【作者】 刘传杨； 李兴东；

【Author】 LIU Chuanyang;LI Xingdong;School of Mechanical Engineering and Automation,BeihangUniversity;National Institute of Metrology;

【机构】 北京航空航天大学机械工程及自动化学院； 中国计量科学研究院；

【摘要】 工业CT技术作为一种无损检测手段在众多领域中得到了广泛的应用。锥束CT因其射线利用率高、扫描时间短等特点逐渐成为CT领域的研究热点。但是,锥束CT的应用也存在一定的局限性,主要包括成像视野小、重建速度慢等。本文提出大视场锥束CT重建的LFOV算法,采用转台偏置的方式进行扫描,利用基于CUDA编程架构的多GPU编程实现重建过程。实验证明本文的方法在大视场锥束CT的重建过程中是有效的,既可以增大成像视野也可以大大提高重建速度。本文的方法可以满足工业应用的需求,可以在工程应用中广泛推广。更多还原

【Abstract】 As a kind of nondestructive test method,industrial computed tomography technology has been widely used in many fields.Cone beam CT gradually becomes the research hotspot in the field of CT due to its advantages such as efficient utilization of X-ray photos and short scan time.However,the application of cone beam CT also has certain limitation,mainly including small field of view and slow reconstruction speed.In this paper,an algorithm for large field of view cone beam CT reconstruction is proposed.Turntable offset scan and multi-GPU programming based on CUDA architecture were used to realize large field of view fast reconstruction.Experimental results proved that the method could not only enlarge the field of view but also greatly improve the reconstruction speed.更多还原