The Study On Edge Extraction Algorithm And Applications On Radiographic Images

Ray imaging technology is a new subject growing point that is based on multi-disciplinary cross and penetration. As a comprehensive high and new technology, ray imaging technology has been widely used in medical, industrial, life science, material science, national defense, security and other fields. Ray can be divided into two categories depending on the type of radiation ray imaging: One kind is particle radiation, including the neutrons, protons, alpha particles, beta particles, etc. Another is electromagnetic radiation, including x-rays and gamma rays. Ray imaging can show the inside information of the tested object through the image, that has extensive and important applications in the industrial inspection and medical tests. But in practical application, the image is not clear due to the limitation of radiation intensity and testing environment, it is difficult to obtain internal implied useful information of the image. As a result, the image processing method has a pivotal position in the preprocessing of ray imaging.The features that show local characteristics of the tested object can’t be got and well observed due to there always exist severely degradation problems in radiography. Therefore, how to quickly and effectively extract ray image texture edge is an important part in the study of X-ray imaging processing method.Local binary pattern(LBP, local binary pattern), is a rapid image texture extraction method, can extract ray image texture edge, but it cannot effectively distinguish the slight variation of superfluous gray because it is sensitive to noise. Now there exist an effectively edge extraction algorithm H-LBP, can extract the image edge effectively by improving the LBP operator. But the algorithm framework is more complex than LBP because of containing too much math floating-point calculations process, and extracting time is longer, which is difficult to meet the requirements of industrial real-time detection. In order to make the detection more quickly and efficiently, this paper presented a new modified accelerated algorithm-AH-LBP by analyzing framework of H-LBP algorithm and optimizing the H-LBP algorithm framework. In addition, this paper also proposed a fast edge extraction method based on local binary pattern, called M-LBP for short in this paper. The new algorithm adopts a counting strategy on the basis of local binary pattern(LBP) to eliminate influence of the tiny gray variation; choose the median of counting results as a global threshold to avoid the complex process of selecting optimal parameters.The actual experiments can illustrate: the accelerated algorithm AH-LBP consumes less time than H-LBP on the premise of the extraction results are same; the new edge extraction method M-LBP can distinguish noise and edge texture dispense with parameter selection, can effectively extract the edge of the X-ray imaging. The above two kinds of edge extraction algorithm have good application prospect in industrial nondestructive testing and medical detection.