Computed tomography resolution enhancement by integrating high-resolution 2D X-ray images into the CT reconstruction

The resolution of computed tomography of flat structures (e.g. printed circuit boards) is limited by the need to rotate the object by at least 180 degrees. 2D X-ray images can achieve higher resolutions but show only a part of the object and contain data for several different layers, making it hard to do further image processing steps like edge detection. This paper presents an approach which acquires both a standard CT image data set and a set of high-resolution 2D images, which are taken with the object much closer to the source, and then merges these data sets by using them as input for an iterative reconstruction algorithm. The resulting high-resolution voxel data set provides high-resolution information in directions parallel to the structure and, unlike the raw 2D images, the layers are separated. The approach presented here is verified using a simulated data set of a printed circuit board with line pair structures and an edge used for determining the modulation transfer function. The improvement of the resolution achieved with this approach is roughly 4.5 for the considered test structure.