Research On Image Reconstruction Algorithm For Propeller Technique

MRI (Magnetic resonance imaging technology) is an important tool for research inthe field of biology and medicine. Magnetic resonance imaging has been widely used,which plays a big role in clinical diagnosis of human disease. MRI has high resolutiontomography，which has no radiation and can be in any direction,etc. With thecontinuous development of imaging techniques, we have become increasinglydemanding on the imaging quality and the image reconstruction speed. The hardwarelevel, the human tolerance limit and gradient switching speed are restricted by imagingtime, which due to that the patients often have independent or autonomous movement，resulting in imaging artifacts that affects the doctor’s diagnosis. How to eliminatemotion artifacts has become the hot point of the technology of magnetic resonanceimaging study.PROPELLER imaging technology has been successfully applied in brain magneticresonance imaging, which has great effect on the elimination of the rigid motionartifacts significantly.Firstly, the article analyzes imaging theory of the MRI and image reconstructionmethods. It research the principle of periodic rotation overlapping parallel lines captureand enhanced post-processing reconstruction algorithm,then analysis on non-Cartesiandata acquisition and reconstruction algorithm.After that,the article realize Propeller image reconstruction algorithm, and deeplyanalysis several key issues for implementation. Because of the particularity of the datacollected for the MRI which is non-Cartesian data, the article explore for theimplementation of the algorithm used in several key operations, including reading rawdata show K trajectory, phasing correction, rotation correction, non-flute Carlreconstruction of the data, the choice of the density compensation function, etc.Finally, the image is reconstructed successfully by the propeller imagingalgorithm.The propeller algorithm is tested by the Shepp-logan model to collect theacquisition of K-space data, the rebuilded results were compared.We realize Propller image algorithm by MATLAB.This software can dealwith.MRD data reading, data accessing, Motion Estimation and Density compensationfunction.It has good help for higher research to deeply study the methods of Eliminationof motion artifacts and Non Cartesian data reconstruction.