Research On Technology Of Identification For Laser Welding Defects Of Aluminum Alloy Based On X-ray Real-time Imaging

Laser welding technology of aluminum alloy is widely used in the manufacture of aeronautic and astronautic structural parts such as aircraft panel structural. For the detection of laser welding defects, the manual evaluated results of traditional X-ray testing is limit to film itself and has a big subjectivity. With the rapid development of new generation X-ray real-time imaging technology and image pattern recognition technology, the computer aid detection for internal defects in laser welded structure of aluminum alloy based on X-ray digital image has become possible. This paper has carried out the research on X-ray real-time imaging technology, image preprocessing and recognition technology. The automatic extraction and identification of laser welding defects of aluminum alloy have been preliminarily realized.First of all, the X-ray real-time imaging system for laser weldment of aluminum alloy with T-joint has been designed and built on the basis of reasonable selection of hardware configuration. Through imaging system, the X-ray real-time imaging technology of laser weldment of aluminum alloy with T-joint has been systematically studied, and the optimal imaging parameters were obtained. The tube voltage is 60 kV, current is 0.3mA, focal length is 350 mm and the best magnification is 2.4.Secondly, aimed at the characteristics of original X-ray image, the image preprocessing module, which includes image gray-scale transformation function, image noise reduction function and image fuzzy enhancement function, is used to improve the quality of X-ray image. Among them, the way of combining single methods of noise reduction has improved the effect of de-noising for mixed noise. The traditional fuzzy enhancement algorithm was subsequently improved to enhance the contrast of image, and weldment X-ray image with good quality has been provided for the extraction and identification of weld defects.Also, considering the analysis to the column gray curve of X-ray image, the methods of curve fitting and gray-value differential have been used to complete the extraction of weld area under complex background. Then, after background simulating by the method of adaptive morphological filtering, the algorithms of image subtraction and iterative threshold segmentation have been used to realize the segmentation of the weld defects. Aubsequently, the algorithms of contour extraction and seed filling was used to realize the extraction of the weld defects.Finally, all kinds of characteristic parameters have been extracted and calculated on the basis of defects region labeling. According to the characteristic parameters of defect and other X-ray image characteristic, expert system based on forward inference engine and fuzzy reasoning was developed to identify and classify the weld defects automatically. Among them, experience and knowledge in knowledge base of expert system exists as the form of rules and users can complete the operation of modifying, adding and deleting to the rules. At the same time, according to relevant standards, this paper has achieved the assessment to circular defects of laser welding of aluminium alloys.Aimed at the characteristics of laser weldment of aluminum alloy with T-joint, this paper has achieved the research on X-ray real-time imaging technology, X-ray image processing and defect identification technology by the experiments on X-ray real-time imaging and image processing simulation through computer programming, which provides a new approach for nondestructive testing of laser welding defects of aluminum alloy.