Font Size: a A A

Research On Rail Fastener Detection Method Based On Computer Vision

Posted on:2017-07-13Degree:MasterType:Thesis
Country:ChinaCandidate:X G ZengFull Text:PDF
GTID:2492305906966049Subject:Instrument Science and Technology
Abstract/Summary:PDF Full Text Request
As the railway transport capacity demand continues to increase,the high-speed railway has become the focus of the direction of railway construction in our country,how to make the inspection of the continuous expansion of the railway,guarantee the railway safety have become increasingly prominent.And the lack of railway fastener will likely lead to the derailment of the railway,causing serious accidents,so the rapid automatic detection of railway fasteners become an important issue.In this paper,the fastener detection system designed before of the subject group is based on the combination of structural light and vision,it has been able to achieve this function,but there are some problems and shortcomings,mainly include:1)System structure is more complex,the structure light positioning system may not produce the signal to trigger the camera when occluded in the fastener.2)Using a camera to capture fastenesr on both sides of the rail,which may reduce the image quality,resulte in false detection and decrease the fastener identification accuracy.3)Not to do the protection of the vehicle visual detection box,which can lead to the the glass measurement window scratched,and need to replace the glass window periodically.In order to solve the above problems,this paper mainly does the following work:1)Optimizing system structure,using only visual sensor to detect the fastener,positioning the fastener by image sequence.2)The hardware system is composed of four cameras to detect the fastener on both sides of the tracks,to avoid the problem of gray difference between the two sides of the image.3)using two kinds of fastener recognition algorithm to detect the image fastener,first use the direction field algorithm to classify the fastener image,and then use the algorithm based on Harris points to two times,improve the accuracy of the identification.4)Increase the protective measures against the vehicle borne vision probe.In order to reduce vibration and shock,the shock absorber is used for vibration reduction,and the model is used to analyze the parameters of the shock absorber.In order to reduce the impact of dust,water droplets,sand and gravel on the measurement of glass windows,the relationship between the structure parameters and the protection ability was researched,and the experimental data and theoretical reference for the design of the hydrodynamic form were analyzed.At last,based on the research of hardware design and algorithm,the experiment device in the laboratory is used to test the algorithm.The feasibility and the performance of the algorithm are verified.
Keywords/Search Tags:fastener detection, phase correlation, Harris, direction field, hydrodynamic form
PDF Full Text Request
Related items