Weak X-ray Images Of Carbon Materials Using Wavelet Analysis To Identify

With the formation and gradual development of modern carbon industry in our country, there is an urgent need for methods in the testing of interior defect for many manufacturers and expenditure departments to control and improve their product quality. However, carbon material is a kind of loose compsition substance, which has uneven density and character and belongs to a half-porous medium. The routine ways of NDT(Non-Destructive Testiong), such as testing of electricity resistance, striking and ultrasonic testing, cannot acquire satisfactory testing outcomed. Many researchs in experiment show that based on the technology of image processing, it is practicable for making use of X-ray in NDT of carbon material.Because the structurial traits of carbon materials, their X-ray testing images are full of strong noise and have inferior quality. What's more, because of the limitation of current technology, the methods of hardware in image enhancing have too much cost and infeior analysis outcome, which limits their application in practical engineering. Therefor, more energy is expended on the software methods in X-ray image processing.The traditional X-ray testing system are not digitalized, which cannot cater to thereafter image processing and computer identification. In order to acquire digital image cheaply, based on the principle and method of X-ray testing, this thesis consummates it and exploits a digital X-ray testing system of low-cost and high reliability. And according to the need of practical produce, a process of online check and automatic judgement of interior defect in carbon material is designed.Aim at the faint character in X-ray testing image of carbon materrial, firstly, traditional processing methods are used to pretreatment and improve the X-ray image. Then, from the aspects of mathematics and filter, the decomposion and reconstrcture of two dimensions wavelet transformation, the best choice of basic wavelet function and the layer of decomposition are studied deeply. Based on this researching, the methods of high frequency enhancing, soft threshold value and reverse mouldingboard of wavelet analysis are deplioted. At the same time, for judge the shape and size of defects, the edge detecting of X-ray testing image is further studied.The experimentation of practical X-ray testing image shows that these methods and ways of wavelet analysis can fundmentally improve the quality of image, enhance defect character markedly and recongnize the defect successfully. These results can provide computer judgement and model matchment of interior defect in carbon material with important technology base.