Design And Application For Pretreatment Of Defect Recognition For X-Ray Weld Inspection System

With the development of X-ray detection and image processing technology, weld defects detection gradually transits from artificial detection to computer intelligence recognition. Analysis and recognition digitized weld image by computer is approved by the people in detection efficiency, economic profits, convenient and pragmatic, etc. This paper is based on the actual project demand, design and develop the system of acquisition and management technology of the effective information from real-time weld image. This system include four modules, they are image acquisition, image preprocessing, image management and defects recognition.Image preprocessing model contains two functions: adjustment of correction value and image filtering. The former means that this system firstly conversion image from the gray level of 65536 to the gray of 256, then adjust to correction value according to actual situation which would have preprocessed to it. This research has put forward eight kinds of image mapping methods, each method is improved though original ones. And finally the result is satisfied using manual adjustment. And image filtering includes image denoising and image enhancement. Image signal is often interfered by many kinds of noise in the process of produced, transmitted and recorded. So it is a very important work to use the suitable method reduction noise. This thesis mainly researches moulding-board filter and median filter in image denoising technique. This article also specifically improved the two enhancement, which set size and numerical matrix freely. The results show that this enhancements have better effect. As the noise image with the edge of fuzzy, this paper use the methods of image enhancement to improve image contrast. Contrast enhancement is the treatment of specific image to highlight some of the information and weaken or remove some of the information does not need. There are two methods of enhancement: the histogram equalization and piecewise linear. This paper compared with two methods among the advantages, disadvantages, use of their scope and value.It is proved that the plan in image pretreatment is feasible based on experiments. This system programmed by Visual C++, and operation was simply and has useful value.