Research On Defect Extraction Of X-Ray Welding Image

X-ray photography defect detection is a widely used detection method in the industry application, with its advantage of straight, reliability, high sensitivity and distinguishability. With the detection result, we can ensure the defect type> positon and size to judge whether the workpiece is qualified. Generally, the X-ray film is judged by the experience specialists under the light.However, by using eyes to observe the film and judge the welding quality, it will take a lot of the work load and bring harm to the eyes due to the hard light as well as low efficiency. On the other hand, the detection level is effected by the subjective factor and astable. As a result, digitzing and automatic judging becomes a tendency.However, due to some characteristics of X-ray detection, X-ray image generated have the characteristic of low contrast, defect edge indistinction, more noise and more background fluctuation, which makes defect extraction and segmentation becomes a big trouble. Therefore, in this paper, Research is done on the X-ray welding defect extraction. By learning the idea from Digital Subtraction Angiography, the original image is used to subtract the image which imitate the background of the original image, then threshold segmentation is implemented on the previous result, then shape of defect can be got.Because the tiff image can't be displayed on the windows platform and the windows platform can just display the images whose gray scale is between 0 and 255, in this article, gray mapping function is designed, which helps to map the gray scale between 0 and 65535 to 0-255.Content researched of this article as follow:1. Research is done on ray mapping of X-ray welding image, by mapping the gray scale to 0-255,which helps the image to display on the windows platform;2. This article leam the idea form application of digital subtraction in angiography, apply digital subtraction to welding defect extraction and do research on X-ray welding defect extraction.Achievements of this article as follow:Implementation of ray transformation of X-ray welding image, which helps the image to display on the windows platform. In the meanwhile, by combining noise removing.background imitation,image subtraction and threshold segmentation, Implementation is done on the defect abstraction,which prepare for the later recognition.