Cone beam CT has a high spatial and temporal resolution, rapid data collection,utilization of high radiation, large scanning range, etc., which is important for medicaland industrial. Cylindrical and flat panel detectors have be more widely used inmedical diagnostic imaging and industrial transmission ray imaging. In this paper,3Dcone-beam CT image reconstruction method is researched systematically, accordingto the characteristics of the cone-beam projection data based on the Cylindrical andflat-panel detectors. The main contents of the paper include the following:To improve the speed of image reconstruction based on fan-beam filtered backprojection, according to the transformation relationship between Cartesian coordinatesand polar coordinates, this paper proposes a fast symmetry reconstruction algorithmfor the back-projection algorithm of the polar coordinates for fan-beam collinearequispaced CT based on the features of trigonometric function. The experimentalresults show that, without sacrificing image quality, compared with the traditionalfiltered back-projection algorithm, this method can improve the reconstruction speedby more than8times. This method has been successfully applied to the reconstructionof3D cone-beam.The cone-beam FDK reconstruction algorithm is introduced the principle of thealgorithm and implementation process based on the scanning cone flat panel detector.According to the transformation relationship between Cartesian coordinates and polarcoordinates, and the symmetry of the pixel location related parameters, this paperproposes a SPFDK reconstruction algorithm. The experimental results showed that,FDK algorithm is optimized by the two kinds of measures, compared with traditionalFDK algorithm, the speed of this method can be improved8times, if using CUDAtechnique, it will be increased40times on GPU without any new error. The cone-beam reconstruction algorithm is introduced based on the scanningcone Cylindrical detector. Firstly, it is introduced he principle of the algorithm andimplementation process, which is named two-dimensional reconstruction algorithmfor fan-beam collinear equispaced CT. Then this paper elaborate AMPR algorithmsand specific implementation steps, the reconstruction plane is not perpendicular to theZ axis, which is a certain title angle. Projection data is disposed to parallel beam in thevertical plane, then the parallel beam reconstruction algorithm is used to reconstructthe image. The experimental results show that, the percentage of detector data usedfor image reconstruction is about90%, the image reconstruction quality isindependent of the pitch within a certain range change of pitch. |