The Design Of Weld Defect Evaluational Software Based On X-ray Real-time Imaging System

With the development of X-ray testing and digital image processing technology,the technology of weld defects test has also gradually transited into computerintelligent identification. Adopting the computer to test and identify the digital imageof the welding line is largely accepted in terms of its testing efficiency, economicbenefits as well as its convenience.This thesis is based on the actual engineering needs and chooses the digital BMPimages which are got when using X-ray to treat the welding line as the study object.Then, visual C++and visualized programmes are designed and also the software ofevaluating the system is designed. This software is mainly composed of such threemodules as image processing, defects identification and defects evaluation.The major contents of the image processing module include the geometrictransformation of the digital image, the image enhancement,image point calculation andboundary extraction. Geometric transformation deals with the translation,rotation,mirrorimage and scaling processing; the image enhancement covers smooth noise reduction andmedian filtering; the image point calculation includes the design of histogram,gray linertransformation,gray window transformation and gray extension.The purpose of design ofhistogram is to calculate the digital image’s gray value and to conduct the binarizationprocessing while adopting liner transformation. Through the window transformation,the welding line and parent metal can be separated. The defects can be revealed byadopting gray extension. When it comes to edge detection operator, there are Robertoperator, Sobel operator, Prewiit operator, Kristch operator and Guass-laplaceoperator. Due to the fact that Guass-laplace Operator is able to reduce the picturenoise while extracting the outline of the defects, this proposed experiment adoptsGuass-laplace operator to extract the defects’ outline.Defects identification module adopts detailing and central marking methods tostatistically analyze the number of defects, then vertical projection is used to swiftlycalculate the average radius, perimeter and area of the defects.Defects evaluation module is designed on the basis of industrial standard ofP.R.C.(JB/T4730.2-2005) and is visualized. When the relevant data are put into thedialogue box of the defects evaluation menu, the welding line defects levels can beaccurately assessed. The results of the experiment shows that this software can effectively identifyand assess the defects and it is with some practical significance.