Research On Artifacts Correction For Industrial Cone-beam Computed Tomography

As a new industrial non-destructive testing technology, cone-beam X-ray Computed Tomography (CT) imaging technology has higher spatial resolution and faster scanning speed, which can be used in three-dimensional non-destructively imaging for industrial parts and precision instruments, and has important applications in the industrial, aerospace and other fields. But because of the problems such as beam hardening and X-ray scattering, the reconstructed 3D images are usually seriously affected by artifacts. So the artifact correction technique for cone-beam CT has great importance to improve the quality of images, and has become one of the difficult and hot spots in the research of industrial CT.In CBCT imaging system for industrial metal parts, scattering artifacts and metal artifacts are two types of common artifacts. So this dissertation focuses on these artifacts and applies digital phantoms and real CBCT experiments to verify mutually, and systematically studies the formative factors of scattering artifact and metal artifact, and then proposes corresponding artifacts correction methods. The contributions obtained in this thesis can be summarized in the following 3 aspects:1. Single factor analysis method is adopted for artifacts in CBCT, the characters of scattering and metal artifacts under different factors are obtained. Because of the complexity of the formative factors of artifacts in CBCT, this thesis investigates scattering artifact and metal artifact with Monte Carlo method, and the formats of artifacts under single factors are given after physics calculation and digital phantoms simulation. Theory analysis and experiment simulation indicate that, the scattering artifacts represent as poor contrast ratio and cup artifacts; the metal artifacts could represent as different forms such as radial artifacts and dark streaking artifacts in different conditions.2. Based on the characters of scattering artifacts in CBCT, a scatter correction method combining beam attenuation grid and scatter sampling is proposed. The correction method divides the X-ray scatter intensity into scatter background and alternate scatter, creates a scatter correction model, obtains scatter 2D distribution, and removes the scatter from original projections. Experiment results indicate that, the proposed scatter method can effectively reduce scatter artifacts in reconstructed images, and the new scatter brought by attenuation grid is avoided.3. In order to eliminate the inherent metal artifacts of flat objects, a correction method based on angular compensation is proposed. Images of flat objects such as Printed Circuit Boards (PCBs) usually have serious metal artifacts, so the characters of metal artifacts for flat objects are investigated and a correction method is proposed to solve this problem. Based on the result of Monte Carlo analysis, the metal artifact correction method considers the effects of beam hardening, scattering and opacity phenomena, obtains correction table from pre-scanning projections, and compensates data by angles in the process of reconstruction. Result of experiments shows that the proposed metal artifact correction method can reduce the dark streaking artifacts dramatically, and rebuild the character information of objects which was contaminated by artifacts. Compared with the interpolation correction method, the proposed metal artifacts correction has a distinct advantage from the speed and effect.