The Algorithms Study Of Reconstructing High Resolution Images From Overlapped Low-resolution Projection Data

Computed Tomography technique(CT) has become a new nondestructive detection method in recent decades. The images reconstructed by CT are of good quality and high resolution. Furthermore, CT images can be processed and analyzed digitally. So CT is widely applied in aviation, astronautics, mechanics, shipping, public security, customs, medicine, and many other areas.The main assumption in the paper is that the linear array of detectors can shift along the line(or arc) of the array with a step which is less than the width of the detectors, and the overlapped low-resolution projection data can be collected by overlapped sampling.In this paper, two novel algorithms of reconstructing high resolution images from both overlapped parallel projection data and overlapped fan beam projection data with low resolution are presented. One is called virtual detector algorithm, and the other is direct reconstruction algorithm. Both of the algorithms with two forms for the parallel data and fan beam data respectively are developed based on the interpolative algebraic reconstruction techniques (IART).The simulation results show that, either for parallel geometry or for fan beam geometry, high resolution images can be reconstructed by means of both the algorithms from the overlapped projection data. Moreover, the quality of the images reconstructed by the virtual detector algorithm from the overlapped low-resolution projection data is almost as good as that of the images reconstructed directly from the high-resolution projection data. Although the resolution of the images reconstructed by the direct reconstruction algorithm is slight lower than that of the images reconstructed by thevirtual detector algorithm, the direct reconstruction algorithm adapts to any kind of overlapped projection data.