The Analysis And Improvement Of The Automated Fluorescent Magnetic Particle Testing System

Fluorescent magnetic particle testing is a NDT method that often used for testing trivial flaws on the surface of ferromagnetic materials in order to control the qualities of product. At present, semiautomatic device for fluorescent magnetic particle testing has been largely used. But the flaws are recognized by manual work. There are a mass of factors which affect the testing and the work condition is atrocious. So it is important to set up a system of automated fluorescent magnetic particle testing. This paper mainly describes the theory of the automated fluorescent magnetic particle testing, including the image sampling, image segmentation, image mathematic morphologic pretreatment, thresholding division, label, eigenvalue extraction and flaw recognition. And analyses all kinds of image processing methods and gives the judges to them. According to the characteristic of the sampled image, I put forward the improved methods about the self-adaptive dynamic thresholding algorithm. The new algorithm can treat with the problem which the noise is distended by the old algorithm and can deal with the losing of information. At last, this paper provides a set of method of the automated fluorescent magnetic particle testing, and puts forward firstly the method about the recognition of air hole and oil spots.