Research And Implementation Of Small Abrasion Measurement System For Steel Rails

With increasing capacity and usage frequency of railway, the speed of steel rail abrasion is faster than ever. The major worn part is rail head. When a steel rail gets abraded, its basal area will be reduced and its intensity and pressure resistance strength will be weakened, which poses great threats to railway safety. Especially, in recent years the government has been putting forward the concept of "Hour Economic Circle". A large number of high-speed railways have been built between cities. The running speed and train frequency have reached highest points ever. Therefore, it is necessary to conduct abrasion examination and test on rails. This work has always been one of the major tasks of the Railway Maintenance Department, which plays an important role in ensuring railway safety.At present, there are lots of methods for detecting rail abrasion, including commonly-used caliper measurement methods, visual analysis methods, detection methods based on sensor technology, digital processing methods based on CCD cameras and so on. The first two methods require much experience and high labor intensity, but they have low efficiency and accuracy, which fails to meet the development need of current high-speed railways. On the basis of previous researches on rail abrasion measurement methods based on computer visual technique and digital processing technique, this paper designs an abrasion measurement system for small-sized steel rails-the Rail Abrasion Measurement system, which can conduct abrasion calculation of a rail's basal area based on digital images.This paper firstly introduces current detection methods for rail abrasion, the background of this subject and the current technical situation of rail abrasion detection, and then describes the computer vision technology and image processing technique used in this system in detail. Camera calibration technique is an important part of this system. Through a variety of analysis and comparison over various calibration techniques, this system chooses Zhang camera calibration method for system calibration. Key technical modules in this system include image pre-processing, camera calibration, contour recovery and image matching. This paper introduces and analyzes each of these technical modules in detail. This system conducts programming based on VC++6.0. The last part of this paper analyzes experiment results of this system, and provides suggestions on further improving functions of the system.