This article deals with the exploitation of magnetic susceptibility artifacts in magnetic resonance imaging (MRI) for the recognition of metallic delivery capsules. The targeted application is a closed-loop position control of magnetic objects implemented using the components of a clinical MRI scanner. Actuation can be performed by switching the magnetic gradient fields, whereas object locations are detected by an analysis of the MRI scans. A comprehensive investigation of susceptibility artifacts with a total number of 108 experimental setups has been performed in order to study scaling laws and the impact of object properties and imaging parameters. In addition to solid metal objects, a suspension of superparamagnetic nanoparticles has been examined. All 3D scans have been segmented automatically for artifact quantification and location determination. Analysis showed a characteristic shape for all three base types of sequences, which is invariant to the magnetic object shape and material. Imaging parameters such as echo time and flip angle have a moderate impact on the artifact volume but do not modify the characteristic artifact shape. The nanoparticle agglomerates produce imaging artifacts similar to the solid samples. Based on the results, a two-stage recognition/tracking procedure is proposed.