| abstract |
The invention belongs to the technical field of instruments used for radiological diagnosis, and discloses a GPU-based fast cone beam computed tomography reconstruction method, comprising: acquiring object measurement data from multiple angles; acquiring various parameters of a computed tomography system; Data preprocessing; weighted filtering of preprocessed data, GPU-based rearrangement filtering algorithm; back-projection of filtered data, GPU-based back-projection using texture mapping theory and the symmetry of the FDK algorithm itself to achieve parallelism Accelerate, complete computed tomography. The present invention is based on the filtering back-projection reconstruction method, utilizes prior information such as system parameters, adopts X-ray irradiation, limited angle measurement, and utilizes GPU strategy to effectively reduce the reconstruction time without changing the original image quality. It has important application value in the field of 3D cone beam reconstruction algorithm of computed tomography. |