This invention discloses an MRI passive shimming optimization method based on the PSA-SQP
hybrid algorithm, comprising the following steps: (1)
Magnetic field mapping to evaluate the uniformity of the
magnetic field: the
magnetic field non-uniformity at a specific sampling location within the imaging area is evaluated using a
magnetic field probe / camera; (2) Establishing an optimization model for passive shimming; (3)
Hybrid algorithm optimization of passive shimming: combining the
pattern search algorithm PSA with sequential quadratic
programming SQP to find the optimal solution for the iron block in the passive shimming
system. PSA is run in advance to perform a global search, obtaining elite individuals as initial values for SQP optimization, and then the optimal solution for the iron block in the passive shimming
system is found by running SQP. This algorithm can achieve
high magnetic field uniformity and also has limitations on the weight of the shimming iron sheet. Importantly, it can achieve a sparse distribution of the shimming iron sheet.