Exploitation of geometric a priori knowledge for limited data reconstruction in nondestructive testing

Computed tomography (CT) is a very powerful tool in medicine and non-destructive testing but is unsuitable for planar objects. A solution can be found in the use of computed laminography (CL), a technique where the object is irradiated by an oblique angle thereby circumventing the problems arising in CT. Due to the limited amount of angular coverage and the special geometric set-up, filtered backprojection methods [4] cannot be employed for the reconstruction in this case. More flexible iterative algorithms like SART (simultaneous reconstruction technique) [1,2] provide an answer to this challenge. One of their important advantages when compared to filtered backprojection methods is their ability to incorporate a priori information about the object into the reconstruction process [6]. Often the object's geometry is known from CAD files or other technical specifications. Especially in the case of limitedangle data, where only a part of the object can be measured, and laminographic geometries, additional information is of great importance. This geometrical a priori knowledge can be exploited to restrict the reconstruction volume to areas where material is definitely present, resulting in correct object contours even in the limited-angle case. This reduces artifacts and increases contrast thereby allowing for a better defect detectability and thus an easier and more reliable inspection of the object