Based On Image Processing

The bullet train plays an important role in china railway passenger transport.Due to the high running speed,the abrasion problem is relatively prominent.So,daily safety maintenance is very important.The shaft-end bolts of the bullet train bogie detection is one of the key tasks among them.At present,the bolt detection is mainly completed by hand.However,large quantity and small size make it difficult to detect.Therefore,the study of an automatic detection method of shaft bolt based on image processing is of great importance to guarantee the driving safety and reduce the labor force.This paper is the continuing research on train operation trouble detection system,through the analysis of shaft end bolt image characteristics,makes the strategy of reducing bolt detection area step by step,and finishes the automatic detection system design of the bullet train shaft-end bolts based on the image processing technology.Firstly,an adaptive extraction method of the bogie area based on the combination of gray projection and Hough transform is proposed based on the characteristics of bogie.The image binarization and morphological pretreatment is used to enhance the regional characteristics of bogie.On this basis,the image preprocessing method of bilateral filtering and Laplacian operator enhancement is adopted for bogie region image,and the effect of noise and light on target detection is reduced.Then,the positioning and extraction of the axle area are realized by using hofhoff circle transform and image statistical information.According to the characteristics of the end cover area,a recognition and extraction method of multi-grayscale template matching end cover based on regional division is proposed,which can effectively realize precise positioning of the end cover area.Next,the bolt detection range is divided according to the position relation between the end cover center and bolt.And the method of multi-grayscale template matching was used to detect the shaft end bolts according to the characteristics of bolt area.At last,an automatic testing system for shaft end bolts of bullet train bogie is designed and constructed according to the bolt detection process.The actual bullet train image is used for detection experiments.The experimental results show that the detection rate of the method for the bolts of the bullet train bogie is 89.26%,and the detection efficiency is greatly improved.An experiment is carried out at the Beijing bullet train section and the detection result meets the actual usage requirements.Thus verifies the design of shaft end bolt detection system has good engineering application value.