Research Of High Quality Slice Images Reconstruction Methods Of Cone-beam CT

Industrial CT(ICT) is a testing technology that obtains the inner information including geometric structure and material ingredient of objects by reconstructing gray images of the slices. Due to the merit of not influencing of by the object's material and structure, ICT is widely applied into non-destructive testing (NDT), evaluation and reverse engineering in many fields. It is considered as the best NDT approach in the modern world. FDK method is the most common kind of 3D image reconstruction algorithm in cone beam CT at present, which is based on a scanning circle and reconstructs images satisfactorily in the small cone angle case. However, with increasing cone angle, FDK algorithm would produce significant image artifacts since the circular trajectory does not satisfy the so-called data sufficiency condition. The cone angle is strictly limited in the application, for example, the large cone angle must be used to scan large workpiece. In the situation, the exact reconstruction algorithm of three nonplanar orbits is studied, and study the calibration method to eliminate artifacts in reconstructed slices with large cone angle in the paper. The main research content and achievement is articulated as follows:1. The Radon transform of the projection is calculated in the filtered process of exact reconstruction, Siddon and Simpson integral are used to sample the pixels to get the projection. In order to get high filtered precision, the resolution of the reprojection images are twice as precise as the projection images that are used to project, therefore the filtered images are high precision and the reconstruction images are improved remarkably.2. The ideal ramp filter and the ideal derivative filter are not implemented, therefore the filters with the rectangle window are used in the filter process, the quality of reconstruction images is degressive because of Gibbs disorder owing to the abrupt truncation. The rectangle window, the sinc window, the hamming window and the new window are added to the ideal filter. The differences of the space-frequency character are compared among the different filters, the result will help people to choose appropriate filter in different reconstruction case.3. The exact reconstruction regions of several kinds of orbits are analyzed by the Radon shell graphic method. It is illuminated that the exact reconstruction region of single circular orbit is zero, and the exact reconstruction regions of the three nonplanar orbits change along with the changing of the orbits' parameters. It is directive meaning for scanning to acquire the exact reconstruction region.4. The main reason is the radon space data of scanned object is not complete, which will result in the image artifacts when using FDK algorithm to reconstruct with moderate or large cone angles. Therefore the cone angle calibration algorithm based Grangeat's formula is presented to compensate the incomplete data of the Radon space. The first sub-result is calculated by applying the FDK algorithm to the projection data acquired by a single circular orbit scan, the second sub-result is a correction received by applying SVFBP algorithm to the missing Radon data acquired by extrapolation, the final result is the sum of the two sub-results. The simulate result demonstrates the algorithm is valid method to reduce artifacts of the images reconstructed by FDK algorithm.