Abstract: A method is disclosed for determining distortion information describing geometric distortion during the recording of magnetic resonance image data with a recording sequence, wherein a magnetic resonance image data set of at least one calibration phantom with a first marker structure having a defined geometry which is visible in magnetic resonance imaging is recorded. The at least one calibration phantom further includes a second marker structure having a defined geometry which is visible in computed tomography imaging. A computed tomography image data set of the calibration phantom is recorded and the distortion information is determined by a comparison of the magnetic resonance image data set with the computed tomography image data set, in particular at least partially automatically.