The Research Of X-ray Negative Digital Image De-nosing Method Based On Wavelet Transform

Radial detection is one of the important methods for non-destructive detection. This technique can reflect the continuity structure in the object by the imaging characteristic and the attenuation of the radial, such as X ray,γ-ray, neutron ray and so on, when it penetrate through the object. As the film data obtained by the traditional detection has those disadvantages such as hard storing, inaccuracy and so on, the digital processing of radial detection becomes the task for discussing.Under the artificial or electron device caused influence, we get the digital images with much noise in the process of radial detection. And this caused serious interruption for the analysis of digital image. The paper aims at the characteristic of X ray digital film and researches for the method of de-noising by the way of wavelet transform.The paper expounds the digital process of X ray film and the feature of the noise in it, by which we estimate the parameter of noise.It investigates the traditional de-noising method for X-ray digital image such as time space and transforms space. And analysis the advantage and disadvantage of the de-noising methods using different wavelet transform. Producing the Gibbs phenomena in the feature-point Produced in the process of traditional wavelet threshold de-nosing, we propose the de-noising method using stationary wavelet transform. It compares with stationary wavelet transform (SWT) and discrete wavelet transform (DWT). The result proves that the former is better. The paper proposes a improved threshold de-noising method based on SWT, which using the sub-image to estimate the parameter of the noise, thus it can select different threshold for different types of X-ray digital film. Experiment results indicate that this method make the image more clear and velvet, and improved the PSNR greatly. This paper applies the principle of Wiener filtering to the process of X ray digital image de-noising based on stationary wavelet, provides a method of X ray digital image de-noising based on stationary wavelet in Wiener filtering. Experiment shows that now method not only can eliminate the Gibbs phenomenon, but also can retains detail information of X ray digital mage.