大型高能工业CT系统的研制与开发是清华大学创建世界一流大学的重点科研项目，同时获得了国家自然科学基金支持。目前基于加速器为辐射源的大型工业CT系统在我国尚处空白，由于美国对能量高于2MV 的无损检测设备进行禁运，独立开发高能工业CT系统对国家安全绝对重要。作者在参与高能工业CT系统预研工作中，较系统地对CT重建相关问题进行了研究，取得一定成果。作者根据原型化算法开发过程的思路设计了一个被命名为STM（Simulative Tomographic Machine）的仿真平台。STM平台具有很好的扩展性和交互性，在平台上可以进行各种算法的研究。断层图像的获取与重建涉及复杂的物理和数学过程，为得到实际可用的高可靠性系统，必须设计并优化多种处理算法。STM平台可以简化算法设计的工作量，同时使算法研究有延续性。同时在该平台实现了CT系统工作过程的模拟，可以单独研究不同的物理现象，便于开展相关研究工作。论文还系统研究了高能工业CT中可能应用的各种常用重建方法，包括傅里叶变换重建、卷积反投影重建、代数重建技术以及统计重建方法等。而后利用集装箱检查系统实现了高能CT扫描，在STM平台上实现上述重建方法，并分析了不同算法在高能CT重建中应用的优缺点，为后面研究工作的开展打下良好基础。论文最后对大型高能工业CT系统性能评价的方法和指标进行了讨论。在分析常规图像指标的定义和计算的基础上，针对工业无损检测需求提出了夹杂可测度和空隙可测度两个新的系统指标定义，并给出相应的理论计算方法。这两个指标评价的结果都是一个无量纲数值，直接反映对相应缺陷检测能力的高低，可直接用于性能比较。论文还分析了实际系统中影响系统缺陷检测能力的各主要因素，包括有效束流宽度、光子统计效应、阵列探测器不一致性、电子学噪声、探测器窜扰以及单脉冲强度稳定性等。同时给出预测实际系统关键指标的理论计算方法，为系统参数设计提供理论依据。作为大型高能工业CT系统的预研课题之一，本文研究工作为后续工作开展奠定了坚实的基础。
Research and development of high energy industrial computed tomography is supported by National Natural Science Foundation (NSFC), it is also the priority project of Tsinghua University.The prototype instrument development of ICT system is a key research work for national defence. The main tasks of the research are to study on the theory of tomograpic reconstruction. The thesis carries out several creative designs and ideas, and puts them into practice.The author has developed the Simulative Tomographic Machine (STM), which can simplify the design procedure of algorithm and make the work continuable, in the research of data processing. We can design many practical methods on the platform of STM. In order to solve the problems brought by instrument implementation separately, We simulated the working process of CT scanning. The general reconstruction algorithms, which may be used in the high energy ICT system, have been studied systematically. The large container inspect system(LCIS) has been used to acquire high energy CT scanning data. Then we test several algorithms such as, Fourier Reconstruction method, Convolution Back-Projection method, Algebraic Reconstruction Technology and Statistic Methods. Trough the analysis of the reconstruct result, we have many benefits for the next step research work.The thesis also discussed how to evaluate the performance of the high energy ICT system. Based on the analyzing of definition and computation of the regular reconstruction image character, proposed two new factors. Then the thesis give the compute method of those factors. Furthermore, we discussed the relationship between the system’s parameters and the inspection factors. We set the research work focus on several parameters, such as, the equation beam width, the photon statistics noise, the inconsistent of array detector, the electronic noise, the crosstalk of detectors, the stabilization of accelerator. Those above work has established a stable base for next step research work.
