Detection On Wagons Bogie Failures Based On Image Processing

Accompanied by the heavy-duty,high-density opening lines of railway wagons,the workload of wagons maintenance is increasing day by day and the traditional methods of wagons inspection have low working efficiency.It is difficult to complete the wagons failure detection efficiently.The fault detection system based on image processing technology can realize the automatic identification of typical failures of wagons and complete the detection of wagons fault further.This paper analyzes the typical fault features of wagons bogie image.Designing a typical fault automatic detection system of wagons bogie based on digital image processing.Achieved the automatic detection of three types of faults which includes center plate screw,across bar and safety chain.The main work of this paper: The image preprocessing of bogie images.It can affect the quality of the bogie images even cause the fasle and missing inspection of bogie fault due to acquisition conditions and environmental factors.Therefore,it needs to be preprocessed to improve the image quality of the collected bogie images.This paper applies median filter algorithm for image smoothing,and uses single scale Retinex algorithm for image preprocessing,that realizes image enhancement of fault images and de-fogging of atomised images.The extraction of bogie images' fault feature.Applying the improved random Hough transform algorithm of circle detection to extract the characteristics of center plate screw,which realizes the failure detection of center plate screw;it completes the detection on bending failure of across bar,combining with the Hough transform line detection algorithm and the deformation characteristic of bending across bar;The structural features of the target area is used to locate the safety chain,and with the help of the feature of the binary image achieve the detectiong on dropped failure of safety chain.