Industrial Ct Broaden Our Methods And Image Artifacts Correction Study

Indstrial CT (computed tomography) is a kind of nondesrtuctive testing technology. Aimed to develop indstrial CT system, we research several problems arising from CT image reconstruction, including the field-of-view extension and image artifacts correction. The main work and innovation of this paper are as follows.(1) In industrial CT, it is often required to inspect large objects using short line-detetor or small size planar-detetor. In this paper, we propose fan-beam RT (rotation-translation) multi-scan with the turntable placed unilaterally, and relative BPF-based (back-projection-filtration-based) accurate reconstruction algorithm. The propose algorithms can reconstruct CT image directly without data rebinning process which is introduced in the existing algorithms. Hence it is of high computational efficiency and image spatial resolution. In addition, for scanning mode with the turntable placed unilaterally, the projection data required for image reconstruction are nearly half reduced. As a consequence, the field-of-view is enlarged under the same times of rotation and translation. Another feature of the proposed algorithm is that it is suitable for acceleration by GPU (graphic processing unit). We generalize the scanning mode and the algorithm above to circular cone-beam CT. The generalized algorithm kept the features of the algorithm for fan-beam RT scanning mode, except that it is approximate for image reconstruction in off-center plane. The algorithms mentioned above are verified by both the simulation and real CT data.(2) For large industrial CT equipment, it is difficult to adjust the scanning machinery parts accurately. In this paper, we analyze formation mechanism of artifacts by simulating reconstruction process with geometric inaccuracy of the machinery parts. And then, we conduct the displaced error of iso-center into line-detector deflection, and then derive a reconstruction formula including a deflection parameter of the line-detector. Finally, we give a method to estimate system parameters included in the reconstruction formula. The numerical experiment results show that the proposed formula enable us to reconstruct an accurate image directly from the data scanned on a CT device with a displaced iso-center and a deflection of the line detector.