The Industrial Detection Of X-Ray Based On Image Processing

As an important branch of machine vision and image processing technology in many industrial sectors, computer application technology have been used widely and successfully,and industrial-ray flaw detection analysis is one of them. To ensure the safety and reliability of industrial products, which allow the enterprises used the testing technology of nondestructive that doesn't destroy the original shape or using function.The using of X-ray technology in the testing technology of nondestructive of approximately 40%, indicating that ray detection occupy a very important position. Traditional X-ray detection and imaging based on the film and artificial,such methods are objective factors, low efficiency, and film injury; There are also many drawbacks at the same time Film Imaging, for example the real-time imaging, requirements of inspection and evaluation,and of the high costs and inconvenience,and the managements. To designed a corresponding defect testing X-ray image processing software platform has become very necessary. This paper focuses on the deficiencies of digital image processing algorithm research, and laid the foundation for the development of the image processing software platform.In this paper some analysis of digital image quality factors of the X-ray detection have been done. And analysised the course of the changing quality in digital convertion. The focus of the pros and cons of the various picture formats in image acquisition and storage aspects have been given, and identified the better picture storage format.This paper have do some studies on the defect image by using the pretreatment and extraction technology. In the image preprocessing, the use of image smoothing and filtering technology in the removal of X-ray detection image noise, and the introduction of non-linear gray-scale transformation and fuzzy enhance image contrast enhancement technology. For expounded Segmentation of the edge and the core idea commonly used methods, introduced the improved algorithm, which focus on multi-scale wavelet-based and algorithm based on morphology. Experiments show that the former approach to multi-scale realization of the defects detected, respectively, the effect of the simple edge detection to the obvious; then an algorithm to eliminate noise drowned and over-segmentation are more obvious role in the resistance to noise algorithm also fairly good results.In the X-ray image of the defect measurement and recognition, the reserch of the defect size measurement technology have been done, and the design of the multi-defect tracking, and filling of muti-defects and the calculation algorithm of characteristic parameter. Experiment proved that the design of the algorithm can extract defect of the basic characteristics of parameters in a better way. This paper established a preliminary defect recognition system, it has laid the foundation for the future development of software platform.