Image Reconstruction Of Industrial Computed Tomography And Three Dimensional Visualization Technology

Industrial Computerized Tomography(ICT). which is short for industrial computerized tomography image reconstruction technique that has risen rapidly in the early1980s and been widely used in industrial nondestructive testing(NDT). is a new imaging technology derived from the combination of radiology and computer technology. Industrial Computerized Tomography (ICT) can obtain the internal structure and information of the workpiece without damaging the shape and structure of the samples by utilizing the penetrability of X or y rays.3-Dimensional visualization technique of industrial CT image reconstruction is able to truly display the3-dimensional structure of the object. Through analyzing these processed images, we can not only get information about defects in2-dimensional slice images, but also find the specific location and distribution of the defects in3-dimensional space in order to deal with the defects.In this paper, we firstly described the fundamental principle of industrial computerized tomography technique (CT), and then discussed several classical methods among some CT reconstruction algorithms. CT reconstruction methods are mainly divided into two categories:the analytic method and the iteration method. The former of two mentioned is one of the most widely used algorithms in reconstruction field, which can be implemented by using the Fourier transform and the Fourier slice theorem. The Filter Back Projection (FBP) reconstruction is the most classical method among analytic algorithms. This method has some advantages of high reconstruction speed and simple calculation, but the reconstructed image will have low resolution and bad reconstruction result because of the perfectionism of this method and the obvious noise characteristics in the projection data. The iteration method consists of algebraic iteration reconstruction algorithm and statistical iteration reconstruction algorithm. Algebraic iteration reconstruction algorithm is similar to the analytic method in essence, and they both are derived from the Radon transform. Statistical iteration reconstruction method is based on the statistical model and can correct the reconstructed images according to optimization criteria.In the reconstruction process of industrial CT images, there will inevitably be some noise that will result in the deterioration of the reconstructed image when the projection data are generated because of the influence of the radiation dose, the equipments and operations. In order to solve this problem, we proposed an improved noise reduction algorithm in projection domain on the basis of partial differential equation in this paper. This method describes the local features of gray value distribution in the neighborhood by introducing the local entropy of the neighborhood sized by3×3. And the concept of the local entropy of image is also introduced into partial differential model. The improved anisotropic diffusion model is employed to reduce noises in projection domain in this paper. Experimental simulation results show that the novel algorithm is able to improve SNR of reconstructed image while preserving the details and edges of the image with better effects.And then, we will describe the surface rendering method and volume rendering method in the process of3-dimensional visualization of industrial CT images Comparison between the two methods will be made. The principle and the implementation of Marching Cubes (MC) which is the most representative algorithm of the surface rendering method will be described in details. And in the following sections, we will write vtkMarchingCubesSpace class that implement the surface rendering method after extracting the isosurface from2-dimensional tomographic image data so that we can solve the issue that the vtkMarchingCubesSpace class in MC algorithm can only extract isosurface from volume data. Thus the VTK libraries will be extended and the implementation of3D visualization will become easier.At the end of this paper, on the basis of aforementioned information we developed and completed industrial CT image reconstruction and3D visualization system on a PC by using VTK software development kit on Visual Studio2008programming platform. This3D visualization system allows3-dimensional volume data to be inputted,2-dimentional industrial CT images to be displayed and digital images to be processed. In addition, the achievement of surface rendering, volume rendering and virtual cutting will be reached, and we can rotate, pan and zoom in or out the displayed images or process them with other interactive operations through the3D visualization system.