Font Size: a A A

Scan Mode And Reconstruction Algorithm Of Large Field Of View In Helical Industrial Cone-beam CT

Posted on:2011-05-11Degree:DoctorType:Dissertation
Country:ChinaCandidate:X B ZouFull Text:PDF
GTID:1118360308457832Subject:Instrument Science and Technology
Abstract/Summary:PDF Full Text Request
Computed tomography (CT), not damaged the inspected object, can obtain two-dimensional (2D) or three-dimensional (3D) images of the internal object. It can directly and clearly display the structure, material and defects of the inspected object with images. Industrial CT has been used in aeronautics and astronautics, military, railway, automobile manufacture, etc. It is an important non-destructive testing technology in industry. Cone-beam CT, acquired data by use of density planar detector, can be used for rapid volumetric imaging with high longitudinal resolution and for efficient utilization of x-ray source. It can obtain more than thousand images with one turn scanning. Compared to cone-beam CT with circular trajectory, helical cone-beam CT can continuously inspect long workpiece, and its trajectory satisfies the sufficient condition of complete data.Conventional helical cone-beam CT assumes that the entire cross-section of the object is illuminated with x-rays at each view angle. The field of view (FOV) is limited by size of the planar detector. In industrial CT, large workpiece is sometimes inspected. At this time, large planar detector must be used. Unfortunately, the size of planar detector is limited by technology. This dissertation mainly studies the scan mode and reconstruction algorithm for inspecting large workpiece with helical cone-beam industrial CT and small planar detector. The main work is as follows.This dissertation deeply analyses the half-cover scan mode of the FOV and reconstruction algorithm for helical cone-beam CT. Before helical scanning, the gantry, which carries the inspected workpiece, translates given distance along the detector row, and more than half of the cross-section of the tested object is covered by x-ray beam at each view angle. Workpiece rotates around the axis of the gantry. The source and detector translate along the axis of the gantry simultaneously. The half-cover data can be obtained. The FOV of the half-cover helical cone-beam CT can be greatest extended up to about two times.FDK algorithm, based on filtered backprojection (FBP), is high efficient, practical and approximate algorithm. This algorithm needs the x-ray beam to cover corss-section completely at each view angle, and the center ray of cone beam intersects and is perpendicular to the axis of gantry. Because the center ray departs from the axis of gantry in half-cover helical cone-beam CT, off-center FDK algorithm is proposed. When reconstructing, data are weightedly backprojected according to the position of projection. Experimental results validate that off-center FDK algorithm of the half-cover helical cone-beam CT can get complete image of the tested object with small helical pitch, and the reconstruction time can be saved about half of the conventional FDK algorithm because of reduction of projections.Contrary to filtered backprojection (FBP) algorithm, backprojection filtration (BPF) algorithm firstly backprojects the differential data to PI line, and then filters along PI line by use of Hilbert inverse transform. It can reconstruct object function from transversely truncated data. This dissertation proposed rebinning-based BPF algorithm for half-cover helical cone-beam CT due to the transversely truncated data at each view angle. The proposed algorithm firstly rebins the half-cover helical cone-beam data to half-cover 2D fan-beam data, and then weighted BPF algorithm is used. Experimental results validate that rebinning-based BPF algorithm can get complete image of the inspected field with less truncated artifact, and reconstruction time is less than the off-center FDK algorithm, although projections are rebinned in proposed algorithm.The cross-section of workpiece is partly covered by x-ray at each view angle in half-cover helical cone-beam CT. Dual-helical cone-beam CT, which can scan large FOV, is proposed. When scanning, gantry and workpiece are translated given distance along detector row, which is similar to the half-cover helical cone-beam CT. Workpiece rotates around the axis of the gantry, and the x-ray source and detector are translated along axis of gantry simultaneously. After the first helical data are acquired, the workpiece is translated again along the detector row, which direction is contrary to the first one. Then the workpiece rotates, and the x-ray source and detector are translated along the axis of the gantry simultaneously. The second helical data can be obtained. Although the cross-section of inspected field is partly covered by x-ray beam at each view angle for each helical scanning, the cross-section can be completely covered by two helical scanning at each view angle. Radius of FOV of the dual-helical cone-beam CT relates to geometry parameters of system. The Radius of FOV can be extended up to 1.77 times with parameters of this dissertation.FDK-type algorithm is proposed for dual-helical cone-beam CT, which does not rebin projections and is high efficient as same as conventional helical FDK algorithm. When reconstructing, each projection is weightedly superposed according to the backprojected position. The reconstruction image can be acquired. Computer simulations validate that good image reconstructed by FDK-type algorithm can be got with appropriate pitch, and the algorithm is low consumed. It can extend the FOV up to 1.77 times.Katsevich algorithm is theoretically exact algorithm. Compared with Grangeat-type algorithm, Katsevich algorithm is high efficient. The dissertation proposed Katsevich-type algorithm for dual-helical cone-beam CT, which does not rebin projections, too. When reconstructing, the backprojected data are weightedly superposed according to proposed dual helical Katsevich-type algorithm. The image of the inspected filed can be obtained. Computer simulations validate that cone-angle, pitch, and increased translation distance have less impact on Katsevich-type algorithm, Although the helical pitch is large, the proposed algorithm can obtain good reconstructed images.This dissertation improves the dual-helical scan mode and proposes dual helical BPF-type algorithm in order to reduce the transversely truncated error. The proposed algorithm firstly backprojects data along PI lines. The differential data are weightedly backprojected, which patches the missing data on the PI line. And then the Hilbert inverse transform is implemented along the PI line. Computer simulations validate that images reconstructed by BPF-type algorithm are good. Cone angle, pitch, and increased translation distance have less impact on BPF-type algorithm, too. The reconstruction images have less truncated artifacts. Furthermore, the FOV can be extended up to about two times if the BPF-type algorithm and improved dual-helical cone-beam are used.
Keywords/Search Tags:Industrial CT, half-cover cone beam, dual-helical cone beam, image reconstruction
PDF Full Text Request
Related items