The reconstruction of PROPELLER MRI data is a typical reconstruction of non-Cartesian data.The sampling density compensation in the current gridding reconstruction is accomplished through the convolution of non-Cartesian data

which requires calculating a large number of distances between each other non-Cartesian positions when the number of data is large such as PROPELLER.Therefore the density compensation is highly computationally intensive.In this paper

a new algorithm was proposed for density estimation based on gridding a unity data vector.Since no more computations of distances were needed

the time complexity of the algorithm was decreased greatly.In the experiments

the proposed algorithm is shown to be able to perform the density compensation much more rapidly than the traditional algorithm while achieving equivalent quality of reconstruction.