Research On Key Technology Of Weld Defect Detection

In the non-destructive testing industry,weld defect testing is an important branch.In this field,radiographic testing has become the most important non-destructive testing methods due to its advantages such as detection of internal defects in welds,detection of almost all types of defects,accurate quantitative analysis,and long-term and low-cost storage of test results.At present,most X-ray imaging equipment uses planar photosensitive elements.However,imaging equipment using linear photosensitive elements has its irreplaceable advantages.Linear scan can greatly improve the resolution of the image,and then significantly improve the quality of the image,making the defect detection results more reliable under high-precision requirements.However,the image obtained by using line scan will have some special problems.This article shows stripe noise on a linear scan image.Its cause is related to the manufacturing process and operating principle of the equipment,and it is unavoidable under the current technical level.In order to detect weld defects through X-ray scanning images,these stripe noises must be reduced to a low level.We designed a denoising method based on image pyramid based on the characteristics of fringe noise,which can effectively filter the fringe noise on the image..In addition,the measurement of the unsharpness of the radiographic image is also a necessary pre-step in the process of weld defect detection.The unsharpness of the image is the minimum actual distance that can be distinguished on the image.The measurement of unsharpness in engineering has undergone a transition from using single-wire image quality meter to using double-wire image quality meter.Based on the research of the relevant standards and the characteristics of the area of double-wire image quality meter on image,we proposed an automatic image unsharpness detection algorithm based on gradient detection and weighted average.This algorithm greatly improves the detection efficiency of image unsharpness.Due to the obvious difference between the gray value of the porosity defect area and the surrounding area,we propose a defect detection algorithm based on connected domain search and edge detection.We use the least squares method for circle fitting,so as to locate the center and edge positions of the porosity defect more accurately.Technological progress should facilitate the automation and intelligence of industrial production.Under the background that related companies lack a weld defect detection platform that integrates image acquisition,image processing,image calibration and other functions,we have developed a highly integrated weld defect detection system.The system applies the algorithms proposed in this article.The system has been put into practical use and received good feedback.