Design And Application Of The Surface Defects Detection System Of Thermal Heavy Rail Based On Machine Vision

Automatic detection of heavy rail surface defects, has long been expected to solve the problem.Especially the online detection of the thermal heavy rail is the focus of attention. Current detection of thermal heavy rail surface defects more uses the method of detection through naked eye or artificial visual detection method, but this method has many weak points, such as high labor intensity, easy to miss detection and lack of guarantee to the quality of detection accuracy. With the development of computer hardware and software technologies, as well as the smart technology of image processing and pattern recognition, the smart work by machine instead of manual work needs more and more. It is on this basis, the paper proposes a surface defects detection system of thermal heavy rail based on machine vision. System has the prominent merits of non-contact, appropriate precision, strong ability for scene antijamming, without interference of the production process and so on.Firstly, this paper introduces the background of the research, showing the inadequacy of the current detection methods of thermal heavy rail surface defects,and the application of machine vision technology in the field of detection of surface defects at home and abroad. It introduces the principle of machine vision detection, according to the detection principle, designs the overall system plan, including the design of lighting methods, the choice of filter, camera lens and image acquisition card,for the problem of massive data of the image,it uses the disk array technology. It describes the morphological characteristics of heavy rail, completing the design of discrete image acquisition unit and installation layout, displaying pictures of each discrete image acquisition unit. It introduces the digital image processing method of image enhancement, segmentation, and edge extraction for thermal heavy rail picture, studies the classic algorithm of threshold segmention, and proposes an algorithm based on Genetic Alogrithm and B spline function aiming to complex flaw edge surrounded by high noise background,and discusses the application feasibility of the defect recognition method based on comparison in quality control of the heavy rail surface. System software is modular in design,it is developed by Microsoft Visual C++6.0. The programming is achieved by using calling MATLAB through proprietary program interface of MatCOM and MATLAB 7.0 colligated the advantage of VC and Matlab. SQL Server 2005 is used to manage system information,equipment information and image data.Various functions module parallel processing is realized using Windows's multithread and multi-view. This model ensures data collecting in time and various operations taking advantage of the Windows. This system has friendly interface.In conclusion, this paper introduces the application of the surface defects detection system of thermal heavy rail for field, to verify the feasibility of the system, and also summarizes the improvement of system, which lays a good foundation for further research.